USF Libraries
USF Digital Collections

Associations between maternal depression and child social competence and display of problem behaviors

MISSING IMAGE

Material Information

Title:
Associations between maternal depression and child social competence and display of problem behaviors a longitudinal investigation of direct, indirect and moderating effects
Physical Description:
Book
Language:
English
Creator:
Zapata, Lauren B
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
Postpartum
Mental health
Child development
Maternal sensitivity
Attachment theory
Dissertations, Academic -- Public Health -- Doctoral -- USF   ( lcsh )
Genre:
government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: Postpartum depression is a significant public health problem facing women, children, and families in the United States with an estimated 10-15% of U.S. mothers experiencing an episode of non-psychotic depression within six months of delivery. The purpose of this study was to investigate the associations between maternal depression during the first three years postpartum and child social competence and display of problem behaviors at first grade. The impact of several characteristics of maternal depression were examined including general exposure, timing of initial onset in the postpartum period, severity of symptoms along the trajectory of initial onset, and chronicity of symptoms. This study also explored the mediating and moderating influences of maternal sensitivity, as well as the moderating influence of exposure to nonmaternal care.This study was based upon secondary analysis of data from the National Institute of Child Health and Human Development Study of Early Child Care (NICHD SECC), a multi-site, prospective, three-phase longitudinal study. The sample consisted of 679 infants and their families that participated in both Phase I and Phase II of the NICHD SECC, had compete data on all variables of interest in the study, and had non-depressed mothers at first grade. Results identified the first six months postpartum as a sensitive period of risk for depression initiation. Severity of symptoms was also found to be important. In some instances depression alone did not increase risk for lower levels of social competence, but severity of symptoms above cut points indicating depression did. Chronic depression at 24 months rather than 36 months postpartum was found to pose the greatest magnitude of negative influence on outcome.
Thesis:
Thesis (Ph.D.)--University of South Florida, 2005.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Lauren B. Zapata.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 514 pages.
General Note:
Includes vita.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001680948
oclc - 62385567
usfldc doi - E14-SFE0001174
usfldc handle - e14.1174
System ID:
SFS0025495:00001


This item is only available as the following downloads:


Full Text

PAGE 1

Associations Between Maternal Depression and Child Social Competence and Display of Problem Behaviors: A Longit udinal Investigation of Dir ect, Indirect and Moderating Effects by Lauren B. Zapata A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Community and Family Health College of Public Health University of South Florida Major Professor: Melinda S. Forthofer, Ph.D. Jeannine Coreil, Ph.D. Stanley N. Graven, M.D. Peter A. Gorski, M.D. Charles S. Mahan, M.D. Date of Approval: April 11, 2005 Keywords: Postpartum, mental health, ch ild development, maternal sensitivity, attachment theory Copyright 2005, Lauren B. Zapata

PAGE 2

Dedication To my boys – Paolo Kent, Noah Alexander, and Mauricio

PAGE 3

Acknowledgements I would like to acknowledge a nd thank, from the bottom of my heart, the following individuals who supported me during this process: Dr. Melinda S. Forthofer, for your e ndless support and assistance with the conceptualization and writing of this doctoral dissertation, te aching me everything I know about research methodology and data analysis, serving as a professional role model, and your dear friendship over the years. Your me ntorship has been invaluable and I owe all my success to you! Dr. Jeannine Coreil for taking me under your wing during my Master’s program, introducing me to program evaluation, challenging me intellectually (especially in theory class!), and your heartfelt support and encourag ement of me as a doctoral student with small children! Dr. Peter A. Gorksi for expanding my knowledge on the importance of human relationships in the growth a nd development of children. Dr. Stanley N. Graven for teaching me to critically analyze public health problems and think outside the box when identif ying effective behavior change strategies. Dr. Charles S. Mahan, my first profes sor in maternal and child health, for introducing me to the topic that would driv e my professional career. Thank you for your positive energy and confidence in my abilities. My parents, Frederick R. Bailey and Bobbie J. Bailey, for their endless support and encouragement of my academic endeavors. Thank you for such an amazing childhood and providing me the self confidence to reach for the stars. Can you believe I am finally done?

PAGE 4

My brother Kent L. Bailey, who was an amazing role model growing up, and his wife Uzma who just makes us all so happy to be around her. My in-laws, Dr. Carlos A. Zapata, Gl oria Zapata, and Melissa Zapata, who have accepted me into their family and provided much love and support through the years. Now there will be three Dr. Zapata’s in the family! And last but not least, my husband and be st friend Dr. Mauric io Zapata. You have always encouraged me to achieve all of my goals, and never once made me feel that my academic and career aspirations fell sec ond to yours. I admire your strength in character and compassion toward living creatures of all shap es and forms. Thank you for your endless support, encouragement, and unders tanding during this painstaking process. This dissertation would not have been possible without you!

PAGE 5

i Table of Contents List of Tables v List of Figures x Associations Between Maternal Depression and Child Social Competence and Display of Problem Behaviors: A Longit udinal Investigation of Dir ect, Indirect and Moderating Effects xvii Chapter One: Introduction.................................................................................................1 Statement of the Problem........................................................................................1 Need for the Study..................................................................................................6 Implication for Public Health..................................................................................7 Overview of Study Methods...................................................................................9 Research Hypotheses............................................................................................10 Delimitations.........................................................................................................11 Limitations............................................................................................................12 Definitions.............................................................................................................13 Chapter Two: Literature Review.....................................................................................18 Postpartum Depression.........................................................................................18 Range of Postpartum Mood Disorders......................................................18 Postpartum Blues............................................................................19 Postpartum Psychosis......................................................................19 Postpartum Depression...................................................................20 Predisposing Risk Factors.........................................................................23 Biological Factors...........................................................................23 Obstetric Factors.............................................................................25 Cognitive or Intrapersonal Factors..................................................25 Interpersonal Factors.......................................................................27 Structural Factors............................................................................30 Personal Psychopathology..............................................................32 Summary.........................................................................................32 Attachment Theory...............................................................................................34 Introduction...............................................................................................34 Historical Context.....................................................................................34 Theoretical Formulations..........................................................................36 Modern Attachment Theory......................................................................38 Attachment, Attachment Behavior, and the Attachment Behavioral System.......................................................................................................38

PAGE 6

ii Development of Attachment.....................................................................40 Maternal Sensitivity and Internal Working Models..................................43 Summary...................................................................................................47 Effects of Postpartum Depression on Child Development...................................48 Introduction...............................................................................................48 Social Competence....................................................................................52 Chronicity.......................................................................................54 Lack of Consideration for De pression Characteristics...................57 Summary.........................................................................................61 Behavioral Difficulties..............................................................................64 Chronicity......................................................................................65 Severity..........................................................................................73 Timing of Onset..............................................................................74 Lack of Consideration for De pression Characteristics...................76 Summary...................................................................................................84 Limitations of Current Research...............................................................88 Conclusions...............................................................................................92 Mechanisms of Influence......................................................................................93 Effects of Postpartum Depression on Maternal Sensitivity......................95 Effect of Maternal Sensitivity on Child Development.............................98 Moderating Influence of Exposure to Nonmaternal Care.......................103 Women in the Workforce..............................................................103 Hierarchy of Attachment...............................................................104 Summary.................................................................................................108 Chapter Three: Methodology.........................................................................................110 Purpose of Study.................................................................................................110 Research Hypotheses..........................................................................................110 Overview of Study Design..................................................................................112 Overview of NICHD Study of Early Child Care................................................112 NICHD SECC Population and Sample...............................................................115 Target Population and Sampling Frame.................................................115 Sampling Plan.........................................................................................116 NICHD Sample: Phase I.........................................................................118 Limitations of Study Sample..................................................................120 NICHD SECC Instrument ation and Study Variable Selection...........................123 Sociodemographic Control Variables.....................................................123 Maternal Depression...............................................................................127 Maternal Sensitivity................................................................................132 Social Competence at First Grade...........................................................138 Display of Problem Behaviors................................................................149 Exposure to Nonmaternal Care...............................................................160 NICHD SECC Data Co llection Methods............................................................161 Data Scoring........................................................................................................164 Data Analysis......................................................................................................166

PAGE 7

iii Univariate and Bivariate Analyses..........................................................167 Multivariate Analyses.............................................................................169 Design Limitations..............................................................................................172 Chapter Four: Results....................................................................................................175 Study Sample......................................................................................................175 Univariate Analyses............................................................................................180 Bivariate Analyses..............................................................................................189 Multivariate Analyses.........................................................................................242 Modifications to the Plan........................................................................242 Overview.................................................................................................243 Ever Exposure to Maternal Depression..................................................244 Hypothesis 1.1...............................................................................244 Hypothesis 2.1...............................................................................247 Hypothesis 3.1...............................................................................256 Hypothesis 4.1...............................................................................258 Timing of Initial Onset............................................................................260 Hypothesis 1.2...............................................................................260 Hypothesis 2.2...............................................................................268 Hypothesis 3.2...............................................................................274 Hypothesis 4.2...............................................................................278 Severity of Symptoms.............................................................................286 Hypothesis 1.3..............................................................................286 Hypothesis 2.3...............................................................................292 Hypothesis 3.3..............................................................................294 Hypothesis 4.3..............................................................................298 Chronicity of Symptoms.........................................................................305 Hypothesis 1.4...............................................................................305 Hypothesis 2.4...............................................................................319 Hypothesis 3.4..............................................................................342 Hypothesis 4.4...............................................................................348 Support for Hypotheses...........................................................................365 Chapter Five: Discussion...............................................................................................369 Synthesis of Research Findings..........................................................................369 Impact of Maternal Depression...............................................................369 Mediating Effect of Maternal Sensitivity...............................................376 Moderating Effect of Maternal Sensitivity.............................................382 Moderating Effect of Exposur e to Nonmaternal Care............................385 Attachment Theory.............................................................................................388 Study Limitations................................................................................................391 Study Strengths...................................................................................................394 Implications for Public Health............................................................................396 Suggestions for Future Research........................................................................408

PAGE 8

iv References .................................................................................................................410 Appendix A: Pearson Correlation Matrices for Maternal Depression Over Time, Maternal Sensitivity Over Time, and Maternal Depression and Maternal Sensitivity Over Time.....................................................................................................437 Appendix B: Partial Model Results from OLS Regression Testing the Impact of Ever Depressed on Child Social Competen ce and Display of Problem Behaviors at First Grade: Significant Main Effects of Ever Depressed...........................................439 Appendix C: Partial Model Results from OLS Regression Testing the Impact of Timing of Initial Onset of Maternal De pression on Child Social Competence and Display of Problem Behaviors at First Gr ade: Significant Main Effects of Timing of Initial Onset............................................................................................................... .442 Appendix D: Partial Model Results from OLS Regression Testing the Impact of Depression Severity Along the Trajectory of Initial Onset on Child Social...................459 Competence and Display of Problem Behaviors at First Grade:....................................459 Significant Main Effects of Depression Severity............................................................459 Appendix E: Partial Model Results from OLS Regression Testi ng the Impact of Depression Chronicity on Child Social Competence and Display of Problem Behaviors at First Grade: Significant Ma in Effects of Depression Chronicity..............467 About the Author...................................................................................................End Page

PAGE 9

v List of Tables Table 1. DSM-IV Criteria for a Major Depressive Episode. .........................................22 Table 2. Selected Characteristics of NICHD SECC Sample from Hospital Screening Recruitment Data............................................................................120 Table 3. Comparison of Phase II Participants Versus Drop Outs..................................122 Table 4. Sociodemographic Control Vari ables Selected for Study Inclusion...............125 Table 5. Items Comprisi ng the CES-D Scale.................................................................128 Table 6. Maternal Depression Variab les Selected for Study Inclusion.........................132 Table 7. Brief Description of Maternal Sensitivity Behaviors Assessed Via the Mother Child Interaction Procedure.................................................................134 Table 8. Measurement of Maternal Sensitivity..............................................................137 Table 9. Coefficient Alpha and Stability Reliability Coefficients for the Elementary Level SSRS...................................................................................141 Table 10. Measurement of Social Competence.............................................................144 Table 11. Brief Description of Item Content by Syndrome Scale of the CBCL/4-18 150 Table 12. Measurement of Problem Behaviors..............................................................154 Table 13. Impact of Study Inclusion Criteria on Final Study Sample Size...................175 Table 14. Comparison of Study Participants with Complete Data Versus Excluded Children and Families with Incomplete Data................................................177 Table 15. Comparison of Study Participan ts with Non-Depressed Mothers at Grade One Versus Excluded Children and Families with Depressed Mothers at Grade One....................................................................................179 Table 16. Sociodemographic Characteri stics of Study Sample (n=679).......................181 Table 17. Means, Standard Deviations, and Ranges for Dependent Variables.............182

PAGE 10

vi Table 18. Categorical Classifications of Child Social Competence and Display of Problem Behaviors.........................................................................................184 Table 19. Univariate Statistics fo r Maternal Depression Variables...............................185 Table 20. Univariate Statistics for Maternal Sensitivity Variables................................186 Table 21. Statistics for Exposure to Nonmaternal Care Variables................................187 Table 22. Univariate Statistics for Expos ure to Center-Based Nonmaternal Care Variables..................................................................................................... ...188 Table 23. Chi-Square Results of Ever Depressed by Sociodemographic Control Variables. ................................................................................................... ...190 Table 24. Chi-Square Results of Maternal Depression at One Month Postpartum by Sociodemographic Control Variables.......................................................191 Table 25. Chi-Square Results of Maternal Depression at Six Months Postpartum by Sociodemographic Control Variables.......................................................192 Table 26. Chi-Square Results of Maternal Depression at 15 Months Postpartum by Sociodemographic Control Variables.......................................................193 Table 27. Chi-Square Results of Maternal Depression at 24 Months Postpartum by Sociodemographic Control Variables.......................................................194 Table 28. Chi-Square Results of Maternal Depression at 36 Months Postpartum by Sociodemographic Control Variables.......................................................195 Table 29. ANOVA Results of Maternal Depression Scores Over Time by Maternal Age.................................................................................................1 96 Table 30. ANOVA Results of Maternal Depr ession Scores Over Time by Sociodemographic Control Variables.............................................................198 Table 31. ANOVA Results of Maternal Sensitivity Scores Over Time by Sociodemographic Control Variables............................................................202 Table 32. Maternal Report of Child So cial Competence by Sociodemographic Control Variables...........................................................................................20 6 Table 33. Teacher Report of Child Social Competence by Sociodemographic Control Variables 207

PAGE 11

vii Table 34. Mother Report of Child Pr oblem Behaviors by Sociodemographic Control Variables...........................................................................................20 8 Table 35. ANOVA Results of Child Social Competence and Display of Problem Behaviors by Maternal Age...........................................................................209 Table 36. ANOVA Results of Child Social Competence and Display of Problem Behavior s by Sociodemographic Control Variables......................................211 Table 37. Ever Exposed to Nonmaternal Care at One Month by Sociodemographic Control Variables...........................................................................................21 5 Table 38. Ever Exposed to Nonmaternal Care at Six Months by Sociodemographic Control Variables..........................................................................................21 6 Table 39. Ever Exposed to Nonmaternal Care at 15 Months by Sociodemographic Control Variables...........................................................................................21 7 Table 40. Ever Exposed to Nonmaternal Care at 24 Months by Sociodemographic Control Variables...........................................................................................21 8 Table 41. Ever Exposed to Nonmaternal Care at 36 Months by Sociodemographic Control Variables...........................................................................................21 9 Table 42. Ever Exposed to Center-Based Nonmaternal Care at One Month by Sociodemographic Control Variables............................................................220 Table 43. Ever Exposed to Center-Based Nonmaternal Care at Six Months by Sociodemographic Control Variables............................................................221 Table 44. Ever Exposed to Center-Bas ed Nonmaternal Care at 15 Months by Sociodemographic Control Variables............................................................222 Table 45. Ever Exposed to Center-Bas ed Nonmaternal Care at 24 Months by Sociodemographic Control Variables............................................................223 Table 46. Ever Exposed to Center-Bas ed Nonmaternal Care at 36 Months by Sociodemographic Control Variables............................................................224 Table 47. ANOVA Results of Ever Exposur e to Nonmaternal Care Over Time by Maternal Age...........................................................................................225 Table 48. ANOVA Results of Ever Exposed to Center-Based Nonmaternal Care Over Time by Maternal Age 226

PAGE 12

viii Table 49. Pearson Correlation Matrices fo r Each Continuous Dependent Variable a nd Maternal Depression Over Time.............................................................228 Table 50. ANOVA Results of Ever Exposed to Nonmaternal Care Over Time by Maternal Depression Scores Over Time........................................................231 Table 51. ANOVA Results of Ever Exposed to Center-Based Nonmaternal Care Over Time by Maternal Depression Scores Over Time.................................233 Table 52. Pearson Correlation Matrices fo r Each Continuous Dependent Variable and Maternal Sensitivity Over Time...............................................................236 Table 53. ANOVA Results of Ever Exposed to Nonmaternal Care Over Time by Maternal Sensitivity Over Time.....................................................................239 Table 54. ANOVA Results of Ever Exposed to Center-Based Nonmaternal Care Over Time by Maternal Sensitivity Over Time.............................................240 Table 55. Summary of Significant Findings from OLS Regression Testing the Impact of Ever Depressed on Child Social Competence...............................246 Table 56. Summary of Significant Findings from OLS Regression Testing the Impact of Ever De pressed on Child Display of Problem Behaviors..............247 Table 57. Tested Associations for Establ ishing a Mediating Effect of Maternal Sensitivity (MS).............................................................................................2 49 Table 58. Mediating Effect of Maternal Sensitivity in the Association Between Ever Exposure to Maternal Depression and Maternal Report of Total Social Competence.........................................................................................250 Table 59. Mediating Effect of Maternal Sensitivity in the Association Between Ever Exposure to Maternal Depression a nd Maternal Report of Assertion..................................................................................................... ...251 Table 60. Mediating Effect of Maternal Sensitivity in the Association Between Ever Exposure to Maternal Depression and Maternal Report of Peer Competence.................................................................................................... 252 Table 61. Mediating Effect of Maternal Sensitivity in the Association Between Ever Exposure to Ma ternal Depression and Ma ternal Report of Total Problem Behaviors.........................................................................................254

PAGE 13

ix Table 62. Mediating Effect of Maternal Sensitivity in the Association Between Ever Exposure to Maternal Depression a nd Maternal Report of Externalizing Problems...................................................................................255 Table 63. OLS Regression Results Testing the Moderating Effect of Maternal Sensitivity on th e Association Between Ever Exposure to Maternal Depression and Child Social Competence and Display of Problem Behaviors..................................................................................................... ..257 Table 64. OLS Regression Results Testing the Moderating Effect of Exposure to Nonmaternal Care on the Association Betw een Ever Exposure to Maternal Depression and Child Social Competence and Display of Problem Behaviors.........................................................................................259 Table 65. Summary of Significant Findings from OLS Regression Testing the Impact of Timing of Initial Onset of Maternal Depression on Child Social Competence and Display of Problem Behaviors.................................262 Table 66. Tested Associations for Establ ishing a Mediating Effect of Maternal Sensitivity (MS) in the Association Between Timing of Initial Onset of Maternal Depression and Child Social Competence and Display of Problem Behaviors.........................................................................................270 Table 67. Mediating Effect of Maternal Sensitivity in the Association Between Timing of Initial Onset of Maternal Depression and Maternal Report of Total Social Competence...........................................................................271 Table 68. Mediating Effect of Maternal Sensitivity in the Association Between Timing of Initial Onset of Maternal Depression and Teacher Report of Total Social Competence...........................................................................272 Table 69. OLS Regression Results Testing the Moderating Effect of Maternal Sensitivity on th e Association Between Timi ng of Initial Onset of Maternal Depre ssion on Child Social Competence and Display of Problem Behaviors.........................................................................................275 Table 70. OLS Regression Results Testi ng the Moderating Effect of Ever Exposure to Nonmat ernal Care on the Associat ion Between Timing of Initial Onset of Ma ternal Depression on Child Social Competence and Display of Problem Behaviors.......................................................................280 Table 71. OLS Regression Results Testi ng the Moderating Effect of Ever Exposure to Cent er-Based Nonmaternal Ca re on the Association Between Timing of Initial Onset of Maternal Depression on Child Social Competence and Display of Problem Behaviors............................284

PAGE 14

x Table 72. Summary of Significant Findings from OLS Regression Testing the Impact of Severi ty of Maternal Depressi on Along the Trajectory of Initial Onset on Child Social Competence and Display of Problem Behaviors..................................................................................................... ..288 Table 73. OLS Regression Results Testing the Moderating Effect of Maternal Sensitivity on the Association Between Depression Severity on Child Social Competence and Display of Problem Behaviors................................296 Table 74. OLS Regression Results Testi ng the Moderating Effect of Ever Exposure to Nonmat ernal Care on the Associat ion Between Depression Severity Along the Trajectory of Initial Onset on Child Social Competen ce and Display of Problem Behaviors...........................................300 Table 75. OLS Regression Results Testi ng the Moderating Effect of Ever Exposure to Cent er-Based Nonmaternal Ca re on the Association Between Depression Severity Along the Trajecto ry of Initial Onset on Child Social Competence and Display of Problem Behaviors......................303 Table 76. Summary of Significant Findings from OLS Regression Testing the Impact of Depr ession Chronicity on Child Social Competence and Display of Problem Behaviors........................................................................308 Table 77. Main Effect of Depression Ch ronicity on Maternal Sensitivity as a Preliminary Step for Establishing Maternal Sensitivity (MS) as a Mediating Effect in the Association Between Depression Chronicity and Child Social Competence and Display of Problem Behaviors................322 Table 78. Main Effect of Maternal Sensitiv ity on Select Dependent Variables of Social Competence and Display of Problem Behaviors as a Preliminary Step for Establishing Mediation.....................................................................324 Table 79. Partial OLS Regression Results Testing the Mediating Effect of Maternal Sensitivity at Six Months in the Association Between Depression Chronic ity at Six Months and Ma ternal Report of Total Social Competence at First Grade..................................................................327 Table 80. Partial OLS Regression Results Testing the Mediating Effect of Maternal Sensitiv ity at 15 Months in the Association Between Depression Chronic ity at 15 Months and Mate rnal Report of Total Social Competence at First Grade...................................................................328

PAGE 15

v Table 81. Partial OLS Regression Results Testing the Mediating Effect of Maternal Sensitivity at 36 Months in the Association Between Depression Chronic ity at 36 Months and Mate rnal Report of Total Social Competence at First Grade..................................................................329 Table 82. Partial OLS Regression Results Testing the Mediating Effect of Maternal Sensitivity at 36 Months in the Association Between Depression Chr onicity at 36 Months a nd Maternal Report of Cooperation at First Grade.............................................................................330 Table 83. Partial OLS Regression Results Testing the Mediating Effect of Maternal Sensitiv ity at 36 Months in the Association Between Depression Chronicity at 36 Months and Matern al Report of Assertion at First Grade................................................................................................. ..331 Table 84. Partial OLS Regression Results Testing the Mediating Effect of Maternal Sensitivity at 15 Months in the Association Between Depression Chr onicity at 15 Months a nd Maternal Report of Responsibility at First Grade.........................................................................332 Table 85. Partial OLS Regression Results Testing the Mediating Effect of Maternal Sensitivity at 24 Months in the Association Between Depression Chr onicity at 24 Months a nd Maternal Report of Responsibility at First Grade.........................................................................333 Table 86. Partial OLS Regression Results Testing the Mediating Effect of Maternal Sensitivity at 15 Months in the Association Between Depression Chronic ity at 15 Months and Mate rnal Report of Self Control at First Grade....................................................................................334 Table 87. Partial OLS Regression Results Testing the Mediating Effect of Maternal Sensitivity at 36 Months in the Association Between Depression Chronic ity at 36 Months and Mate rnal Report of Self Control at First Grade....................................................................................335 Table 88. Partial OLS Regression Results Testing the Mediating Effect of Maternal Sensitivity at 24 Months in the Association Between Depression Chronic ity at 24 Months and Mate rnal Report of Peer Competence at First Grade.............................................................................336 Table 89. Partial OLS Regression Results Testing the Mediating Effect of Maternal Sensitivity at 36 Months in the Association Between Depression Chronic ity at 36 Months and Mate rnal Report of Peer Competence at First Grade.............................................................................337

PAGE 16

vi Table 90. Partial OLS Regression Results Testing the Mediating Effect of Maternal Sensitivity at 15 Months in the Association Between Depression Chroni city at 15 Months and T eacher Report of Total Social Competence at First Grade..................................................................338 Table 91. Partial OLS Regression Results Testing the Mediating Effect of Maternal Sensitivity at 36 Months in the Association Between Depression Chronic ity at 36 Months and Mate rnal Report of Total Problem Behaviors at First Grade..................................................................339 Table 92. Partial OLS Regression Results Testing the Mediating Effect of Maternal Sensitivity at 24 Months in the Association Between Depression Chr onicity at 24 Months a nd Maternal Report of Externaliz ing Problem Behaviors at First Grade...........................................340 Table 93. Partial OLS Regression Results Testing the Mediating Effect of Maternal Sensitivity at 36 Months in the Association Between Depression Chr onicity at 36 Months a nd Maternal Report of Externaliz ing Problem Behaviors at First Grade...........................................341 Table 94. Summary of Findings from OL S Regression Testing the Moderating Effect of Maternal Sensitivity on the Associ ation Between Depression Chronicity on Ch ild Social Competence and Display of Problem Behaviors..................................................................................................... ..344 Table 95. Significant Interaction Terms of Depression Chronicity and Maternal Sensitivity (MS).............................................................................................3 46 Table 96. Summary of Findings from OL S Regression Testing the Moderating Effect of Ever Exposure to Nonmaternal Care on the Association Between Depre ssion Chronicity on Child Social Competence and Display of Problem Behaviors.......................................................................351 Table 97. Significant Interaction Terms of Depression Chronicity and Ever Exposure to Nonmaternal Care on the Association Between Depression Chronicity and Child Social Competence and Display of Problem Behaviors.........................................................................................353 Table 98. Summary of Findings from OL S Regression Testing the Moderating Effect of Ever E xposure to Center-Based Nonmaternal Care on the Association Be tween Depression Chronicity on Child Social Competen ce and Display of Problem Behaviors...........................................359

PAGE 17

vii Table 99. Significant Interaction Terms of Depression Chronicity and Ever Exposure to Center-Based (CB) Nonmaternal Care on the Association Be tween Depression Chronicity and Child Social Competen ce and Display of Problem Behaviors...........................................361 Table 100. Summary of Study Findings........................................................................366 Table A1. Depression Over Time, Matern al Sensitivity Over Time, and Maternal Depressi on and Maternal Sensitivity Over Time..........................................438 Table B1. Partial Model Results Testi ng the Effect of Ever Depressed on Maternal Report of Social Competence Including (1) Total Social Competence; (2) Assertion; and (3) Assertion 440 Table B2. Partial Model Results Testing th e Effect of Ever Depressed on Maternal Report of Child Display of Problem Behavi ors Including (1) Total Problem Behaviors; (2) Internalizing Problems; and (3) Externalizing Problems..................................................................................................... ..441 Table C1. Partial Model Results from OLS Regression Testing the Impact of Timing of Initial Onset of Maternal Depression on Maternal Report of Total Social Competence..............................................................................443 Table C2. Partial Model Results from OLS Regression Testing the Impact of Timing of Initial Onset of Maternal Depression on Maternal Report of Assertion................................................................................................. ..445 TableC3. Partial Model Results from OL S Regression Testing the Impact of Timing of Initial Onset of Maternal Depression on Maternal Report of Peer Competence...........................................................................................447 Table C4. Partial Model Results from OLS Regression Testing the Impact of Timing of Initial Onset of Maternal Depression on Teacher Report of Total Social Competence...............................................................................449 Table C5. Partial Model Results from OLS Regression Testing the Impact of Timing of Initial Onset of Maternal Depression on Teacher Report of Self Control:................................................................................................. ..451 Table C6. Partial Model Results from OLS Regression Testing the Impact of Timing of Initial Onset of Maternal Depression on Teacher Report of Peer Competence:.........................................................................................453

PAGE 18

viii Table C7. Partial Model Results from OLS Regression Testing the Impact of Timing of Initial Onset of Maternal Depression on Maternal Report of Total Problem Behaviors:..............................................................................455 Table C8. Partial Model Results from OLS Regression Testing the Impact of Timing of Initial Onset of Maternal Depression on Maternal Report of Externalizing Problem Behaviors:................................................................457 Table D1. Partial Model Results from OLS Regression Testing the Impact of Severity of Matern al Depression Along the Trajectory of Initial Onset on Matern al Report of Total So cial Competence..........................................460 Table D2. Partial Model Results from OLS Regression Testing the Impact of Severity of Matern al Depression Along the Trajectory of Initial Onset on Maternal Report of Self Control..............................................................461 Table D3. Partial Model Results from OLS Regression Testing the Impact of Severity of Ma ternal Depression Along th e Trajectory of Initial Onset on Maternal Report of Peer Competence...........................................463 Table D4. Partial Model Results from OLS Regression Testing the Impact of Severity of Ma ternal Depression Along th e Trajectory of Initial Onset on Teacher Report of Peer Competence:............................................464 Table D5. Partial Model Results from OLS Regression Testing the Impact of Severity of Ma ternal Depression Along th e Trajectory of Initial Onset on Maternal Report of Total Problem Behaviors...............................465 Table D6. Partial Model Results from OLS Regression Testing the Impact of Severity of Ma ternal Depression Along th e Trajectory of Initial Onset on Maternal Report of Externalizi ng Problem Behaviors:.................466 Table E1. Partial Model Results from OLS Regression Testing the Impact of Depression Chr onicity on Maternal Repor t of Total Social Competence:..................................................................................................4 68 Table E2. Partial Model Results from OLS Regression Testing the Impact of Depression Ch ronicity on Maternal Re port of Cooperation:.........................470 Table E3. Partial Model Results from OLS Regression Testing the Impact of Depression Ch ronicity on Maternal Re port of Assertion:.............................472 Table E4. Partial Model Results from OLS Regression Testing the Impact of Depression Chr onicity on Maternal Repor t of Responsibility:......................474

PAGE 19

ix Table E5. Partial Model Results from OLS Regression Testing the Impact of Depression Se verity on Maternal Repor t of Self Control:............................476 Table E6. Partial Model Results from OLS Regression Testing the Impact of Depression Chroni city on Maternal Report of Peer Competence:.................478 Table E7. Partial Model Results from OLS Regression Testing the Impact of Depression Chronicity on Teacher Report of Total Social Competence:......480 Table E8. Partial Model Results from OLS Regression Testing the Impact of Depression Chr onicity on Teacher Report of Self Control:...........................482 Table E9. Partial Model Results from OLS Regression Testing the Impact of Depression Chroni city on Teacher Report of Peer Competence:..................484 Table E10. Partial Model Results from OLS Regression Testi ng the Impact of Depression Chronici ty on Maternal Report of Total Problem Behaviors:..486 Table E11. Partial Model Results from OLS Regression Testi ng the Impact of Depression Chroni city on Maternal Report of Internalizing Problem Behaviors:.................................................................................................. ..488 Table E12. Partial Model Results from OLS Regression Testi ng the Impact of Depression Chr onicity on Maternal Report of Externalizing Problem Behaviors:................................................................................................. ..490

PAGE 20

x List of Figures Figure A. Conceptual Framework for Test ing the Effect of Maternal Depression on Child Social Competence and Display of Problem Behaviors with Mediating and Moderating Influences..............................................................95 Figure B. The NICHD SECC Sampling Process............................................................118

PAGE 21

xvii Associations Between Maternal Depression and Child Social Competence and Display of Problem Behaviors: A Longit udinal Investigation of Dir ect, Indirect and Moderating Effects Lauren B. Zapata ABSTRACT Postpartum depression is a si gnificant public health problem facing women, children, and families in the United States with an estimat ed 10-15% of U.S. mothers experiencing an episode of non-psychotic depression within si x months of delivery. The purpose of this study was to investigate the associations be tween maternal depre ssion during the first three years postpartum and child social comp etence and display of problem behaviors at first grade. The impact of several characte ristics of maternal depression were examined including general exposure, timing of initial onset in the postpartum period, severity of symptoms along the trajectory of initial onset and chronicity of symptoms. This study also explored the mediating and moderating infl uences of maternal sensitivity, as well as the moderating influence of exposure to nonmaternal care. This study was based upon secondary analysis of data from the Nati onal Institute of Child Health and Human Development Study of Early Child Care (NIC HD SECC), a multi-site, prospective, threephase longitudinal study. The sample consisted of 679 infants and their families that participated in both Phase I and Phase II of the NICHD SECC, had compete data on all variables of interest in the study, and had non-depressed mo thers at first grade. Results identified the first six months postpartum as a sensitive period of risk for depression initiation. Severity of symptoms was also found to be important. In some instances

PAGE 22

xviii depression alone did not increase risk for lower levels of so cial competence, but severity of symptoms above cut point s indicating depression did. Chronic depression at 24 months rather than 36 months postpartum wa s found to pose the grea test magnitude of negative influence on outcome. Maternal sens itivity partially mediated the negative impact of maternal depressi on on child outcome only when averaged across the first 36 months postpartum, or when assessed at la ter time points in the postpartum period (24 or 36 months. All mediations were modest in magnitude. Higher levels of maternal sensitivity buffered the negative impact of maternal depression among earlier onset episodes (six months), and ever exposure to nonmaternal care by 24 months was found to buffer the negative impact of chronic depression at 24 months on both mother and teacher reports of social competence.

PAGE 23

1 Chapter One: Introduction Statement of the Problem While more common than gestational diabetes, preeclampsia, and preterm deli very, postpartum depression has received much less attention in contemporary medical literature, training, and clinical practice. (Leopold & Zoschnick, 1996). Postpartum depression is a significant public health problem facing women, children, and families in the United States. Although the National Center for Health Statistics (NCHS) does not routinely colle ct prevalence estimate s of postpartum mood disorders in women, it is estimated that 10-15% of U.S. mothers experience an episode of non-psychotic depression within six months of delivery (O'Hara, 1995a; O'Hara & Swain, 1996), with most episodes developing within the first 2-6 weeks postpartum (Hendrick & Altshuler, 1999). It is also suggested that rates of postpar tum depression may be grossly underestimated due to stigma associated with mental disorders, lack of awareness among women, and lack of screening and diagnosis by medical professionals. Further, many women not meeting diagnostic criteria for non-psychotic depression during the postpartum period according to the 4th edition of the Diagnostic and Statistical Manual of Mental Disorders ( DSM-IV ) (American Psychiatric Association, 1994) also experience significant morbidity associated with mild symptoms of depression. Postpartum depression is an affective mood disorder that develops during the postpartum period and is characterized by f eelings of hopelessness, worthlessness, inadequacy as a parent, sleep and appetite disturbances, and impaired concentration

PAGE 24

2 (Miller, 2002; Misri, 2002). Other comm on symptoms described by women experiencing postpartum depression include excessive anxi ety over the child’s health, unc ontrollable crying, irritability, and lack of interest in the baby, (Le opold & Zoschnick, 1996; Spinelli, 1998; Walther, 1997). The average duration of postpartum depression is three months (Stover & Marnejon, 1995), however about one -half of women experiencing depression during the first three months postpartum c ontinue to be depressed at six months postpartum (Whiffen & Gotlib, 1993), and sympto ms may persist for a year or longer. Untreated postpartum depression can have l ong-term adverse effect s including recurrent depression for the mother; and emotional, behavioral, cognitive, and interpersonal problems later in life for the child (Miller, 2002). Morbidity associated with postpartum m ood disorders is often exacerbated by a lack of consensus among medi cal professionals in recognizi ng postpartum depression as a distinct and separate health risk to wome n, children, and families, as compared to nonpsychotic depression occurring at other times during a woman’s life. Rates of major and minor depression following delivery for chil dbearing women do not significantly differ from rates of major and minor depression for non-childbearing women (O'Hara, 1995b); moreover, the clinical manife stations of postpartum depre ssion have been found to be similar to those for non-postpartum depres sive episodes (Cooper, Campbell, Day, Kennerley, & Bond, 1988; O'Hara, Zekoski, Phi lipps, & Wright, 1990). Thus, there is disagreement about the importan ce of childbirth as a precipita ting factor in the onset of the depressive episode (Taylor, 1996). Although the second edition of the Diagnostic and Statistica l Manual of Mental Disorders (DSM-II) (American Psychiatric Associ ation, 1968) included a distinct

PAGE 25

3 diagnosis for postpartum mood disorders, “294.4 – Psychosis with Childbirth”, subsequent editions have eliminated this category under the premise that no compelling evidence supports postpartum depression as a distinct mental condition (Stern & Kruckman, 1983). The 4th edition of the DSM (American Psychiatric Association, 1994) classifies women with postpartum depression as experiencing a mood disorder (either major depression or bipolar disorder) with postpartum onset. The diagnostic classification is identical to individuals experiencing major de pression or bipol ar disorder outside of the postpartum wi ndow; however, the clarifier postpartum is used to provide additional diagnostic information to describe th e onset of occurrence w ithin the first four weeks postpartum. The limitation of sympto m onset to the fourth postpartum week is considered problematic to some clinicians because many consider symptom onset within the first 12 postpartum months related to childbirth, thus cases of depression with a postpartum onset (Misri, 2002; Se idman, 1998; Spinelli, 1998). Further, lack of or delay in diagnosis of postpartum de pression may be related to health care providers misinter preting reported symptoms su ch as sleep disturbance, decreased energy, or appetite/w eight change, all included as diagnostic criteria for major depression according to the DSM-IV as normal experiences of exhaustion and body changes due to delivery and childcare (Riecher -Rossler & Hofecker, 2003). For example, appetite and sleep disturbances fatigue, changes in weight an d anxiety over child’s health may all be experienced as signs of postpartu m depression but may also be interpreted as normal responses to pre gnancy and childrearing. An additional obstacle facing women in the diagnosis and treatment of postpartum depression is a delay in help s eeking behavior among suffering women for a

PAGE 26

4 number of reasons. According to a survey conducted by the Nati onal Mental Health Association on public attitudes and beliefs a bout clinical depressi on, more than one-half of women reported denial as a barrier to treatment and 41% of women cited embarrassment or shame as a barrier to treatme nt (National Mental Health Association, 1996). In a working-class Scottish samp le, one study found that women often do not realize they are suffering from depression but instead believe that they have failed as a mother, or fear treatment would lead to stigma tization as being mentally ill or result in being separated from their child (McIntosh, 1993). In summary, due to varying acceptance of postpartum depression as a separate and distinct entity from non-postpartum depr ession, discrepancies in diagnostic criteria across editions of the DSM, misinterpretations of reported symptoms as normal experiences for postpartum women, and feeli ngs of shame, stigma, and embarrassment experienced by women, postpartum depressi on often goes undiagnosed and untreated, which contributes to substantial morbidity each year in the lives of women, children, and families. Despite evidence of no significant differe nce between rates and symptom patterns of major and minor depression among childbearing and nonchildbearing women, research on postpartum depression as a distinct health risk is warranted for several reasons. First, women experiencing depres sion during the postpartum period are at increased risk for future depressive episodes. Studies have shown a significantly higher percentage of postpartum depressed women experiencing a subsequent major or minor depression during longitudinal follow up period s ranging from four and a half to five years compared to postpartum non-depresse d women (Cooper & Murray, 1995; Philipps

PAGE 27

5 & O'Hara, 1991). Given th at a large proportion of de pressions go undetected and untreated (Shapiro et al., 1984), and childbear ing women are often in regular contact with health care professionals (bot h obstetricians/gynecologists a nd pediatricians), targeted efforts to educate, screen, and refer for pos tpartum depression may help alleviate not only current, but future morbidity associat ed with depressive symptoms. Second, findings from one British study s uggest that depressions originating in the postpartum period take a unique course, de spite the exclusion of a separate category for childbirth-related mood disorders in the DSM-IV. Cooper and Murray (Cooper & Murray, 1995) found that among women whose postpartum depression was their first lifetime episode, the symptoms had a shorter dur ation, they were significantly less likely to experience a subsequent non-postpartum depression, and were more likely to experience a subsequent postpartum depressi on, as compared to postpartum depressed women who had experienced previous depressi ons. Further, in addition to classical symptoms of depression, intense mood sw ings, preoccupation with worries about the baby’s well-being as well as the woman’s own ability as a mother, have been described in episodes of postpartum depression (Brockington, 1996). Future research in this area will allow clarification regarding the extent to which postpartum mood disorders represent a significant health risk apart from depression occurring at other time periods and help establish routine screening for depression during the peri natal period as a national standard. Most important and critical for the aims of this study are the serious implications postpartum depression has on the mother’s abilit y to function effectively in the role as mother and caregiver. Postpartum depression poses a serious threat to the development

PAGE 28

6 of the child during infancy and early toddler hood. Research has in creasingly shown that the infant environment during the first years of life contributes significantly to optimal brain development, mother-i nfant bonding, infant capacities to trust and relate with others, and infant cognitive functioning incl uding focused learning and problem-solving skills (Karr-Morse & Wiley, 1997). Given a si tuation where a mother is depressed and limited in her ability to provide for and nurture her infant, negative developmental outcomes may emerge in the child (Jacobse n, 1999). Thus, although rates and symptom patterns of depression may not differ among postpartum and non-postpartum women, children’s needs for warm, sensitive and res ponsive caretakers warrant further research and attention to depression occurr ing during the postpartum period. Need for the Study Although specific symptoms of depression have been described and used to develop diagnostic criteria including depressed mood, fatigue, and feelings of worthlessness or guilt, it is important to stress that not all women experience the same symptoms of postpartum depression, and these symptoms may vary from mild to severe, and vary in symptom duration. Although postpar tum depression can manifest in different ways for individual women, the experience is nonetheless distressing and has the potential to damage the woma n’s self-esteem as a parent and her ability to care for, nurture, and bond with her infant, potentially l eading to negative outcomes in the child. This potential for negative developmental outcomes however is not equal for all postpartum women, given the heterogeneity in women’s experiences with depression. A tremendous limitation in the current knowledge base is the lack of consideration of the wide va riation in women’s experien ces with depression and the

PAGE 29

7 implications of such variation for understa nding the effects of postpartum depression on child development. For future research end eavors including the pres ent dissertation, it is critical to consider the course and character istics of the depressive episode, including timing on initial onset, severity of sympto ms, and chronicity of symptoms, when investigating the causal associations between postpartum depression and child development. One additional limitation of current research is the relatively little exploration of mediating and moderating infl uences between the impact of postpartum depression on child development, including social competence and display of problem behaviors. To identify effective strategi es for promoting child development in the context of postpartum depression, it is crit ical to understand the i ndividual, social and environmental circumstances in which exposure is likely to lead to negative outcomes. Implication for Public Health Understanding which aspects of postpartum depression pose the greatest risk to child social competence and display of problem behaviors is crucial at this time, not only to advance the current knowledge in the fiel d, but also to fully develop the prevention implications of early detec tion and treatment of postpartum depression. Public and mental health professionals need to k now if all women experiencing postpartum depression need to be targeted, or only t hose experiencing severe, longer duration, or early onset episodes. It is also importa nt to elucidate, in terms of prevention implications, which contextual risk factors increase the likelihood of negative outcomes, and what factors includ ing maternal sensitivity and expos ure to nonmaternal care mediate and moderate the association, given that not all children of depressed mothers experience adverse outcomes. For example, if matern al sensitivity is found to mediate the

PAGE 30

8 relationship between postpartu m depression and child social competence and display of problem behaviors at first gr ade, treatment options for women experiencing postpartum depression may include a parenting component aimed to enhance levels of sensitivity and responsivity, in addition to psyc hotherapy or pharmacology therapy. By more fully understanding the character istics of postpartum depression that place children at risk for developmental de ficits, as well as testing the mediating influence of maternal sensitivity and mode rating influence of nonmaternal care, the present dissertation aims to contribute to th e field of public health by: (1) advancing the state of knowledge; and (2) informing the deve lopment of health promotion programs and efforts to enhance child social competence and display of problem behaviors in the context of maternal depression.

PAGE 31

9 Overview of Study Methods This study is based upon secondary analysis of data from the National Institute of Child Health and Human Development Study of Early Child Care (NICHD SECC). The NICHD SECC is a multi-site, prospective, th ree-phase longitudinal study designed to explore the effects of expos ure to nonmaternal care dur ing infancy, early childhood, and middle childhood on child developmental outcomes. Ten study sites were selected to pa rticipate in the NICHD SECC based on competing scientific merit. A sampli ng frame of 8,986 women giving birth during selected 24-hour sampling periods at particip ating hospitals was es tablished. A threestage sampling plan through which a total of 1,364 families with full-term healthy newborns were enrolled to participate in Phase I. During Phase II, 1,223 families continued to participate in the study and were followed through first grade. Data for this study will consist of infants a nd their families who participated in both Phase I and Phase II of the NICHD SECC, and had complete contex tual data available for the variables of interest.

PAGE 32

10 Research Hypotheses The research hypotheses for this study are as follows. 1. Maternal depression is asso ciated with child social competence and display of problem behaviors at first grade. 1.1. Maternal depression, in general, during the first three years postpartum is negatively associated with child social competence and display of problem be haviors at first grade. 1.2. Timing of initial onset of mate rnal depression, during the first three years postpartum, is associated with child social competence and display of problem behaviors at first grade, with earlier onset episodes having th e greatest negative impact. 1.3. Severity of maternal depression along the trajectory of initial onset is associated with child social competence and display of problem behaviors at first grade, with severity of symptoms having the greatest negative imp act with earlier onset episodes. 1.4. Chronicity of maternal depression is associated with child social competence and display of problem behaviors at first grade, with chronicity of symptoms having th e greatest negative impact at later time points. 2. The relationship between maternal depression and child social competence and display of problem behaviors at first gr ade is mediated by maternal sensitivity.

PAGE 33

11 3. The effect of maternal depression on child social competence and display of problem behaviors at first grade is st ronger among mothers who rate low in maternal sensitivity. 4. The effect of maternal depression on child social competence and display of problem behaviors at first grade is weaker among children exposed to nonmaternal care. Delimitations Delimitations describe the population to which study results may be generalized (Locke, Wyric Spirduso, & Silv erman, 2000). The following delimitations are imposed on this study: 1. Results are only generalizable to the population of infants and their related families born in participating hospitals within the ten sites during 1991, given use of non-probabilistic methods for sample determination. 2. Results are only generalizable to infants born to mothers over the age of 18, those fluent in English, those with no serious medical conditions requiring postpartum intensive care, those with no disability including deafness or blindness, and those with no reported substance a buse, due to sample exclusi on criteria. Mothers were not excluded if they had a previous history of psychiat ric illness. 3. Results are only generalizable to single ton infants, those who experienced no medical complications requiring hospita lization beyond seven days, and those residing in neighborhoods considered sa fe for home visitation based on police reports, due to sample exclusion criteria.

PAGE 34

12 4. Results are only generalizable to infants and their families who participated in both phases of data collection and had nonmissing data for every variable included in the study. 5. Results are only generalizable to infants of mothers who did not have elevated depression scores at first grade. Limitations Limitations refer to limiting conditions or re strictive weaknesses inherent in the study design (Locke et al., 2000). The follo wing are limitations of this study: 1. Families that participated in both phases of data collection may be different from families that only participated in ph ase one and were lost to attrition. 2. Families in the study sample, based on avai lability of complete contextual data and non-presence of maternal depression at grade one, may be different from families excluded from the study sample due to some missing data or presence of maternal depression at first grade. 3. The study is limited to the scales, subscales and items that were included in the NICHD SECC. 4. Assessment of maternal depression using the Center for Epidemiologic Studies Depression Scale (CES-D) (Radloff, 1977) is limited to measurement of depressive symptomatology onl y, not clinical depression. 5. The assessment tool for maternal sensitivity was developed for the purposes of the NICHD SECC. The procedure entails traine d coders to rate videotaped play situation between mother and child. No psychometric evidence including interrater reliability was presen ted in NICHD SECC documentation.

PAGE 35

13 6. Data collection techniques including ho me interviews and laboratory-based observation may have increased social de sirability and resulted in decreased variation of some study constructs. 7. Data from Phase I began in 1991. Infa nts born in 2005 may be different from infants born in 1991 Definitions 1. Antenatal Period – the period from conception until the onset of labor, approximately 40 weeks. 2. Attachment –the deep and enduring c onnection or emotional bond established between infant and caregiver during th e first years of the child’s life. 3. Attachment Behavioral System – a behavioral system that mediates the display of attachment behaviors; functions to mainta in a sufficient level of proximity to the attachment figure for protection. 4. Attrition – the loss of subject s during the course of a study that may be a threat to the validity of study conclusions. 5. Attachment Behaviors – behaviors displayed by infants and young children including crying, smiling, vocalizing, clinging, and following, that mediate the formation of an attachment bond between child and caregiver. 6. Depression Chronicity – re fers to the duration of depression symptomatology; exposure can be either continuous or intermittent. 7. Depression Severity – the degree of ma gnitude of depression symptomatology.

PAGE 36

14 8. Depression Timing of Initial Onset – refers to the time period depression symptoms first emerged; operationalized in this study to represent the first time depression symptoms emerged in the postpartum period. 9. Depressive Symptomatology – symptoms of depression; may or may not be indicative of clinical depression. 10. Display of Problem Behaviors – problematic child behaviors indicative of adverse social development; operationalized in this study as both externalizing (undercontrolled) and internalizing (overcont rolled) problems. 11. Externalizing Problems – undercontrolled pr oblem behaviors that include a broad array of aggressive behaviors, antisocia l characteristics, and hyperactivity; operationalized in this study as deli nquent and aggressive behaviors; 12. Internalizing Problems – overcontrolled pr oblem behaviors that include a broad domain of symptoms related to depr ession, anxiety, social withdrawal, and somatic complaints; operationalized in this study as withdrawn behaviors, somatic complaints, and anxious/depressed behaviors. 13. Internal Working Model –an individual ’s internalized representation of a caretaker’s accessibility and responsivene ss based on prior inte ractions with the caregiver; serves as a template of interaction guiding future social relationships. 14. Intrusiveness – a dimension of maternal se nsitivity that represents the degree to which a mother imposes her agenda on her child as opposed to interacting in a way that provides a sense of control to the child. 15. Maternal Depression – depression occurring in mothers; experience of symptoms not limited to the postpartum period.

PAGE 37

15 16. Maternal Sensitivity –a mother who notices infant signals, interprets signals accurately, considers the infant ’s perspective, and responds to signals in a prompt and appropriate manner; operationalized in this study to repr esent the following maternal behaviors: sensitivity to nondistress, positive regard for the child, and intrusiveness. 17. Monotropy – the concept that infants tend to have a single primary attachment figure, usually the mother. 18. Nonmaternal Care – childcare provided by individuals other than the child’s mother. 19. Perinatal Period – the time period surrounding childbirth. 20. Positive Regard for the Child – a dimension of maternal sensitivity that is represented by the quality and quantity of a mother’s expressions to her child that connote positive feelings. 21. Postnatal Period – the period from the deliv ery of the placenta through the first six weeks after delivery; also known as the postpartum period. 22. Postpartum Blues –a mild adjustment di sorder experienced du ring the postpartum period resulting from feelings of le tdown after the em otionally charged experience of birth where the mother may feel “down,” cry for no apparent reason, experience instability of mood, sa dness, irritability, s ubjective confusion, insomnia, and anxiety. 23. Postpartum Period – the period from the de livery of the placenta through the first six weeks after delivery; also known as the postnatal period.

PAGE 38

16 24. Postpartum Depression – an affective m ood disorder that develops during the postpartum period and is characterized by f eelings of hopelessness, worthlessness, inadequacy as a parent, sleep and ap petite disturbances, and impaired concentration. 25. Postpartum Mood Disorders – emotional difficulties experienced by postpartum mothers associated with pregnancy; includes postpartum blues, postpartum psychosis, and postpartum depression. 26. Postpartum Psychosis – the most severe ty pe of postpartum mood disorder that is characterized by unstable mood, disorganiz ed behavior with confusion, sleep disturbances, delusions, and auditory hallucinations. 27. Sensitivity to Nondistress – a dimension of maternal sensitivity; extent to which mother-child interaction is characteri zed by prompt and appropriate maternal responses to the child’s social ge stures, expressions, and signals. 28. Singleton Infant – an infant born not as part of a multiple birth. 29. Social Competence – a complex construct representing five dimensions including peer relationships, self-management, acad emic skills, compliance, and assertion; operationalized in this study as cooperati on, assertion, responsibi lity, self-control, and peer competence. 30. Social Development – the way in which indi viduals’ interactions with others and their social relationships grow, change, a nd remain stable over the course of life. 31. Sociodemographic Variables – control vari ables including child sex, child birth order, maternal age, maternal race, maternal ethnicity, maternal education,

PAGE 39

17 maternal marital status, presence of hus band or partner in the home, and whether or not the family is on public assistance. 32. Syndrome Scales – broad dimensions of problem behaviors including withdrawn, somatic complaints, anxious/depressed, social problems, thought problems, attention problems, delinquent beha vior, and aggressive behavior;

PAGE 40

18 Chapter Two: Literature Review The purpose of this literature review is to introduce the reader to postpartum depression and attachment theory, and to exam ine the literature exploring the relationship between postpartum depression and child social competence and display of problem behaviors. Specific emphasis in this latter section will emphasize the need for further research to more fully consider the wide variation in women’s experiences with postpartum depression, as related to child social competence and display of problem behaviors. In addition, this literature revi ew will present empiri cal evidence supporting the conceptual framework guiding this study. In particular, resear ch on the effects of postpartum depression on maternal sensitivity and the effects of maternal sensitivity on child development will be reviewed. Lastly, literature on the effects of nonmaternal care will be examined. Postpartum Depression Range of Postpartum Mood Disorders There are three types of psychiatric m ood disorders experienced by women during the postnatal period including postpartum bl ues, postpartum depression and postpartum psychosis. These types of mood disorders va ry in terms of symptoms, severity and duration of illness, as well as treatment and prevention strategies. All three types of mood disorders will be described below, with differences between the types highlighted. The proposed study will focus on postpartum depression because of its significant prevalence and risk to child development.

PAGE 41

19 Postpartum Blues. The least severe and most transient perina tal mood disorder is postpartum blues. It is estimated that between 39-85% of all new mothers experience postpartum blues (Chokka, 2002; Kennerly & Gath, 1989; O'Hara, Schlechte, & Lewis, 1991) with rapid onset that usually begins on the first post partum day and peaks between the third and seventh postpartum day (Ugarriza & Robins on, 1997). Postpartum blues is a mild adjustment disorder resulting from feelings of letdown after the emotionally charged experience of birth where the mother may feel “down,” cry for no apparent reason, experience instability of mood, sadness, irri tability, subjective confusion, insomnia, and anxiety (Beck, 1992; Miller, 2002) The onset of postpartum blues often coincides with the beginning of lactatio n (Miller, 2002), and is thus believed to be a ssociated with abrupt hormonal changes. Symptoms of postpart um blues usually subside without medical treatment within two weeks; however, h ealth professionals should monitor women experiencing postpartum blues because they often preclude pos tpartum depression, a more severe psychiatric disord er. One in five women with postpartum blues are likely to develop major depression (Chokka, 2002). Postpartum Psychosis. Postpartum psychosis is the most seve re type of psychiatric mood disorder experienced by women during the perinatal pe riod. It is much less common than postpartum blues occurring in only 0.1-0.2% of all deliv eries (Chokka, 2002; Gitlin & Pasnau, 1989). Although rare in comparis on to other mood disorders, postpartum psychosis is noteworthy because incidence rate s of psychosis in the postpartum period are 12-14.5 times higher than incidence rates of psychosis occurring outside the postpartum

PAGE 42

20 period (Nott, 1982), suggesting that the e xperiences of pregna ncy, childbirth, and motherhood compound risk for this psychiatric disorder. Postpartum psychosis is characte rized by unstable mood, disorganized behavior with confusion, sleep disturbances delusions, and auditory hallucinations (Beck, 1992; Holden, 1991; Ugarriza, 1995). Women that experience delusions or hallucinations frequently focus on suicide, harming the infant, or the infant being demonic (Leopold & Zoschnick, 1996). Child ren of mothers experiencing postpartum psychosis are at risk of severe child a buse and death. As compared with women experiencing non-psychotic postpartum de pression, those experiencing postpartum psychosis that harbor thoughts of harming th eir infants are more likely to act on these thoughts (Attia, Downey, & Oberman, 1997). On set of symptoms usually occurs within the first postpartum month (Attia et al ., 1997; Chokka, 2002; Gitlin & Pasnau, 1989), and aggressive treatment is required includi ng psychiatric hospitalization and medication therapy. Postpartum Depression. More severe than postpartum blues but less severe than postpartum psychosis, postpartum depression is the focus of this di ssertation. As mentioned previously, 10-15% of all mothers experience symptoms of postp artum depression (O'Hara, 1995b; O'Hara & Swain, 1996), with a recurrence rate of 20-30% with subsequent pregnancies (Walther, 1997). Unlike postpartum blues, postpartu m depression does not generally subside without medical treatment within two weeks. Symptoms can persist up to six weeks or longer with an average duration of thr ee months (Gitlin & Pasnau, 1989; Stover & Marnejon, 1995). A study on severe postpartum depression revealed th at if symptoms go

PAGE 43

21 untreated, 25% of women remained depre ssed one year later (Gregoire, Kumar, & Everitt, 1996). Another study found that abou t half of women experiencing depression during the first three months postpartum c ontinued to be depressed at six months postpartum (Whiffen & Gotlib, 1993). Onset of postpartum depressive symptoms can occur within days of delivery or within the first postpartum year, with most episodes beginning between the 6th and 12th postpartum week (Chokka, 2002). The DSM-IV requires symptom onset within the first postpartum month and symptom dur ation of at least two weeks for a clinical diagnosis of postpartum depression (American Psychiatric A ssociation, 1994). Many clinicians argue that this window of time for symptom ons et is too narrow. Table 1 outlines the DSM-IV criteria for a major depressive episode as ad apted from Misri and Kostaras (Misri, 2002).

PAGE 44

22 Table 1 DSM-IV Criteria for a Major Depressive Episode A. Five or more of the following symptoms must be present daily or almost daily for at least 2 consecutive weeks: 1. Depressed mood* 2. Loss of interest or pleasure* 3. Significant increase or decrease in appetite 4. Insomnia or hypersomnia 5. Psychomotor agitation or retardation 6. Fatigue or loss of energy 7. Feelings of wo rthlessness or guilt 8. Diminished concentration 9. Recurrent thoughts of suicide or death B. The symptoms do not meet the crite ria for other psychiatric conditions. C. The symptoms cause significant impairme nt in functioning at work, school, and social activities. D. The symptoms are not caused directly by a substance or general medical condition. E. The symptoms are not caused by bereavement after the loss of a loved one. *At least one of the five symptoms must be #1 or #2. As summarized in the above table, sy mptoms of postpartum depression may include depressed mood, loss of interest in usually pleasurable activities, difficulty concentrating or making decisions, psychomotor agitation or retarda tion, fatigue, changes in appetite or sleep, feelings of worthlessness or guilt, and thoughts of suicide or death. Others common symptoms described by wome n experiencing postpartum depression include excessive anxiety over the child’s h ealth, uncontrollable cryi ng, irritability, lack of interest in the baby, exagge rated highs and lows, and lack of interest in sex (Leopold & Zoschnick, 1996; Spinelli, 1998; Walther, 1997). As previously discussed, although specif ic symptoms of postpartum depression have been described and used to develop diagno stic criteria, it is important to stress that not all women experience the same symptoms nor do symptoms present and persist with

PAGE 45

23 uniform intensity. Given the heterogeneity in women’s experien ces with depression, research exploring the effects of postpartu m depression on child development needs to consider the course and characterist ics of the depressive episode. Predisposing Risk Factors The results of research on the etiolo gy of and risk factors for postpartum depression have been inconsistent and cont radictory. Responsible for much of the variation in findings is lack of consistent operational definitions for postpartum depression. The distinction between postp artum mood disorders including postpartum blues and postpartum depression is often not made. Some studies use self-report screening measures as estim ates of postpartum depression, whereas others use more stringent diagnostic criteria. The followi ng is a summary of th e various influencing factors explored to describe and explain the onset of pos tpartum depression found in the literature. Biological Factors. Available evidence supports a biological etiology of postpartum depression as a result of the dramatic hormonal changes that occur in a woman’s body following childbirth. Research has shown that fo llowing childbirth, blood plasma levels of progesterone and estrogen sharply drop to prepregnancy levels due to the removal of the placenta following delivery. The mechanisms of influence are thought to operate much like mood disturbances occurring during the premenstrual cycle and menopause -time periods during which hormonal fluctuations are high and unpredictable. Supporting a hormonally driven hypothesis of postpartu m depression etiology, Nott et al (Nott, Franklin, Armitage, & Gelder, 1976) found th at women who experienced the greatest

PAGE 46

24 drop in progesterone levels after delivery we re more likely to se lf-report depressive symptomatology within 10 days of deliver y, however the association was weak. Subsequent studies failed to find a statistically significan t relationship between blood plasma progesterone levels and postpartum depression (Heidrich et al., 1994; O'Hara, Schlechte, Lewis, & Varner, 1991); however, one study did find a slight association between the degree to which salivary progester one levels dropped from the antenatal to postnatal period and postpartu m blues (Harris et al., 1994). Research investigating the influence of rapid drops in estrogen following childbirth has also yielded inconsistent findi ngs. Several studies reported no difference in total estradiol levels or changes in estr adiol levels between wo men with and without postpartum depression (Gard, Handley, Pars ons, & Waldron, 1986; Harris et al., 1989; Nott et al., 1976). However, another study did find significantly lower levels of estradiol in postpartum depressed women compared to non-depressed women at 36 weeks gestation and two days postpartum (O'Hara, Schlechte, Lewis et al., 1991). Another study that was double-blinded and placebo-co ntrolled found that estradiol patches significantly reduced postpartum depressive symptoms (Henderson, Gregoire, Kumar, & Studd, 1991), thus meriting the potential a ssociation between es trogen levels and postpartum depression. Other bi ological variables investigat ed as causative factors of postpartum depression have included levels of cortisol, prolactin, thyroid hormones, and beta endorphins, all of whic h have yielded mixed results (Hendrick & Altshuler, 1999; O'Hara, 1997). In summary, there is little c onsensus as to the associations between biological factors and postpartum depression.

PAGE 47

25 Obstetric Factors. Obstetric stressors or complications su ch as cesarean section and delivery by forceps have been linked to postpartum m ood disturbances. However, results are contradictory and conclusions regarding the association between obstetric complications and postpartum depression cannot be draw n. Murray and Cartwright (Murray & Cartwright, 1993) found that delivery by ces arean section or forceps was positively associated with postpartum depression among women with a pr evious history of depression. However, other studies f ound a decreased prevalence of postpartum depression among mothers who experienced obs tetrical complicati ons (Paykel, Emms, Fletcher, & Rassaby, 1980; Pitt, 1968). This may be due to the fact that greater social support and aid was offered to women undergoing traumatic birth experiences. Still other studies found no association between obstetric factors and postpartum depression (O'Hara, Neunaber, & Zekoski, 1984; Whiffen, 1988). Cognitive or Intrapersonal Factors. Cognitive explanations for postpartu m depression focus on personality characteristics and psychological vulnerabilities of the mother as related to the onset of postpartum depression. Role conflict and feelin gs of maternal inadequacy are frequent complaints among women experiencing postp artum depression. Motherhood is a significant transition in one’s life and cogni tive explanations fo r postpartum depression highlight ambivalence toward the maternal role, particularly among women who were employed outside of the home before pregna ncy and are not returning to work. The change in role is not sufficient in providing the mother with the feelings of self-worth and status that she received from her outside employment. This role conflict is thought to be

PAGE 48

26 associated with postpartum depression (Hock & De Meis, 1990). An additional study on the qualitative experiences of women with postpartum depression revealed several areas of role conflict, each focusing on the woman’s desire to be the “perfect mother.” One resounding theme from narratives was the conf lict with balancing breastfeeding with employment outside the home (Mauthner, 1999). In addition, the experience of childbirth is thought to r eawaken unresolved issues related to the mother’s relati onship with her own mother. The transition into motherhood exacerbates preexisting conflicts she experienced as a child related to unmet dependency needs and may cause the woman to internaliz e negative expectations for herself as a mother (Scott, 1992). Such feelings may l ead to performance anxiety and increased feelings of failure and despondency. Evid ence supporting the “ghosts from the nursery” hypothesis has shown that prevalence rate s of postpartum depression are higher among women who report poor relationshi ps with their own mothers and negative early parentchild relationships as compared to mothers who report healthy relationships and positive early parent-child relationships (Bir tchnell, 1988, 1993; Kumar & Robson, 1984). Furthermore, incongruence between fant asized images of motherhood and the experience of motherhood as a reality has been hypothesized as a predictive factor for postpartum depression (Beck, 2002; Scott, 1992) In the U.S. population, pregnancy and childbirth are culturally constr ucted as positive and joyful experiences. In addition, women are expected to have “maternal in stincts” and be fulfilled and innately knowledgeable in the role of mother. Given a situation where ther e is great disconnect between one’s expectations of motherhood and reality, vulne rability to postpartum depression is believed to be increased. Berggren-Clive (Berggren-Clive, 1988) suggests

PAGE 49

27 that the “myths” of motherhood created by our society set exp ectations that are impossible for mothers to attain, placing them at risk for mental health disturbances. The women interviewed in one study described the incongruity between th eir expectations of motherhood and the reality related to labor and delivery, life with their infants, themselves as a mother, their relationships with their partners, s upport from family and friends, life events, and physical changes (Berggren-Clive, 1988) Another study of women’s lived experiences with postpartum depression identified internal conflicts revolving around women’s expecta tions that they could cope with their new infants although the reality was that they needed he lp; and their expectat ions that motherhood would be happy and joyful when in fact they were experiencing depression and unhappiness (Mauthner, 1999). Interpersonal Factors. Interpersonal factors such as marital problems, family interaction, companionship, and social support have also b een implicated as causative f actors in the development of depression in postpartum women. The quality and stability of the relationship between the mother and her spouse has been shown to be a significant predictor of postpartum depression (Gotlib, Whiffen, Wallace, & M ount, 1991; Kumar & Robson, 1984; Whiffen, 1988). Others, however, have found marital di scord to be associated with postpartum depression only in the presence of other stre ssful life events (Paykel et al., 1980). The results of research investigating the re lationship between spousal problems and postpartum depression consiste ntly support the no tion that the relationship prior to pregnancy is as crucial to the mother’s mental health as the relationship following pregnancy. The proposed association between marital discord and postpartum depression

PAGE 50

28 is thought to be related to the availability of soci al support and aid with childcare duties. In addition, mothers who are depressed may be less likely to engage in reciprocally supportive relationships or even recognize positive reinforcement from a spouse’s supportive behavior (Scott, 1992). Mothers’ relationships with family memb ers including the infant have also been suggested as important factor s in postpartum depression wi th evidence supporting infant temperament as a contributing factor in th e onset of depressive symptoms in the postpartum period (Cutrona & Troutman, 1987; Thomas, Chess, & Korn, 1982; Whiffen & Johnson, 1998). Mismatch between parenting style and infant temperament may lead to increased frustration in the parental role and result in feelings of failure. Incessant crying and disturbed sleeping patterns of the in fant may weaken a mother’s resistance to postpartum depression particularly when she is sleep-deprived hers elf. There is support that the stress created by a temperamentally difficult in fant predicts subsequent depression in the postnatal period even after initial levels of depr ession are controlled (Terry, Mayocchi, & Hynes, 1996). Cutrona and Troutman (Cutrona & Troutman, 1987) also found evidence between an infant’s fu ssiness, crying, and othe r difficult behaviors and levels of postpartum depression both di rectly and through the mother’s perceived self-efficacy as a parent. Social support has been identified as an important factor influencing health and wellness, in particular decreasing levels of postpartum depression (Mueller, 1980). Social support may be conceptualized as help with household chores, help with childcare duties or simply listening to the mother and her experiences. Unlike research investigating the influence of other correlates of postpartum de pression such as biological

PAGE 51

29 and obstetric factors, findings from res earch linking social support and postpartum depression are rather consistent Several studies have found an association between lack of perceived social support during pregnancy and postpartum depression (Colllins, Dunkel-Schetter, Lobel, & Scrimshaw, 1993; Cutrona, 1984; Cutrona & Troutman, 1987; Terry et al., 1996); with the support provided by the baby’s father particularly important (Colllins et al., 1993). An additional line of evidence strong ly linking social support and postpartum depression is the reported low rates of postpartum depression found in non-Western countries where levels of social suppor t are much higher (Mauthner, 1995; Wile & Arechiga, 1999). In fact, many scholars beli eve postpartum depression to be a cultural artifact of Western nations. In a wide range of cultures including China, the Philippines, Kenya, Nigeria, Mexico, Guatemala and Jamai ca, the amount of emotional and practical support extended to pregnant women and young mothers is much greater than that received in the United States. The support is provided throughout pregnancy and the postnatal period, and in some cultures even throughout the child’s life. Female friends and relatives including the wo man’s mother and her mother-in-law provide the emotional and physical support. In the United States, not only are values such as privacy and independence highly valued cont ributing to lower levels of social support received during the perinatal period, but also extended familie s are often geographically unavailable to mothers. One study investigating cultura l variations in prevalence of postpartum depression related to social support between the U.S. and Japan found that mothers who were depressed received less support from their families during pregnancy and were more

PAGE 52

30 likely to be responsible for housekeeping and household management tasks compared to non-depressed mothers (Murata, Nodaoka, Morioka, Oiji, & Saito, 1998). The tradition in Japan is for women to live in the home of her husband’s family after delivery where she receives help caring for the new infant, a nd is not expected to assist with household chores such as cooking and cleaning. Desp ite the trend of traditional Japanese living arrangements falling out of fa vor, 51% of the study sample lived in households composed of three generations. Although the results of the study indicate d a lower rate of postpartum depression in Japan than in the Un ited States (consistent with the cultural variation trend), the researcher s found a higher rate (7.2%) than what ha d been previously published (3.1%) in Japan. Although the differe nce may simply result from the use of variable methodologies and assessment tools, the difference may also highlight a shift away from the more traditional three-generation living arrangements to more Western living arrangements characterized by isolati on from extended family and lower levels of familial support. Structural Factors. Macro-level or structural f actors such as financial stre ss, housing, availability of food and shelter, access to health care, and co mmunity violence have been identified as important factors related to poor mental he alth in general, and in relationship to depression during the postpartum period. Negati ve life events and chronic stressors such as unemployment and housing difficulties ha ve been found to increase the likelihood of postpartum depression (Brown & Harris, 1978). Structural factor s are particularly influential if they interfere with the mother’s ability to care and provide for her infant, as

PAGE 53

31 is the case with severe financial difficulty, lack of food and shelter, or abandonment by the baby’s father (O'Hara, 1997). One study investigating the relationshi p between chronic life stressors and postnatal depression among lo w-income women from Montreal found that maternal health problems, housing difficulties, lack of money for basic needs, and frequent conflicts with network members were asso ciated with postpartum depression at six months (Seguin, Potvin, St-Denis, & Loiselle 1999b). Measures of chronic stress were also assessed at 30 weeks ge station, three weeks postpartu m and nine weeks postpartum to allow the researchers to investigate ch anges in exposure to stressors over time. Longitudinal analysis revealed that more wo men expressed stress related to lack of money for basic needs at six months than previo us data collection point s. Further, of the 38.2% of women identified as experiencing pos tpartum depression at six months, more than half of the depressed women were ne w cases. The finding that nearly half of depressed cases at six months postpartum we re new cases, and a higher percentage of women reported stress related to financial strain at this time point suggest that structural factors including lack of money ma y impact later onset depression. A simultaneous study by the same res earchers compared low socioeconomic status women with a group of higher socio economic status women (Seguin, Potvin, StDenis, & Loiselle, 1999a). Findings revealed differential duration and severity of depressive symptoms. Whereas prevalence of depression for high socioeconomic status women decreased from three weeks postpar tum to nine weeks postpartum (22.2% 8.3%), rates for low socioeconomic status wo men decreased only slightly between the third and sixth week postpartum (23.7% 21.3%). Furthermore, among high

PAGE 54

32 socioeconomic status women, no women were cl assified as being se verely depressed at nine weeks postpartum, wh ereas 1.3% of low socioeconomic status women were classified as severely depressed. Lastl y, only one new case (2.8%) of depression was detected at nine weeks postpartum for women of higher socioeconomic status as compared to ten (12.5%) new cases among women of lower socioeconomic status. Thus it appears that the experien ces of postpartum depression, particularly duration of symptomatology, differ for women of hi gh and low socioeconomic status. Personal Psychopathology. The most well accepted correlate of postpartum depression is personal psychopathology or depression prior to pregna ncy and/or during pregnancy. Research findings have consistently found that women who have e xperienced prior psychiatric mood disturbances are at increased risk for postpartum depression (Campbell, Cohn, Flanagan, Popper, & Meyers, 1992; Wats on, Elliott, Rugg, & Brough, 1984). It is estimated that 60% of the women who deve lop postpartum depression have a positive psychiatric history (Paykel et al., 1980; Watson et al., 1984; Whiffen & Gotlib, 1993). Summary. In summary, a host of factors have been implicated in research studies as predisposing risk factors for the onset of pos tpartum depression, with varying levels of empirical support. Despite contradictory evid ence, it is largely accepted by researchers in the field that biological and obs tetric factors contribute little understanding to the onset of depression during the postnatal period. There appears to be moderate evidence for an association between intrapers onal factors includi ng role conflict with the woman’s new role as mother, unresolved issues with her own mother when she was a child, and

PAGE 55

33 incongruent expectations of motherhood with the reality of caring for a new child, and postpartum depression. Interpersonal factors have also been sugge sted as predisposing risk factors, with marital discord having a particul arly strong impact. Available social support to women is another well-accepted determinant of th e onset of postpartum depression, with instrumental support being mo re important than emotional or information support. Cultural variations in the prevalence of postpartum depression also maintain the importance of social support in the deve lopment of postpartum depression, where nonWestern cultures that provide extensive suppor t to the new mother have relatively low levels of reported depression during the pos tpartum period as compared to the U.S. Other factors that have moderate ev idence for an association with the development of postpartum depression are m acro-level factors including unemployment, financial strain, and housing di fficulties. These factors appear to be most related to depressive episodes that begin later in the postpartum period, and have longer durations. Lastly, the most well accepted predictor of pos tpartum depression is a previous history of psychiatric illness includi ng previous episodes of postp artum depression with prior pregnancies. In investigating the impact of postp artum depression on child developmental outcomes, it is crucial to include as many of the identified correlates of postpartum depression found in the l iterature as possible, to control for their potential impact on child development. For example, many of predispos ing risk factors for postpartum depression including marital discord, low levels of per ceived social support, and financial strain, may also have a direct effect on child development.

PAGE 56

34 Attachment Theory Introduction Attachment is described as the deep and enduring connection or emotional bond established between infant and caregiver during the first year s of the child’s life (Klaus, Kennell, & Klaus, 1995; Levy & Orlans, 1998). The mother is most often the first attachment figure with whom an emotiona l bond develops, however, infants and young children can establish attachment relationshi ps with alternative caregivers who have a regular presence in the child’s life. Accordi ng to attachment theory, infants are social from birth and pre-adapted to interact with and respond to a prim ary caregiver (Bowlby, 1969). Infants display attachment behaviors su ch as crying or clinging to seek proximity to the mother, which from an evolutionary standpoint promotes species survival. Attachment theory also purports that adult ca regivers are equipped w ith a repertoire of care taking behaviors to complement infant attachment behaviors (Bowlby, 1969). The mutual responsiveness and interaction between child and caregiver is believed to be critical for the formation of h ealthy attachment relationships. A mother that responds in a consistent, appropriate, and prompt manner to her infant’s cues and behaviors will foster the development of a positive emotional bond between herself and the child. Historical Context John Bowlby, a child psychiatrist and founder of attachment theory, is considered one of the most important psychiatrists of the twentieth century (S torr, 1992). Although trained in psychodynamic theory, a perspective that believes behavior is motivated by inner forces, memories, and conflicts of wh ich a person has little awareness or control (Feldman, 1999), Bowlby was dissatisfied w ith many aspects of the psychoanalytic

PAGE 57

35 community particularly the neglect of the role of real environmental trauma in the genesis of normal development (Holmes, 1995). Thr ough his work and research experiences, Bowlby was certain that environmental events such as real trauma and loss played a critical role in personality development. Bowlby concluded th at significant early deprivation resulted in a va riety of severe developmental problems including lack of empathy, display of conduct disorders, inab ility to give and receive affection, and attention deficits (Bowlby, 1951). He bega n to formulate his own ideas converging several disciplines within the biological a nd social sciences in cluding psychoanalysis, ethology, and control systems theory to de velop a new paradigm for developmental psychology (Ainsworth, Blehar, Waters, & Wall, 1978c). During his exploration of alternative models of human behavior, Bowlby was influenced by ethology, the zoological study of animal behavior concerned with the adaptive or survival value of behaviors with in an evolutionary history (Hinde, 1989). Ethology focuses on genetically programmed behaviors known as instincts, which are inherited by animals through their parents and open to natural selec tion and modification. According to ethology, animal infants are biolog ically prepared to actively contribute to the establishment of a bond with their careg iver in order to promote the chance for individual genes to survive (Hinde, 1989) After reading research on imprinting behaviors of ducklings and gos lings that noted the early fo llowing behaviors of ducklings and baby geese to promote survival given th at the behavior ensured close proximity, Bowlby supported the notion that human infants were like animal infants, equipped with a set of innate behaviors to promote proxim ity to the caregiver in order to increase chances of survival.

PAGE 58

36 Theoretical Formulations Attachment theory was introduced to the psychoanalytic community through a series of papers presented to the British Psychoanalytic Society and later published entitled “The Nature of the Child’s Tie to His Mother” (Bowlby, 1958), “Separation Anxiety” (Bowlby, 1959), and “Grief and Mourning in Infancy and Early Childhood” (Bowlby, 1960). These papers were the firs t formulations of the major tenets of attachment theory and were later expanded into entire volumes “Attachment” (Bowlby, 1969), “Separation” (Bowlby, 1973), and “L oss” (Bowlby, 1980), the three volumes collectively known as “The Trilogy.” In the first paper presented in 1957, “The Nature of the Child’s Tie to His Mother,” Bowlby rejected the well-accepted ps ychoanalytic notion that the mother-infant relationship was rooted in the infant’s n eed for and the mother’s provision of food (Lamb, Thompson, Gardner, & Charnov, 1985a). Drawing from Lorenz’s research that noted instinctual bondin g without feeding among newly ha tched goslings (Lorenz, 1952), Bowlby supported the notion of an innate pr opensity for contact with the mother independent of the infant’s need for food (Holmes, 1993; Meins, 1997). Bowlby suggested that infants display instinctual responses such as sucking, clinging, following, gazing, smiling and crying, to maintain close contact with the mother. The second paper “Separation Anxiet y” presented in 1959 highlighted the failure of traditional psychological theories including psychoanalysis to explain the dramatic responses of infants and young childre n when separated from the mother figure (Bowlby, 1988). Contemporary theories proposed that fear was arous ed in individuals only in situations perceived as painful or dangerous. Given separation from the mother

PAGE 59

37 figure was not considered to be a painful or dangerous situation, theorists posited that some other situation was really the cause of the fear res ponse. Bowlby however, taking an ethological approach, proposed that man like other animals responds with fear to certain situations not because they carry a high risk of pain or danger, but because they signal an increase in risk (Bowlby, 1988). Bowlby explains that infants and young children respond to separation fr om a caregiver much in the same way as man responds to sudden movement or a marked change in level of sound or light (Bowlby, 1988). From an ethological viewpoint, display of such behaviors has survival value. Working with Robertson (Roberts on & Bowlby, 1952), the two researchers generalized the sequence of responses s een in young children during a temporary separation from the mother figure. The seque nce included protest (related to separation anxiety), despair (related to grief and mourning), and denial or detachment (related to defense mechanisms). Bowlby maintained that infants and children experience separation anxiety when the instin ct to be near the mother is activated, but the attachment figure is not available. Bowlby asserted th at excessive separati on anxiety was due to adverse family experiences such as repeated threats of abandonmen t or rejection by the parents. He also noted that se paration anxiety could be low or absent altogether if the child has achieved a level of maturity or pseudoindependence as a defense mechanism to repeated separations (Bretherton, 1995). The third and most controversial pape r “Grief and Mourning in Infancy and Early Childhood” was presented to the Psyc hoanalytic Society in 1959 and published in 1960. In this paper, Bowlby suggested that separated and bereaved infants and small children experience grief and mourning no less inte nsely than healthy adults. He asserted

PAGE 60

38 that grief and mourning develop when the ins tinct desiring proximity to the mother is activated but unable to be term inated given the unavailabil ity of the mother. It is proximity to the mother, not food th at drives the behavioral pattern. Modern Attachment Theory Attachment, Attachment Behavior, a nd the Attachment Behavioral System. To understand attachment theory, it is fi rst critical to dist inguish between the interrelated concepts of attachment, attachme nt behavior, and the attachment behavioral system. Attachment is the psychodynamic component of attachment theory and represents the deep and enduring connecti on or emotional bond established between infant and caregiver during the first years of the child’s life (Levy & Orlans, 1998). Attachment refers to the state and quality of the bond between mother and child that is formed over time through infant display of and mo ther response to attachment behaviors. Bowlby believed that a specific unbroken bond to a particular person was essential for nonpathological psychological and social development (Bowlby, 1969; Levy & Orlans, 1998). Attachment behaviors are the behavioral patterns displayed by infants and young children that mediate the form ation of an attachment bond be tween child and caregiver. Examples of attachment behaviors incl ude crying, smiling, vocalizing, clinging, and following. Probably the most fundamental aspect of attachment theory is its focus on the biological basis of infant attachment behavior s to achieve close proximity to the mother (Cassidy, 1999). From an evolutionary standpo int, Bowlby asserted that genetic selection favors innate infant attachment behaviors be cause they increase the likelihood of close mother-infant proximity, which in turn provides protection fro m danger and thus

PAGE 61

39 promotes survival. Because the functional efficacy of infant attachment behaviors depends on the promptness and appropriatene ss of the caregiver’s response, mutual responsiveness and interaction become critical in the formation of the attachment relationship between mother a nd child (Lamb et al., 1985a). Just as attachment behaviors mediate the formation of the attachment relationship between mother and child, Bowlby argued that attachment behaviors are mediated by a behavioral system (Bowlby, 1969). An ethological concept describing species-specific behaviors that lead to pred ictable outcomes or “set-goa ls” (Cassidy, 1999), behavioral systems serve a particular biological functi on to promote survival. For example, the attachment behavioral system functions to ma intain a sufficient level of proximity to the attachment figure for protection. To e xplain the dynamics operating within the behavioral system, Bowlby drew upon control sy stems theory to assert that the system operates much like a thermostat continua lly evaluating both exogenous and endogenous factors to maintain its “set goal” (Ain sworth, 1964b; Lamb et al., 1985a; Marrone, 1998a). For example, Bowlby suggested that the degree of proximity sought by an infant during activation of the atta chment behavioral system would depend upon internal conditions such as infant illness, fatigue, a nd hunger, as well as external environmental conditions such as novelty of physical surr ounding, presence of threatening stimuli, and absence of mother figure. When the set-goal is set widely due to a level of comfort, a child may venture a substantial distance be fore the set-goal is exceeded and the attachment behavior activate d. The child’s mother may need to only smile or nod in reassurance to terminate the ch ild’s attachment behavioral system. Similarly, when the attachment behavioral system is activated to a high degree of intens ity, the set-goal may

PAGE 62

40 be close bodily contact and the attachment be havior will not cease until this set-goal has been achieved (Ainsworth et al., 1978c). In summary, the attachment behavioral system dictates the display of infant attachment behaviors, which in turn medi ate the development of an affectional bond between mother and child. Although an abstract psychodynamic construct, the attachment relationship has been described as having three key f eatures including: (1) infant proximity seeking to a preferred figure; (2) the secure base effect or using the caregiver as a springboard for curiosity and exploration; and (3) separation protest (Weiss, 1982). Development of Attachment. Bowlby outlined four phases in which the attachment relationship between child and primary caregiver develops over time; howev er, he was careful to note that there are no sharp boundaries between phases given in dividual heterogeneity (Bowlby, 1969). The first phase “Orientation and Si gnals without Discrimination of Figure” (from birth to four to six weeks) (Lamb et al., 1985a) is ch aracterized by infant s emitting attachment behaviors such as tracking movement with eyes, grasping, reaching, smiling, or babbling with other persons present in the environmen t without the ability to discriminate one individual from another. The infant appears content as long as someone satisfies his or her needs and responds to his or her signals. This phase is also know n as the “Initial Preattachment Phase” (Ainsworth et al., 1978c). As infant visual and auditory senses ma ture, the infant begins to show a clear preference for a small number of familiar individuals including parents and regular caretakers, particularly the mother. This phase, “Orientation and Signals Directed

PAGE 63

41 Towards One or More Discriminated Figures” (t o about six or eight months of age), is typified by infants who soothe more quick ly and respond more positively with some adults compared to others (Lamb et al., 1985a). Infants are also able to discriminate between one familiar attachment figure and another exhibiting different proximitypromoting behaviors (Ainsworth et al., 1978c). Characteristic of th is phase are infants who listen for and respond differently to thei r mothers’ voices and cry differently when their mothers depart, as compared to othe r individuals. In a study of infant-mother interaction among dyads from Uganda, Africa, Ainsworth observed that Ganda infants, when held by someone else, kept orienting th emselves toward their mothers (Ainsworth, 1964a). Bowlby asserted that although infant s will continue to accept the attention and nurturance from less preferred individuals or even strangers, these ch ildren clearly prefer interaction with a small number of persons in particular the mother figure. This phase is also referred to as the “Phase of Attachme nt-in-the-Making” (Ain sworth et al., 1978c). The third phase “Maintenance of Proximity to a Discriminated Figure by Means of Locomotion as Well as Signals” (six months to three years) is represented by increased infant discrimination for inter action with a preferre d caregiver, as well as display of an extended repertoire of responses to include following a departing mother, greeting her upon return, and using the mother as a base from which to explore (Bowlby, 1969). Bowlby declares that the frie ndly and undiscriminating accepta nce of attention from nonpreferred individuals present in phase two no longer exists as infants become wary of strangers and treat them with increased caution. According to Bowlby, it is not until phase three that true emotional attachment between infant and caregiver emerges given it marks the onset of the infant’s behavior

PAGE 64

42 becoming organized on a goal-corrected basi s (Bowlby, 1969). The display of specific attachment behaviors is no longe r critical, but instead it is the set-goal and the child’s plan for achieving it that becomes significant. Bowlby highlights that in measuring the affectional bond between mother and child, it is the charac teristic way in which a child learns to organize his or her behavior in reference to an attachment figure that is of importance, not the intensity or frequency of display of attachment behaviors (Ainsworth et al., 1978c; Bowlby, 1969). Phase three also marks the beginning of in fant protest when infants are separated from their attachment figures (Bowlby, 1973). Separation protest is th e child’s attempt to summon the absent attachment figure back to reestablish close proximity. It requires infant comprehension that th e attachment figure continues to exist when not visible, audible, or palpable (Lamb et al., 1985a), a cognitive capacity known as object permanence. Separation protest is the concep tual basis of Mary Ainsworth’s empirical method of assessing attachment bonds between mother and child known as the Strange Situation. Phase three of attachment de velopment coincides with child locomotion marking the beginning of a more active and i ndependent role of th e infant within the attachment behavioral system. Through loco motion, the child can now seek to maintain proximity to the caregiver inde pendent of the attachment figu re’s responsiveness to his or her behavioral cues (Holmes, 1993). Phase th ree is also known as the “Phase of Clearcut Attachment” (Ainsworth et al., 1978c). The last phase of attachment develo pment “Formation of a Goal-Corrected Partnership” occurs sometime during the third or fourth year of the child’s life when the child’s egocentricity decreases and his or he r ability to infer the feelings, motives and

PAGE 65

43 plans influencing the behavior of others incr eases. At this time, children acquire the capacity to recognize that othe rs, including their attachment figures, have desires and needs of their own which are to be consider ed during patterns of interaction (Marvin, 1977). As a child’s cognition develops and becomes more sophisticated through maturity, the child begins to understand th e mother’s set-goals and her plans for achieving them and alters his or her behavi or accordingly. The ch ild-mother relationship develops into a partnership. Maternal Sensitivity and Internal Working Models. Infant development of an attachment re lationship to a primary caretaker is a normative developmental process. Almost a ll infants develop an affectional bond with a primary caretaker, even in families where abuse or neglect is present, and attempt to use that caretaker as a source of comfort and reassurance when faced with challenges or threats from the environment (Bowlby, 1956; Bretherton, 1993; Weinfield, Sroufe, Egeland, & Carlson, 1999). Nonethel ess, the distinction between the presence and quality of an attachment relationshi p is critical, as not all at tachment relationships are equal in terms of the comfort a nd reassurance that they afford. Bowlby asserted that variation in atta chment relationships resulted not from individual traits of th e infant or primary caregiver, such as temperament, but instead from dyad-specific transactional patterns (Bowl by, 1969; Weinfield et al., 1999). He argued that because infant behavior is adapted to complement a caretaker’s contingent and appropriate response to infant cues, an in fant faced with a caretaker who responds inconsistently or inappropriately learns th at his or her proximity -promoting attachment behaviors are functionally ineffective (Lamb et al., 1985a). Bowlby suggested that an

PAGE 66

44 infant who lacks confidence in the mother’s responsiveness might have a more narrowly defined set-goal for proximity as well as a te ndency to perceive situational events more negatively, as compared to an infant with a more responsive caretaker (Bowlby, 1969). The child’s appraisal processes underlying the attachment behavioral system are affected, greatly impacting the quality of the attachment formed. Bowlby went further to assert the impo rtance of caregiver responsiveness by suggesting that infants construc t internal working m odels, or internali zed representations of the caretaker’s accessibility and respons iveness, based on their history of prior interactions with the caregiver (Bowlby, 1973, 1980). The function of internal working models is to forecast and inte rpret the behavior of the careg iver, as well as to plan the child’s behavior in response to the caregiver (Bre therton, 1993). Internal working models are organized around basic questions of trust, security, and well being such as: “Will my signals of need and distress be understood a nd responded to? What kind of responses can I anticipate? Will I feel better or worse as a result of the response I get? Which of the many signals I give get attention, and is it th e kind of attention I want?” (Lieberman & Zeanah, 1995). Children of responsive mother s expect to be fed when hungry, comforted when distressed, and played with when aler t and ready for interaction (Gowen & Nebrig, 1997). They learn that they are worthy, expect others to respond to them in a positive manner, and engage in behaviors that elicit further care and acceptance (Crockenberg & Leerkes, 2003). Infants of unresponsive mothers whose needs often go unmet or experience rough or intrusiveness treatment, be gin to expect caregivers to be hostile, controlling, unresponsive and undependable (Gowen & Nebrig, 1997); and view

PAGE 67

45 themselves to be unworthy, unloved, and expect others to treat them as such and behave in ways that elicit further rejecti on (Crockenberg & Leerkes, 2003). Bowlby felt that once established, internal working models were fairly resistant to change and guided all future in teraction with the car egiver (Lamb et al., 1985a). For this reason, an infant who has become reluctant to seek or accept comfort from a caregiver who has rejected or inconsis tently responded to the child is not likely to respond when the previously unresponsive caregiver suddenly become s available (Bowlby, 1973; Bretherton, 1993). The internal working mode l not only represents the template of interaction between mother and child, but also provides the child with notions of self in relation to others guiding fu ture relationship development and interaction patterns with individuals in the social envir onment (Bowlby, 1973; Thompson, 1999). Although Bowlby emphasized the theore tical significance of caretaker responsiveness in the formati on of an attachment bond between mother and child, it was Mary Ainsworth who provided the first empirical evidence of its importance. Ainsworth applied the term “maternal sensitivity” to re present what Bowlby referred to as caretaker responsiveness and defined a sensitive mother as one who noticed their infant’s signals, interpreted them accurately by taking the infa nt’s perspective, and responded reasonably promptly and appropriately to those cues (Ainsworth, Bell, & Stayton, 1974). She was careful to note that although an insensitive mother may display unpleasant or mean behaviors toward the infant, insensitivity rath er implies that the caregiver is not reading and supportively responding to the infant’s states or goals (Ainsworth et al., 1974; Bretherton, 1993).

PAGE 68

46 In her naturalistic observa tional studies of mother-i nfant dyads from Uganda (Ainsworth, 1963) and Baltimore (Ainsw orth, Blehar, Waters, & Wall, 1978a), Ainsworth noted striking indi vidual differences in the se nsitivity in which mothers responded to their infants’ signals, as related to attachment style. In Uganda, Ainsworth noted that mothers who were classified as se nsitive were more likely to have children who were securely attached; as compared to mothers who appeared indifferent or imperceptive to their infants’ behaviors and n eeds, classified as insensitive, who were more likely to have children who were ins ecurely attached (Ainsw orth, 1963; Bretherton, 1995). In Baltimore, Ainsworth observed distinct differences in maternal sensitivity to infant cues during face-to-face interactions, feeding, close bodily contact, and crying episodes, which were associated with more harmonious mother-infant relationships at 54 weeks. Sensitive mothers were described as adjusting and pacing their own behaviors in response to their infants, a display of effec tive functioning of the attachment behavioral system. For example during face-to-face inte ractions, when mothers meshed their own playful behavior with that of their infants, the infants responded with joyful bouncing, smiling and vocalizing; when mothers however in itiated face-to-face in teractions silently or with an unsmiling expression, the interacti on was often muted and brief (Ainsworth et al., 1978a; Blehar, Lieberman, & Ainsworth, 1977). Similar results were found for observations of feeding, close bodily contact, and crying ep isodes. Results from the study revealed that maternal sensitivity during the firs t quarter of observation was correlated with positive mother-child interact ion during the fourth qua rter, as well as to infant behavior in a laboratory procedure known as the Strange S ituation (Ainsworth &

PAGE 69

47 Wittig, 1969), designed to examine the balance of attachment and exploratory behaviors under conditions of low and high stress in or der to classify a child’s attachment relationship with a primary caregiver, the ea rliest formation and organization of the child’s internal working mode l (Marrone, 1998b). Ainsworth concluded that infants of sensitive mothers were more likely to be s ecurely attached accord ing to her developing classification system of attachment styles, and concluded a direct correlation between maternal sensitivity and attachment quality. Summary. In summary, the development of an attachment relationship or emotional bond between mother and child is mediated by th e child’s display and mother’s response to attachment behaviors such as clinging and cr ying, which serve a biol ogical need for the infant to achieve protective proximity to the mother. Infant display of attachment behaviors is dictated by the attachment behavioral system which operates much like a thermostat taking into consideration internal and external factors to help the infant achieve a set level of proximity to the car egiver. Although nearly all infants become attached to a primary caregiver, not all attach ment relationships are equal with maternal sensitivity acting as the moderating influence. Sensitive mothers notice infant signals, interpret signals accurately by taking the infa nt’s perspective, and respond promptly and appropriately to these signals, promoting th e development of a healthy mother-infant attachment relationship. These early caregivi ng experiences, whether with a sensitive or insensitive mother, forms the basis of an inte rnal working model of self and others that serves as a template for future social inte ractions and guides the child in meeting later social and emotional demands in life. Infant s with a secure internal working model of

PAGE 70

48 relationships seek and begin to expect suppor tive and satisfying en counters with old and new partners; just as children with insecure internal working models may anticipate less support and actually deter suppor t from others due to distrust and uncertainty in relationships (Thompson, 1999). Effects of Postpartum Depression on Child Development Introduction Postpartum depression has serious impli cations for the mother’s ability to function effectively in her role as mother and caregiver and pos es a threat to the development of her child. Mothers of young in fants, particularly primiparous mothers, must adjust to their baby and learn to unders tand their infant’s cues and needs (Campbell & Cohn, 1997). According to att achment theory, mothers need to respond to their infants in both a sensitive and consistent manner to promote optimal psychological health. This task may be difficult for mothers experien cing postpartum depression who may feel overwhelmed by the responsibilities requi red during the transition to motherhood (Belsky, Rovine, & Fish, 1989; Klaus et al., 1995). Substantial evidence regarding the asso ciations between postpartum depression and adverse outcomes in children has accumulated during the past two decades. Developmental outcomes examined have incl uded a wide range of measures including social competence (Carter, Garrity-Rokous Chazan-Cohen, Little, & Briggs-Gowan, 2001; Hay & Pawlby, 2003; Luoma et al., 2001; Murray, Sinclair, Cooper, Ducournau, & Turner, 1999; NICHD Early Child Care Research Network, 1999b); behavioral difficulties (Brennan, Hammen, Andersen, & Bo r, 2000; Caplan, Cogill, Alexandra,

PAGE 71

49 Robson, & Katz, 1989; Carter et al., 2001; Civic & Holt, 2000; Dawson et al., 2003; Hubbs-Tait et al., 1996; Lyons-Ruth, Easterbrook s, & Cibelli, 1997; Murray et al., 1999; NICHD Early Child Care Research Networ k, 1999b; Phillipps & O'Hara, 1991; Shaw & Vondra, 1995); mother-child interaction and at tachment relationships (Beck, 1996; Carter et al., 2001; Donovan, Leavitt, & Walsh, 1998; Edhborg, Lundh, Seimyr, & Widstrom, 2001; Frankel & Harmon, 1996; Hipwell, Goossens, Melhuish, & Kumar, 2000; Jameson, Gelfand, Kulcsar, & Teti, 1997; Karl, 1995; Livingood & Daen, 1983; Murray, 1992; Murray, Fiori-Cowley, & Hooper, 1996; NICHD Early Child Care Research Network, 1999b; Popper, Ross, Cohn, & Campbell, 1993; Teti, Gelfand, Messinger, & Isabella, 1995); cognitive functioning (Kurstje ns & Wolke, 2001; Murray, 1992; Murray, Fiori-Cowley et al., 1996; Murray, Hipwell, & Hooper, 1996; NICHD Early Child Care Research Network, 1999b; Petterson & Burk e Albers, 2001); and child psychopathology (Essex MJ, Klein MH, Miech R, & NA, 2001; Hammen & Brennan, 2003). Given that the focus of this study is on child social co mpetence and display of problem behaviors, particularly social competence and display of problem behaviors at first grade, the literature exploring these two outcomes will be reviewed in greater detail. Empirical evidence from studies inve stigating the impact of postpartum depression on child social competence and beha vioral difficulties (as well as other child developmental outcomes) is quite consistent. As a result, the focus of researchers is no longer on whether or not depression has an e ffect on the child, but instead understanding the characteristics of depression that pose the greatest risk, and understanding factors that mediate and moderate the association.

PAGE 72

50 Much of the literature in this area has neglected to fully consider the wide variation in women’s experien ces with postpartum depression and how the nature and course of depressive episodes may differe ntially impact the likelihood of negative outcomes. For example, depression may va ry in severity from mild with little impairment to severe with more dist urbed functioning (Hammen & Brennan, 2003). Chronicity of symptomatology is another char acteristic of depressi on that likely impacts the degree to which negative child outcomes surface, with infants exposed to longer durations of depressive ill ness in the mother having the worst outcomes (Campbell & Cohn, 1997). Stein and colleagues (Stein, Lu, & Gelbery, 2000) have applied the concept of “weathering” in their study documenting a relationship between chronic homelessness and adverse birth outcomes in women to represent the physic al consequences of chronic homelessness on health. Similarly, depres sion chronicity in mothers may pose a “weathering” effect on child social developm ent. Timing of initial onset during the postpartum period is another cri tical attribute of maternal depression that warrants more investigation. Developmental psychology ha s long considered the importance of sensitive periods in a child’s developmenta l trajectory (Hay, 1997) ; however, literature investigating relationships be tween postpartum depression and child outcomes has rarely considered the timing of initial onset relative to the infant’s age and specific developmental needs at that period. Understanding which aspects of postpartum depression pose the greatest risk to child development is crucial at this time not only to advance the cu rrent knowledge base in the field, but also to fully develop the pr evention implications of early detection and treatment of postpartum depressi on. Public and mental health professionals need to know

PAGE 73

51 if all women experiencing postpartum depre ssion need to be ta rgeted, or only those experiencing severe, longer duration, or early onset episodes. Elucidating which contextual risk factors in crease the likelihood of nega tive outcomes, as well as understanding what factors including maternal sensitivity and exposure to non-maternal care mediate and moderate the association, is also important for future development of public health interventions. The following sections will summarize the empirical evidence investigating the impact of postpartum depression on child soci al competence and beha vioral difficulties. Close attention will be given to the characte ristics of depression considered across the studies including timing of onset severity, and chronicity, if considered at all. Although a substantial number of studies have included one or more of these finite aspects of depression in their investigations, operat ionalization has not b een consistent nor treatment during analysis sophisticated, and thus will be critically anal yzed in the review. Further, given that the studi es vary in terms of design, sample population, timing and measurement of depression, timing and meas urement of child outcome, inclusion of contextual risk factors, and inclusion of mediating and mode rating influences, attention to these differences will be highlighted where n ecessary. The primary goal of this review is not to compare and contrast results acro ss studies, although importa nt, but instead to highlight the need for further research investigating the effects of timing of initial onset, severity, and chronicity of postpartum depr ession on child developmental outcomes; and the need to more thoroughly test the influences of medi ating and moderating factors among a large sample of mother and child dyads participating in a multi-site prospective longitudinal study.

PAGE 74

52 Social Competence Social development is defined as the way in which individuals’ interactions with others and their social relati onships grow, change, and remain stable over the course of life (Feldman, 1999). During the preschool and early elementary periods, socially competent children begin to see their peer s as individuals with whom they form friendships based on trust and shared intere sts. Play becomes more constructive and cooperative, and social skills are import ant for the maintenance and growth of relationships. The terms social compet ence and social skills are often used interchangeably, but social competence is actu ally the broader term that includes social skills as well as other behavioral, cognitive, and emotional adjustment characteristics, which are necessary to develop and maintain adequate social relationships (Merrell, 1999c). Social competence is a complex construct th at is largely accepted to represent five dimensions including peer relationships, self -management, academic skills, compliance, and assertion (Caldarella & Me rrell, 1997). Peer relationships is dominated by social skills reflecting children who are positive with their peers engaging in behaviors such as complimenting, praising others, offering help, and inviting others to play or interact (Merrell, 1999c). Self-management is of ten used to describe children who are emotionally well-adjusted given they are able to control their temper, follow rules and limits, compromise with others, and receive criticism well. Academic skills are characteristic of youth who are considered to be independent and productive workers given they tend to accomplish tasks independe ntly and carry out teacher directions. Compliance describes children who get al ong with others by following rules and

PAGE 75

53 expectations, appropriately use free time, and share. Last, as sertion reflects children who are outgoing or extroverted; they initiate conversations with others, acknowledge compliments, and invite others to interact in social situations. Social competence is important to unde rstand and promote healthy development of given literature in the fiel ds of child development, e ducation, and psychology support long-term ramifications of ade quate social skills and peer relationships during childhood. It has been established that social competence lays the foundation for interpersonal interactions later in life (Asher & Taylor, 1981); and has a significant impact on academic success during the school years (Walker & Hops, 1976) including school readiness (Carlton & Winsler, 1999) and pl acement into special education classes (Resnick et al., 1998; Resnick et al., 1999). Socially compet ent kindergartners have been shown to be more successful in developing positive attitudes about and adjusting to school, as well as earning better grades and achieving more, as compared to less competent children (Birch & Ladd, 1997; Ladd, Birch, & Buhs, 1999). Conversely, deficits in social skills and p eer relationships have been show n to be strong predictors of mental health problems later in life (C owen, Pederson, Babigan, Izzo, & Trost, 1973; Roff, 1963); as well as school difficulties, deli nquency, drug abuse, and school drop out rates (Gagnon, Craig, Tremblay, Zhou, & V itaro, 1995; Haapasalo & Tremblay, 1994; Roff, Sells, & Golden, 1972; Roff & Sells 1968; Tremblay, Ragani-Kurtz, Masse, Vitaro, & Pihl, 1995). Although there is abundant re search investigating th e impact of postpartum depression on social and emotional devel opment of children, ther e is a paucity of literature that focuses specifical ly on social competence. Much of the research in the area

PAGE 76

54 of child social and emotiona l development explores the im pact of postpartum depression on the mother-child interpersonal relationshi p including attachment security. Although a critical aspect of social and emotional development, this study focuses on social competence rather than attach ment security in order to investigate the impact of depression on child development at first grad e, to capitalize on available longitudinal data. Further, there is substantial evid ence of a strong associa tion between postpartum depression and mother-infant in teractions includi ng the attachment relationship, and very little research investigating the influence on social competence and only two that take into consideration the course of de pression on child social competence. Chronicity. One longitudinal study has investigated the impact of postpartum depression chronicity and child social competence am ong a large (n=1,215) multi-site sample of mother and infant dyads (NICHD Early Ch ild Care Research Network, 1999b). Child social competence at 36 months was meas ured via the Adaptive Social Behavior Inventory (ASBI) (Hogan, Scott, & Bauer, 1992), which yields scale scores for expression (sociability and empathy) a nd compliance (prosocial engagement and cooperation). The expression and compliance sc ales were summed together to generate a single social competence variable. Postpart um depression was assessed via the Center for Epidemiological Studies Depression Scal e (CES-D) (Radloff, 1977) at one, six, 15, 24, and 36 months. Despite longitudinal data on postpartum depressi on, chronicity was operationalized as a single vari able with women classified as (a) never depressed if they scored below threshold valu es on the CES-D at every assessment; (b) sometimes depressed if they scored above threshold valu es at least once, but fewer than four times;

PAGE 77

55 and (c) chronically depressed if they had elevated CESD scores on four or five assessments. This operationalization of chronicity is limiting given it reduces longitudinal data by analyzing it in a cross se ctional manner, and also because it neglects the potential importance of sequential episod es of chronic depression. For example, women with elevated scores on the CES-D at six and 36 months would be classified the same as women with elevated scores at si x and 15 months, although logic implies that a child exposed to sequential episodes may suffe r more adverse outcomes than a child who is exposed to symptoms for a period of time, followed by remission, than a recurrent episode. Results from multivariate ANCOVA analyses revealed that, after controlling for maternal education, child sex, birth orde r, and site, women who reported being chronically depressed rated their children as less socially competent than women who reported depression some of the time; and wo men who reported being depressed some of the time rated their children as less social ly competent than women who never reported feeling depressed. The researchers next te sted the mediating effect of maternal sensitivity on the association between chr onic depression classifi cation and maternal reports of social competence at 36 months revealing no evidence of an indirect relationship. Tests for a moderating influen ce of maternal sensitivity however did result in significant findings reveali ng that depression chronicity had a stronger influence when maternal sensitivity was low. Comparisons of chronicity group scores when maternal sensitivity was high indicated that on ratings of social competence, children whose mothers were in the two depression groups (s ometimes or chronic) did not differ from

PAGE 78

56 those whose mothers were never depressed, whereas the group differences were large when maternal sensitivity was low. Another study from Finland (Luoma et al., 2001) investigated the impact of postpartum depression as measured via the Edinburgh Postnatal Depression Scale (EPDS) (Cox, Holden, & Sagovsky, 1987) duri ng pregnancy, one week postpartum, two months, six months, and eight to nine years postpartum on ch ild social competence when the children were eight to ni ne years old. The sample size participating in all waves of data collection was 147 mother-infant dyads. Child social competen ce was measured via the Child Behavior Checklist/4-18 (CBCL/418) social competence scale (Achenbach, 1991), which is based on maternal reports of the child’s activities (hobbies, tasks, chores), functioning in social relationships (w ith peers, siblings, a nd parents), and school achievement. Logistic regression analyses were conduc ted adjusting for maternal education, marital status, family socioeconomic status, number of children in the family, and child sex. Although five assessments of depressi on were ascertained (prenatal, one week postpartum, two months, six months, and eight to nine year s postpartum) three dichotomous variables were included in th e model – presence of prenatal depression, presence of postpartum depression at two mont hs, and presence of depression at eight to nine years. Results revealed that matern al depression at two months significantly predicted low levels of child social compet ence, controlling for pr enatal depression and concurrent depression at eight to nine years. Prenatal depression and current depression at eight to nine years were not associated with low levels of social competence, in the full analytic model. These results must be cons idered with discretion however given that the

PAGE 79

57 authors do not mention consideration of collin earity between the multiple assessments of depression being included in a single analytic model, thus estimates are likely to be biased. Lack of Consideration for Depression Characteristics. One study (Carter et al., 2001) intended to examine the effect between postpartum depression severity and child social competence at thirty months usi ng the Infant-Toddler Social and Emotional Assessment (ITS EA) (Carter & Briggs-Gowan, 1993), but investigated only the impact of a lifetime history of depression due to sample size limitations. Postpartum depression was sc reened for using the CES-D followed by a diagnostic interview using the Structured Clinical Interview for the DSM-III-R-NonPatient Version which also assessed lifetim e history of depression (SCID-NP) (Spitzer, Williams, Gibbon, & First, 1992) during the pren atal period, four months postpartum, and 14 months postpartum. Depression seve rity was operationalized by classifying women into one of three categories: (1 ) no lifetime psychopathology; (2) lifetime depression only; and (3) lifetime depression a nd at least one other psychopathological condition such as anxiety, eating disorder, an d/or substance abuse (comorbid group). The researchers were particularly intere sted in exploring significant moderating effects of child sex, hypothesizing increased risk for poor social competence in boys of depressed mothers when compared to girls of depressed mothers. Due to limitations in power given the small sample of 69 moth er-infant dyads, the researchers claimed inability to concurrently examine de pression severity group (no psychopathology, depression only, or comorbid group) and infant sex. The two depression groups (depression only and comorbid women) we re thus collapsed together, no longer

PAGE 80

58 considering severity of symptomatology but instead a lifetime history of depression in women (prior to childrearing and postpa rtum). Results from multivariate ANOVA analyses controlling for income revealed no si gnificant differences in maternal ratings of social competence. Another longitudinal investigation explor ed the association between postpartum depression and child social competence at age five among a sample of primiparous mothers (Murray et al., 1999). Child social competence at age five was assessed via time-sampled observational ratings of child beha vior at school during free play. For each thirty-second time period, observers rated child ren on the following be haviors: (1) social play that consisted of play with other ch ildren including prete nd play; (2) passively watching other children; and (3) interacting w ith the teacher. Two additional events were also recorded including the occurrence of fr iendly approaches to the study child by other children, and for the subgroup of children who received such approach es, the nature of the child’s response (either positive or nega tive). Postpartum depression was assessed using the Edinburgh Postnatal Depression Scale (EPDS) (J. Cox et al., 1987) at six weeks postpartum. Those identified with elevated levels of symptomatology were diagnosed using the Standardized Psychiatric Inte rview (SPI) (Goldberg, Cooper, Eastwood, Kedward, & Shepherd, 1970). All participants were also interviewed using the lifetime version of the Schedule for Affective Disord ers and Schizophrenia (SADS-L) (Endicott & Spritzer, 1978) at 18 months and 5 years to retrospectively a ssess whether or not women had experienced depression in the interval between follow-up interviews. Although the study design was longitudinal w ith prospective assessment of social competence at five years, depression afte r six weeks was assessed retrospectively

PAGE 81

59 requiring validity of recall. Women were cla ssified as either experiencing postpartum depression or being nondepresse d for analysis purposes. Results from multivariate analyses revealed that whether or not children engaged at all in social play with other children, or received friendly approaches for play was unrelated to exposure to postpartum de pression. However, among children who did receive a friendly approach from another ch ild, the quality of re sponsiveness differed according to exposure to postpartum depressi on, with children of depressed mothers being more likely than controls to respond negatively. The resear chers then examined whether the quality of attachme nt to the mother at 18 months mediated the association between exposure to postpartum depression a nd response to social approaches with no significant association detected. In add ition, no significant rela tionships were found between exposure to postpartum depression and passively watching while other children play or interacting with the teacher. Hay and Pawlby (Hay & Pawlby, 2003) inve stigated the long-term impact of postpartum depression on prosocial behaviors, particularly cooperation, in children at age four and 11 among a sample of 149 urban London families. Prosocial behavior at age four was rated via observation during ad ministration of the McCarthy Scales of Children’s Abilities (McCarthy, 1972) and thro ugh an experimental cooperation task. This task involved mother and child pairs to cooperate together in copying a picture of a house using an Etch-A-Sketch toy. Mother and ch ild were each assigned a dial to use in constructing the house and were told that e ach person could only use their dial, thus requiring both mother and child to work together to meet an interdependent goal. At age 11, maternal and child reports of prosocial behavior were assessed using the Strengths

PAGE 82

60 and Difficulties Questionnaire (SDQ) (Goodman & Gotlib, 1999) which measures five aspects of prosocial behavior including: (1) consideration of other people’s feelings; (2) sharing with other children; (3) helpful if some one is hurt, upset or feeling ill; (4) kind to younger children; and (5) volunteering to help others including pare nts, teachers, and other children. Postpartum depression wa s assessed using the Clinical Interview Schedule (SCI) (Goldberg et al., 1970) during pregnancy, three and 12 months postpartum. Retrospective assessment of depr ession of maternal depression was assessed using the Schedule for Affective Disorder s and Schizophrenia (SADS-L) (Endicott & Spritzer, 1978) at four and 11 years postpartum. Women were classified as being either depressed or nondepressed during the postpartu m period (three months to 11 years) for analysis. Results from ANOVAs revealed that there was no association between maternal reports of depression during the first 11 years of the child’s life and ratings of cooperation at age four. Child ren of depressed mothers were no more or less cooperative than other children. Tests for an interacti on between child sex and mother’s history of depression also showed no signi ficant relationships. Furthe r, there was no significant association between mother history of depr ession and prosocial behavior at age 11. There was however a significant interaction between informant (mother and child) and mother history of depression; mothers with a history of depression ra ted their children as less prosocial but the children of depressed mothers saw themselves as more prosocial than other children. Findings suggest th at perhaps nonmaternal reports of social competence (and other child behaviors for that matter) may be more appropriate in the context of maternal depression given that women with a history of depression or

PAGE 83

61 currently experiencing depression may be more likely to rate their children negatively (Meredith & Noller, 2003). Summary. In summary, there is mixed empiri cal support for a relationship between mother’s postpartum depression and child soci al competence with litt le consideration of the impact of various characteri stics of depression on social competence. In fact only one study investigated a specific feature of pos tpartum depression, chronicity of exposure, which did find a positive relationship between chronicity of symptoms and child social competence at 36 months. Levels of child social competence we re lower among women classified as chronically depressed compared to those classified as sometimes depressed, and women classified as sometimes depressed rated their children as having lower levels of social competence compared to women who were never depressed. All other studies reviewed investig ating a relationship between postpartum depression and child social competence did not consider the heterogeneity of women’s experiences with depression during the pos tpartum period. These studies employed a dichotomous measure of depression (whether or not symptoms were present); some of which were lifetime histories of depression a nd not exclusively depression occurring in the postpartum period. Findings from these investigations we re mixed likely due to methodological differences among the various studies. In additio n to differences in the way and timing in which postpartum depression was assessed, child social competence was also measured using various tools and at va rious time points along the ch ild’s trajectory. Some studies utilized structured quantitative surveys incl uding the social competence scales of the

PAGE 84

62 Child Behavior Checklist, the Infant-Toddler Social and Emotional Assessment, and the Adaptive Social Behavior Inventory, whereas others employed qualitative measures including the Etch-A-Sketch shared task and obs ervation of child social play at school to develop ratings of child c ooperation and sociability. Studies employing observational assessmen ts by an independent rater of child social competence largely failed to find signifi cant associations be tween child exposure to postpartum depression and social comp etence, with the exception of comparisons between children of depressed and non-depr essed mother when approached by other children in free play. These children we re found to respond more negatively when approached by other children to engage in pl ay than children of nonde pressed mothers. In fact, none of the 28 children of nondepresse d mothers reacted negatively to social approaches by other children whereas ni ne of 30 children of depressed mothers responded negatively. Quantitative assessments of social competence also failed to find strong support for a relationship between postpartum depression and social competence. One study did report a positive relationship between postpartum depression at six months and child social competence at eight to nine years as measured via the CBCL/4-18 social competence scale; however, the regression model built to explore the relationship simultaneously included additional assessmen ts of maternal depression without any collinearity diagnostics reported. Given that these repeated measures of depression were highly likely to be correlated with one anot her, research findings may lose credibility during interpretation.

PAGE 85

63 Another interesting finding from the literature on the effects of postpartum depression and child social competence is the notion that alternative raters of competencies may be more appropriate in th e context of postpartum depression. It has been documented that mothers experiencing c oncurrent depression, and even those with a history of depression may perceive their child more negatively. When maternal and child reports of child social competence were assessed via a structur ed rating tool and compared with one another, mothers with a history of depression ra ted their children as less prosocial than mothers without a hist ory, and children of depressed mothers saw themselves as more prosocial than other children. Although it is not clear which assessment was more accurate, this finding sugg ests that future re search should attempt to corroborate maternal reports of competen cies with additional rater reports such as teachers, fathers, or other caregivers. Little empirical research has consid ered mediating and moderating effects on the relationship between postpartum depressi on and child social competence. Perhaps once the literature base has increased in qua ntity and strengthened in quality, these influences can be investigated further. On e factor limiting the inve stigation of mediating and moderating influences in the current literature is relatively small sample sizes. One study reviewed above did investigate the medi ating and moderating effects of maternal sensitivity on chronicity of postpartum de pression on social competence at 36 months. This study utilized a large sample of 1,215 mo ther-infant dyads that were followed from one month postpartum to 36 months. There was no evidence of a me diating effect of maternal sensitivity, but a moderating influence was detected. In dyads where maternal

PAGE 86

64 sensitivity was low, chronicity of postpar tum depression had a gr eater impact on child social competence than in dyads wh ere maternal sensitivity was high. Further research is needed exploring the impact of postpartum depression on child social competence using larger samples gi ven inconclusive findings at this time. In addition, greater attention to the heterogeneity of women’s experiences with depression including chronicity, severity, and timing of onset is imperative to identify differential mechanisms of influence. Currently, different features of depression have rarely been considered. Although maternal reports of child behaviors are often easiest to obtain, future research should utilize multiple raters of child competencies, if possible, to reduce reporting bias of maternal re ports in the context of postp artum depression. Further, exploring mediating and modera ting influences on the relationship between postpartum depression and child social competence would contribute to more targeted prevention efforts aimed to promote the healthy developm ent of children in the context of maternal depression. Behavioral Difficulties Display of problem behaviors in ch ildren is a sign of adverse social development. Researchers interested in unde rstanding developmental tr ajectories of child behavioral problems often utilize the predom inant paradigm of classifying behavioral difficulties as either externalizing (undercontrolled) or internalizing (overcontrolled) problems. Externalizing problems include a broad array of aggressive behaviors, antisocial characterist ics, and hyperactivit y (Cicchetti & Tot h, 1991). Youth with externalizing problems have cons istently been described as being aggressive, disruptive, acting-out, defiant, oppositional, and hyperactive (Merrell, 1999a). Internalizing

PAGE 87

65 problems include a broad domain of sympto ms related to depression, anxiety, social withdrawal, and somatic complaints (Me rrell, 1999b). Both internalizing and externalizing problems are considered to be deficits in a child’s social development. Young children with behavior problems are at increased risk for mental health disorders and antisocial behavior in early adulthood (Capsi, Moffitt, Newman, & Silva, 1996). Understanding the relationship betw een postpartum depression and behavioral difficulties in children holds prevention implications for adult psychopathology and antisocial behaviors. Chronicity. Several studies have investigated the impact of chronicity of maternal depressive symptoms and behavioral difficulties in chil dren ranging from 18 months to seven years of age. Despite variation in measuremen t of behavior problems, study design, and operationalization of chronicity of depressi ve symptoms, there is moderate empirical support for a relationship between chronic pos tpartum depression and display of problem behaviors from toddlerhood to ear ly elementary school age. One longitudinal study (Murray, 1992) conducted among a sample of primiparous mothers investigated the impact of chr onic maternal depression through six weeks postpartum on maternal reports of child beha vior problems at 18 months of age assessed via the Behavioural Screening Questionnair e (BSQ) (Richman & Graham, 1971). The BSQ was designed to assess feeding difficu lties, sleep problems, temper tantrums, excessive dependency, miserable mood, peer re lationships, and lack of compliance. Although the study design was longitudinal with prospective assessment of behavior problems at 18 months of age, chronicity of depression was determined using both

PAGE 88

66 retrospective and prospective means. Wome n were recruited for participation in the study during the hospital stay following birt h and completed the Edinburgh Postnatal Depression Scale (EPDS) (J. Cox et al., 1987) at six weeks postpartum. Those identified with elevated levels of symptomatology were interviewed using the Standardized Psychiatric Interview (SPI) (Goldberg et al., 1970) and classified as experiencing depression according to Research Diagnosti c Criteria (RDC) (Spitzer, Endicott, & Robins, 1978). All participants were also in terviewed using the lifetime version of the Schedule for Affective Disorders and Schi zophrenia (SADS-L) (E ndicott & Spritzer, 1978) to retrospectively assess whether or not women had experienced a previous history of depression prior to childbirth. Both asse ssments of depression (six weeks postpartum and retrospective recall of a lif etime history) were used to classify women into one of four categories of chronicity : (a) no previous history of depression or depression since delivery (n=42); (b) no previous history of depression, but history since delivery (n=40); (c) a previous history of de pression, but no depression sin ce delivery (n=14); and (d) a previous history of depression and depr ession since delivery (n=21). Although operationalization of chronicity consisted of multiple time assessments, one required validity of recall due to the re trospective nature of inquiry; and chronicity was analyzed cross-sectionally given data from pre-preg nancy and 6 weeks postpartum were collapsed into a single variable. Results of chi-squares tests revealed that although women who had experienced postpartum depression (postpartum depres sion only and those with postpartum depression and a previous history) were more lik ely to have an infant with at least some degree of behavioral disturbance as compar ed to women who had never been depressed,

PAGE 89

67 there were no detectable differences between mothers who had experienced only postpartum depression compared to those w ho had experienced postpartum depression and had a previous history of depression. Thus chronicity of symptoms, as defined in this study, did not appear to impact adverse beha viors in 18 month-old children. No other influences were considered given the bi variate nature of the data analysis. Another longitudinal study investigated th e impact of chronic maternal depression symptoms on child problem behaviors at ag e three among a sample of 124 children of mothers with and without a history of de pression (Dawson et al., 2003). Problem behaviors were assessed via the Child Be havior Checklist (CBCL) (Achenbach, 1979) and the Child Adaptive Behavior Inventor y (CABI) (Cowan, Cowan, Schulz, & Heming, 1994); in which scores from the two instrument s were combined to assess internalizing problems, externalizing problems, and total behavior problems. Postpartum depression was measured using the Center for Epidem iological Studies Depression Scale (CES-D) (Radloff, 1977) to identify potential cases, and evaluate d diagnostically using the Structured Clinical Interview for DSM-III-R (SCID) (Spitzer, Williams, Gibbon, & First, 1989). Measurements were conducted at 14 months, 24 months, and three and a half years postpartum. At each assessment, mothers were interviewed about their experiences with depressive symptoms not only during th e past week as the CES-D is intended, but for the time interval between consecutive a ssessments, thus larg ely retrospective in nature. Further, despite longitudinal da ta on postpartum depre ssion, chronicity of depression was reduced to a sing le variable and analyzed in a cross-sectional manner. Women were categorized into three groups including nondepressed, remitted depression (those experiencing depression during the first tw o years of their child ’s life), and chronic

PAGE 90

68 depression (those experiencing recurrent depres sion during the first three and a half years of their child’s life). An additional measure of depressi on chronicity was computed based on the total number of months childre n were exposed to postpartum depression during the first three and a half years. Results from one-way ANOVAs revealed si gnificant main effects of maternal chronicity group for internalizing, externa lizing, and total problem behaviors; with chronically depressed mothers having child ren with higher le vels of total and internalizing problems than children of mothers with remitted depression and nondepressed mothers. Mothers with remitted depression had the highe st mean level of externalizing problems, although the differen ce between remitted and chronic depression groups was small. Multiple regression analys is was also conducted revealing that after controlling for the number of months of expos ure to postpartum depression after the age of two years, the number of months of postp artum depression during the child’s first two years of life remained a significant predicto r of both externalizing and total problem behaviors at age three and a ha lf years, but not for internal izing problems. These results are problematic to interpret however give n that the authors do not mention any consideration of collinearity between the number of months of depression during the first two years of the child’s life and the number of months of depression after the age of two; variables included in the multivariate model that are extremely likely to have a high correlation and thus result in biased estimates. The researchers next investigated the medi ating effect of contextual risk factors on the relationship between the number of m onths of postpartum depression during the first three and a half years and total probl em behaviors, via path analyses. Results

PAGE 91

69 revealed that after controlling for contextual risk factors such as marital adjustment and available social support, the direct relationship between th e total number of months the child was exposed to postpartum depression and total problem behaviors was no longer significant. This latter findi ng suggests that chronic exposur e to postpartum depression may exert its influence on child behavior pr oblems through its asso ciation with a less supportive and more stre ssful environment (Dawson et al., 2003). Yet another longitudinal study investiga ting the impact of postpartum depression chronicity and child display of problem be haviors was undertaken utilizing the National Institute of Child Health and Human Deve lopment Study of Early Child Care (NICHD SECC) (NICHD Early Child Care Research Network, 1999b). One advantage of this study over previous studies discussed includes a large sample size of mother-infant dyads (n=1,215) from multiple sites across the nation. Child problem behaviors were assessed via maternal reports at age three using th e Child Behavior Checklist – 2/3 (CBCL-2/3) (Achenbach, Edelbrock, & Howell, 1987) wh ich derived scale scores for social withdrawal, depression, sleep problems, somatic problems, aggressive behavior, and destructive behavior; and the Adaptive Social Behavior Inventory (ASBI) (Hogan et al., 1992) which yielded a scale scor e for disruption. To measur e problem behaviors at age three, the scales from the CBCL-2/3 and th e ASBI were combined into one behavior problem composite. Postpartum depression wa s assessed via the CES-D at one, six, 15, 24, and 36 months. Once again, chronicity of depression was reduced to a single variable despite rich data on symptomatology over the firs t three years. Women were classified as (a) never depressed if they scored belo w threshold values on the CES-D at every assessment; (b) sometimes depressed if they sc ored above threshold values at least once,

PAGE 92

70 but fewer than four times; and (c) chronica lly depressed if they had elevated CES-D scores on four or five assessments. In addi tion to the loss of data richness by collapsing longitudinal data into a single variable, this operationalization of de pression chronicity also neglects the potential importance of se quential episodes of chronic depression. For example, women with elevated scores on the CES-D at six and 36 months would be classified the same as women with elevated scores at six and 15 months, although logic implies that a child exposed to sequential episodes may suffer more adverse outcomes than a child who is exposed to symptoms for a period of time, followed by remission, then a recurrent episode. Results from multivariate ANCOVA analyses revealed that after controlling for maternal education, child sex, birth orde r, and site, women who reported being chronically depressed rated th eir children as more problematic than women who reported depression some of the time; and women who reported being depressed some of the time rated their children as more problematic than women who neve r reported feeling depressed. The researchers next tested the mediating effect of maternal sensitivity on the association between chronic depression cl assification and problem behaviors at 36 months, revealing no evidence of an indirect relationshi p. Tests for a moderating influence of maternal sensitivity also did not result in significant findings. Civic and Holt (Civic & Holt, 2000) evaluated the asso ciation between postpartum depressive symptoms and child behavior problems am ong a large (n=5,303) nationally representative sample of U.S. mo thers of normal birthweight infants from the 1988 National Maternal and Infant Health Surv ey. Women completed the initial survey at a mean of 17 months postpartum and we re followed up in 1991 where mothers were

PAGE 93

71 asked to report the frequency of problem beha viors in their children including ratings of temper tantrums, degree of child happine ss, degree of difficulty managing the child, fearfulness, and child difficulty getting along with other children. A composite problem behavior measure was created. Children were considered to display problem behaviors if their mother reported the presence of at least three of the five problems described above. Postpartum depression was assessed using the CES-D in 1988 and 1991 and women were classified into the following ch ronic categories: (a) depressed on both the 1988 and 1991 surveys; (b) depressed on the 198 8 survey but not the 1991 survey; (c) depressed on the 1991 survey but not the 1988 survey; and (d) not depressed on either survey. Bivariate analyses revealed that women in each depression category (1988 only, 1991 only, and 1988 and 1991) were more likely to have a child with problem behaviors as compared to nondepressed women; with women who were depressed at both time points reporting a higher percentage of children with behavior problems (23.2%) compared to 15.3% of women who were depressed in 1991 only, 11.2% of women who were depressed in 1988 only, and 6.5% of women who were nondepressed. Logistic regression analyses controlling for househol d income, maternal education and marital status revealed that depression in 1988 only was a significant predictor of child problem behaviors in 1991, depression in 1991 only was a significant predicto r of child problem behaviors, as well as depre ssion in 1988 and 1991, as compared to the reference group of nondepressed women. Notable to emphasize is the increase in odds ratios across categories of chronicity. For example, women depressed in 1988 only were 1.6 times more likely to rate a child as displaying problem behaviors compared to nondepressed women; women depressed in 1991 only were 2.3 times more likely to rate a child as

PAGE 94

72 displaying problem behaviors compared to nondepressed women; and women depressed in 1988 and 1991 were 3.6 times more likely to rate a child as displaying problem behaviors compared to nondepr essed women. The increase in risk across categories of chronicity provides evidence of an association between postpartum depression chronicity and display of problem behavior s at a mean of 36 months pos tpartum. No mediating or moderating influences were investigated. Another study (Brennan et al., 2000) inve stigating the impact of depression chronicity on child behavioral problems f ound a positive association between the number of times a mother reported elevated depres sion scores during four time points (during pregnancy, immediately postpartu m, six months postpartum, and five years) via the Delusions Symptom-States Inventory (DSSI) (Bedford & Foulds, 1978). Scores ranged from zero to four. Child behavior at age five was assessed using the Child Behavior Checklist –4/18 (CBCL/4-18) (Achenb ach, 1991), which measured aggression, oppositional behavior, hyperactivity, anxiety, w ithdrawal, and depression. Results from multiple regression analysis controlling for covariates including child sex, parity, maternal age, maternal education, marital st atus, and family home structure revealed a significant main effect of nu mber of postpartum depression episodes during the first five years on child problem behavior scores at age five. As with earlier studies reviewed in this section, the internal va lidity of operationalization of chronicity of depression was limited due to crude reduction of longitudinal assessments were collapsed into a single variable. No mediating relationships were explore d. Depression severity (a continuous measure of the worst maternal depressive episode ranging from zero to seven) moderated

PAGE 95

73 the association between number of depres sion episodes reported during the first five years and child problem behavior s at age five resulting in a si gnificant interaction effect. The researchers acknowledged the problem of collinearity by including both depression chronicity and severity in a single model give n they were correlated 0.81 with each other. As a result, they created a single variable encompassing both severity (moderate vs. severe) and chronicity (one re port vs. two or more reports) These categories consisted of: (a) neither severe nor chronic; (b) chroni c but not severe; (c) severe but not chronic; and (d) both chronic and severe. Results from a 2X2 ANCOVA revealed a significant interaction effect between ch ronicity and severity with children of mothers who had experienced both chronic and severe depressive symptoms having significantly more behavior problems than children in th e other three comparison groups. Although a definite contribution to the knowledge base co ntributing to understa nding the effects of women’s experiences of postpartum depres sion on child behavioral difficulties, collapsing longitudinal data for cross-secti onal analysis limits the full potential of prospective longitudinal studies. Severity. The Brennan study (Brennan et al., 2000) al so explored the di rect effect of postpartum depression severity on child prob lem behaviors at ag e five revealing a significant association between the worst D SSI score across the four assessment time points and total problem behaviors assessed using the CBCL/4-18. Op erationalization of depression severity in this way is limiting gi ven it does not consider the possibility that depression severity experienced at various points along the developm ental trajectory may have differential impacts on display of problem behaviors.

PAGE 96

74 Timing of Onset. In the study by Civic and Holt (Civic & Holt, 2000) described above that evaluated the association between chronicity of postpartum depressive symptoms and child behavior problems among mother-infant dyads participating in the 1988 National Maternal and Infant Health Survey, the way in which chronicity was operationalized and treated during analyses also enabled the researchers to dr aw conclusions regarding the timing of symptom onset. Recall that postpartum depression was assessed using the CES-D in 1988 and 1991 and women were cl assified into the following chronic categories: (a) depressed on both the 1988 a nd 1991 surveys; (b) depressed on the 1988 survey but not the 1991 survey; (c) depr essed on the 1991 survey but not the 1988 survey; and (d) not depressed on either survey. Results from separate logist ic regression models revealed that after adjusting for household income, maternal education and marital status, wome n who experienced depression in 1988 only, women who experien ced depression in 1991 only, and women who experienced depression at both time points, were significantly more likely to have a child with problem behaviors as compar ed to the reference group of nondepressed women; with increasing likelihood of problem be haviors across chronici ty classifications. What was interesting in this study was the finding that women who were depressed in 1991 only had a higher likelihood of having a chil d with problem behaviors as compared to nondepressed women (OR=2.3), than wome n who were depressed in 1988 only as compared to nondepressed women (OR=1.6). This finding suggests that later onset depression (mean assessment at 36 months postpartum) may pose a greater risk to potential adverse outcomes related to behavioral difficulties than earlier onset depression

PAGE 97

75 (mean assessment at 17 months postpartum). Although important to consider in future research endeavors, this finding may also be a methodological artifact of the timing in which assessments were taken – particul arly the simultaneous measurement of postpartum depression and maternal reports of problem behaviors. Given that depressed individuals are more likely view their world more ne gatively than nondepressed individuals (Meredith & Noller, 2003), it is probable that mothers experiencing depression in 1991 would be more likely to al so report more negative behaviors in their children. An additional study considering the impact of timing of symptom onset on child problem behaviors has previously been revi ewed in both the chronicity and severity sections (Brennan et al., 2000). In this study, depression was assessed using the DSSI during pregnancy, immediately postpartum, six months postpartum, and five years postpartum; child behavior at age five was assessed using the CBCL/4-18. The authors reported investigating the impact of timing of symptom onset on chil d behavior problems only among women who had a single elevated score during the five-year assessment period, to avoid the confounding effect of severi ty and chronicity of depression. Results from ANCOVA analyses adjusting for child se x, birth order, maternal age, maternal education, family income, and maternal marital status indicated a significant main effect of timing of moderate leve ls of postpartum depressive symptoms on child problem behaviors, as well as a significant main effect of timing of severe levels of postpartum depressive symptoms. Post hoc tests revealed that behavior problem scores were higher for children whose mothers report ed moderate levels of depressive symptoms only at six months or at five years than for children whose mothers reported moderate levels only

PAGE 98

76 during pregnancy or at birth. Further, there was a significan tly higher level of problem behaviors in children whose mothers reported se vere levels of depressive symptoms only at age five than in children whose mother s reported severe levels only during pregnancy, only at birth, or only at six months of age. Although significant contributions to the literature, these results shoul d be interpreted with caution and replicated with studies incorporating more frequent assessments of postpartum depr ession with more equal time intervals in between. It is likely that child behavior problems were higher among mothers with moderate levels of symptoms w ith onset occurrence at six months or five years postpartum due to the fact that advers e outcomes are theorized to be related to postpartum depression, not depression during pregnancy that does not continue into the postpartum period, nor depression in the imme diate postpartum peri od that more likely represents postpartum blues rather than pos tpartum depression which has been discussed to be a mild and temporary condition in whic h symptoms often subside within two weeks without medical help. The latter finding th at higher levels of problem behaviors in children were detected among mothers who reported severe levels of depressive symptoms only at age five may once agai n be a methodological artifact of assessing problem behaviors simultaneously with matern al depression (depressed mothers are more likely to rate their children more negativ ely in terms of problem behaviors than nondepressed mothers). Lack of Consideration for Depression Characteristics. Most studies of associations betw een postpartum depression and child developmental outcomes fail to consider th e characteristics of postpartum depression in relation to differential risk for adverse outcome s. A review of this literature will be

PAGE 99

77 presented below. Although a few studies fa il to show a positive association between postpartum depression and child proble m behaviors during toddlerhood or early elementary school age, there is substantia l research supporting an association. The same sample of mother and infant dyads described in Murray’s (Murray, 1992) study on the chronicity of depressive symptoms through six weeks postpartum on behavior problems at 18 months were follo wed-up at five years postpartum to assess behavioral adjustment (Murray et al., 1999) via the Rutter A2 ques tionnaire (Rutter, Tizard, & Whitmore, 1970), which assesses child neurotic and antisocial behaviors. The mother’s psychiatric history since the 18month interview was also taken using the SADS-L, once again requiring va lid recall of depressive symptoms experienced during the past three and a half years. Episodes of depression during this time frame were recorded along with information on the timi ng of onset and remission of symptoms to enable the construction of a c ontinuous measure of chronicity consisting of the overall duration in months of the child’s exposure to maternal depression during the first five years of life. Although the researchers inte nded to examine the impact of postpartum depression chronicity on behavior problems at five years, its skewed distribution due to the high number of women who had not e xperienced any depression during the study period was reported as the reason for inst ead employing a dichotomous measure of postpartum depression as the major independent variable under scru tiny. Results from multivariate analysis revealed a significant main effect of postpartum depression during the first five years and total problem behavi ors (neurotic and antisocial) at age five, controlling for social class and child sex. Th e mediating effects of parental conflict and attachment security at 18 months were ne xt investigated revealing no associations.

PAGE 100

78 Another study (Carter et al., 2001) intended to exam ine the effect between postpartum depression severity and display of problem behaviors at thirty months using the Infant-Toddler Social and Emotional A ssessment (ITSEA) (Carter & Briggs-Gowan, 1993) and the CBCL/2-3 (Achenbach et al., 1987). The ITSEA assesses externalizing problems (activity/impulsivity, aggre ssion/defiance, peer aggression, and depression/withdrawal), in ternalizing problems (inhibition/separation, fears, and depression/withdrawal), and regulation problems (sleep, eat ing, and toileting). The CBCL/2-3 provides scale scores for both ex ternalizing and inte rnalizing problem behaviors. Postpartum de pression was screened for us ing the CES-D followed by a diagnostic interview using the Structured Clinical Interview for the DSM-III-R-NonPatient Version which also assessed lifetime history of depression (SCID-NP) (Spitzer et al., 1992) during the prenatal period, 4 months postpartum, and fourteen months postpartum. Depression severity was operati onalized by classifying women into one of three categories: (1) no lifetime psychopathol ogy; (2) lifetime depr ession only; and (3) lifetime depression and at least one other ps ychopathological condi tion such as anxiety, eating disorder, and/or substa nce abuse (comorbid group). The researchers were particularly intere sted in exploring significant moderating effects of child sex, hypothesizing increased risk for problem behaviors in boys of depressed mothers when compared to girls of depressed mothers. Due to limitations in power given the small sample of 69 moth er-infant dyads, the researchers claimed inability to concurrently examine de pression severity group (no psychopathology, depression only, or comorbid group) and infant sex. The two depression groups (depression only and comorbid women) we re thus collapsed together, no longer

PAGE 101

79 considering severity of symptomatology but instead a lifetime history of depression in women (prior to childrearing and postpartum). Results from multivariate ANOVA analyses controlling for income revealed significant differences in CBCL/2-3 externaliz ing and internalizing problems with boys and girls in the depressed gr oups having higher levels of problem behaviors as compared to boys and girls in the nondepressed group. A significant interacti on between depression status and child sex was also detected with daughters of depressed mothers having higher externalizing scores compared to daughter s of nondepressed moth ers. Daughters of depressed mothers had scores comparable wi th those of all boys and daughters of nondepressed mothers had the lowest scores. There were no significant differences for the internalizing scale, or the problem behavior scales of the ITSEA. Previous literature has suggested a greater biological vulnerabil ity in boys to experiences of postpartum depression with boys having highe r rates of adverse outcomes when exposed to maternal psychopathology (Murray & Cooper, 1997). Results from this study s uggest that in the context of maternal depression, girls may be equally at risk for adverse externalizing problem behaviors given their rates were si gnificantly higher than girls of nondepressed mothers and comparable to rates of boys from both the depressed and nondepressed groups. Shaw and Vondra (Shaw & Vondra, 1995) explored the impact of postpartum depression on behavior problems at age three among a sample of 100 low-income mother-infant dyads. Problem behaviors were assessed using the CBCL/2-3, which generated scale scores for internalizing and externalizing problems. Postpartum depression was assessed via the Beck’s Depression Inventory (BDI) (Beck, Ward,

PAGE 102

80 Mendelson, Mock, & Erbaugh, 1961) at recru itment (6-11 months), 12, 18, 24, and 36 months. Women were asked to report the prevalence of depressive symptoms experienced during the past six months rather than the past week to increase the time period in which depression was ascertained. Three variables of de pression were created based on the repeated measures of the BDI including year one depression (women with elevated scores at recruitmen t or 12 months); year two depr ession (women with elevated scores at 18 and 24 months); and year three depression (women with elevated scores at 36 months). Separate multiple regression mode ls were run for each depression variable. Although the researchers do uniquely consider th e presence of depression at various time points (years one to three) al ong the child’s developmental tr ajectory, the presence of depression within a given year of the child’s life does not ne cessarily represent the timing of initial onset – just the presence of symptoms in a given year. Analyses were conducted separately fo r boys and girls. Regression models among boys revealed a significant association of depression during ye ar three only with externalizing problems at 36 months, afte r controlling for a composite measure of attachment security across 12 and 18 months of age. For internalizing problem behaviors, only year one depr ession scores were significant after controlling for family income, maternal education, and quality of the home environment at 15 months. For girls, BDI depression scores were not signifi cant during any year. These findings suggest an increased vulnerability of boys to the exposure of postpartum depression. The finding that year three depression wa s related to externalizing probl em at 36 months is likely due to maternal reporting bias given there is ev idence that maternal depression may inflate maternal reports of child behavior pr oblems (Goodman & Brumley, 1990; Radke-

PAGE 103

81 Yarrow, Cummings, Kuczynski, & Chapman, 1985). Although a ttachment security was included in the model predicting externaliz ing problems and separate analyses were conducted for both boys and girls, no fo rmal hypotheses addressed mediating and moderating influences, nor were contextual ri sk factors consistently included in the models. In a correlational study among a small sample of 34 mother-infant dyads, maternal depression assessed at 44 and 54 months via the CES-D were significantly correlated with both internalizing and external izing problem behaviors as measured using the CBCL/4-18 at 54 months of age (Hubbs-Tait et al., 1996). Results from regression analyses controlling for sex revealed that ma ternal depression at 44 months significantly predicted externalizing problem behaviors at 54 months, and that maternal depression at both 44 and 54 months were significant predicto rs of internalizing problem behaviors at 54 months. However, these results should be interpreted with caution given that the authors do not consider the high collinearit y between depression scores at 44 and 54 months. Including two variables in a single m odel that are highly correlated with one another yields biased estimates. A preschool follow-up study of infants of dysphoric mothers was conducted when the children were a mean age of 3.9 year s. Postpartum depression (Field, Lang, & Martinex, 1996) was measured at three m onths using the BDI, and child problem behaviors were assessed using the CBCL-2/ 3. MANOVA results reve aled a significant main effect of postpartum depression for child depression (internalizing), aggressive behavior (externalizing), a nd destructive behavior (ext ernalizing); with dysphoric mothers rating their children as having highe r levels of depression, aggressive, and

PAGE 104

82 destructive behaviors, as compared to nondys phoric mothers. Women in this sample were not asked about current levels of de pression at the time of maternal report of problem behaviors. In a longitudinal investig ation of 92 women and thei r first-born children, no association was found between postpartum de pression and child problem behaviors at four years old (Caplan et al., 1989). Problem behaviors were assessed via maternal report using the Behavioural Screening Questi onnaire (Richman & Graham, 1971), which assesses feeding difficulties, sleep problem s, temper tantrums, excessive dependency, miserable mood, peer relationships, and lack of compliance. A psychiatrist who was blinded to the mothers’ psychiatric historie s also rated children on problem behaviors. Although mothers who were concurrently depr essed at 4 years reported significantly more behavioral difficulties in their ch ildren, no evidence supported a relationship between postpartum depression and be havioral problems at age four. Another study found no asso ciation between postpartu m depression and child problem behaviors at age four and a half (Phillipps & O'Hara, 1991). Women were initially recruited in to the study during the second trim ester of pregnancy and followedup at four and a half years to assess probl em behaviors using the CBCL/4-18. Postpartum depression was assessed using the BDI and a se mi-structured interview adapted from the Schedule of Affective Disorders and Schiz ophrenia (Endicott & Spri tzer, 1978) at both recruitment and the follow-up interview. Bi variate analyses revealed that postpartum depression was not significan tly associated with scores on the intern alizing or externalizing scales of the CBCL/4-18; how ever RDC-defined mate rnal depression at four and a half years was significantly associated with both internalizing and

PAGE 105

83 externalizing behaviors, and postpartum de pression was significantly associated with later depression. The researchers concluded that although a direct association between postpartum depression and child problem be haviors was not dete cted, postpartum depression may pose a risk to child behavior problems given it increases a woman’s risk for later non-postpartum depressi on. It is also possible that maternal depression at age four and a half was found to be associated with both internaliz ing and externalizing problems due to biased maternal report s in the context of current depression. One last study to be reviewed investigated the impact of maternal depression at 18 months and five years postpartum asse ssed via the CES-D and internalizing and externalizing problems at age seven measur ed using the CBCL/4-18 among a small highrisk sample (Lyons-Ruth et al., 1997). Mothers and teachers both completed the CBCL/4-18 to reduce reporting bias among mothers concurrently experiencing depression; and analyses we re conducted to predict both continuous scores on the CBCL/4-18 and dichotomous scores of clinical significance. Women were classified as depressed if they had experienced depressi on at either 18 months or five years. Regression results controlling for a cumula tive demographic risk variable, infant sex, maternal psychosocial problems, and infa nt attachment at 18 months revealed a significant association between maternal depres sion at 18 months or five years and both internalizing and externalizi ng child behavior problems at age seven, as reported by the mother (both continuous values and dichotom ous scores of clinical significance). Maternal depression at 18 m onths or five years was also found to be significantly associated with teacher reports of internal izing and externalizing behaviors (continuous scores) but only predictive of internalizi ng clinical scores. Given that maternal

PAGE 106

84 depression at 18 months and five years wa s predictive of both ma ternal and teacher reports of problem behaviors at age seven pr ovides strong evidence that an association is likely to exist since the influence of matern al reporting bias found in many studies was removed. The results also suggest that despite an associatio n between maternal depression and continuous problem behavior scores, maternal reports of externalizing behaviors reaching clinical significance may be an overestimate, given maternal depression did not significantly predict te acher reports of exte rnalizing behaviors reaching clinical significance. It has alrea dy been reviewed that depressed mothers tend to rate their children higher in te rms of externalizing behaviors. Summary. In summary, there is strong empiri cal support for relationships between postpartum depression in general, and chronici ty of depressive symptoms in particular, and display of problem behavi ors in children. Only a few studies have failed to find a significant association, two of which employed bivariate an alytic techniques, and one of which the long-term assessment of problem behaviors was made at 18 months postpartum, possibly too early to detect si gnificant differences in child behaviors. There is some evidence for relationships between symptom severity and timing of onset and later behavioral di fficulties in children, although studies investigating these associations are far fewer. Only one study explored the impact of symptom severity revealing a positive association between the worst episode of maternal depression during the five-year follow-up period and problem behaviors in children at age five. Two studies were reviewed that explored the impact of timing of symptom onset on display of problem behaviors. One ran se parate logistic regression models for women

PAGE 107

85 experiencing depression in 1988 only, 1991 only, and 1988 and 1991. Results found that all three classifications of postpartum depr ession significantly pred icted child problem behaviors at 1991, with increasi ng odds ratio values across categories (women depressed in 1991 only had higher odds ratio values th an women depressed in 1988). Although on the surface findings suggest that later onset de pression may pose a greater risk to child problem behaviors than earlier onset de pression, this finding is more likely a methodological artifact of obtaining maternal reports of problem behaviors in 1991, in which depression was also assessed. The second study found contrasting results making it difficult to draw clear conclusions regarding the relationship be tween timing of onset and child problem behaviors. In this study, the impact of onset was only explored among women experiencing a single elevated rating of depression during the study period. Research findings revealed that problem behaviors were only higher for children whose mothers reported being moderately depressed at six m onths or five years, compared to mothers who reported being moderately depressed during pregnancy or immediately postpartum. Furthermore, problem behaviors were only higher for children whos e mothers reported being severely depressed at five years, comp ared to mothers who reported being severely depressed during pregnancy, immediately post partum, or six months postpartum. Once again, significant associations with depression onset at five years (for both moderately and severely depressed mothers) are likely related to maternal bias in reporting of child behaviors in the context of pos tpartum depression rather than a true association between later onset depression and child problem be haviors. The finding of an association between depression onset at six months for m oderately depressed mothers is evidence of

PAGE 108

86 a “sensitive” period of onset, however this time point was only compared to pregnancy, immediately postpartum, and five years postpar tum. Further research with more finite assessments of depression is needed to more clearly explicit critical periods of development in which symptom onset incr eases risk for adverse child outcomes. Overall there is strong em pirical support for a rela tionship between postpartum depression and child problem behaviors, with the potential of an increased risk for internalizing problem behaviors as compared to externalizing problem behaviors. This latter point is still under debate however due to mixed findings. Although many studies find an association between postpartum depr ession and internalizing problems, only a few studies have also documented a positive association with externalizing problems. The contributions of these studies are lim ited due to methodological flaws. One study concurrently assessed postpartum depressi on along with maternal reports of child problem behaviors making the assessment vul nerable to reporting bias. Mothers' experiences with depression upwardly bias th eir ratings of externalizing behaviors. Another study found that maternal reports in dicated a relationship between depression and externalizing problems reaching clinical si gnificance, but teacher reports did not. This finding suggests that perhaps maternal reports are inflated in the context of depression and must be interpreted in light of other individuals’ rati ngs of child problem behaviors. Nevertheless, most studies utilized a total pr oblem behavior score as the major dependent variable (including both internalizing and externalizing problem behaviors) and found positive associations. There is also room for debate regard ing differential vulnerability of boys and girls to exposure of postpartum depressi on. Although boys are generally accepted as

PAGE 109

87 being at higher risk for adverse outcomes than girls in the context of postpartum depression, one study did find a significant in teraction effect between exposure to postpartum depression and child sex, with daughters of depressed mothers displaying higher levels of externalizi ng problems compared to daught ers of nondepressed mothers. These rates were however similar to rates of externalizing problems reported for all boys, despite their exposure to postpartum depression. Another study did not find an increased risk for girls, but instead for boys. E xposure to depression during the first year postpartum predicted internaliz ing problems among boys only. As mentioned above, maternal report s of problem behaviors simultaneously with postpartum depression may inflate rate s of adverse outcomes in children in the context of postpartum depression. When expl oring the impact of exposure to postpartum depression at 18 months or five years on problem behaviors reaching clinical significance, maternal reports predicted both internalizing and externalizing problems, whereas teacher reports only predicted internal izing problems. Mother s with a history of depression tended to overreport the prevalence of problem behaviors in their children. Another study exploring the impact of tim ing of depression onset on total problem behaviors revealed significant findings onl y among children of mothers who were severely depressed at five y ears, the same time point in which maternal reports of problem behaviors were ascertained. This finding was likely a me thodological artifact given that maternal depression and child prob lem behaviors were measured concurrently. More studies also investigated the ro le of mediating and moderating effects on the relationship between postpartum depressi on and display of problem behaviors in children as compared to those investigati ng the impact of postpartum depression on child

PAGE 110

88 social competence. Such investigations were far from uniform however necessitating future research to more thoroughly explore th ese influences to be tter direct prevention efforts. Nevertheless, empirical evidence from this review suggests that contextual risk factors such as marital adjustment, social s upport, and exposure to li fe stresses partially mediate the impact of postpartum depressi on on child problem behaviors. Attachment security and maternal sensitivity did not a ppear to mediate the relationship, nor was maternal sensitivity found to moderate the impact of postpartum depression on child problem behaviors. Limitations of Current Research One significant limitation in the current literature investigating the influence of postpartum depression on social competence and behavior problems is the wide variation in operationalization of matern al depression. For example, some studies use indices of self-report symptom checklists to measure pos tpartum depression, whereas others utilize more stringent diagnostic criter ia. Self-report assessment tool s have been criticized for over-diagnosing “depression” while failing to identify some women who meet formal diagnostic criteria (Campbell & Cohn, 1991; Gotlib, Whiffen, Mount, Milne, & Cordy, 1989); and have also been associated with nondepressive psychiatric disturbances such as anxiety (Garrison & Earls, 1986). Appli cation of diagnostic criteria to define postpartum depression on other hand is time and labor intensive gi ven it requires a oneon-one in-depth interview by a trained prof essional to assess symptomatology and provide a diagnosis. This may not be f easible in large community samples. Even comparisons among studies utilizing th e same type of assessment tool for depression (self-report symptom checklist or diagnostic criteria) are complex given the

PAGE 111

89 wide variety of measurement instruments and interviews found in the literature. Examples of self-report measures used to assess postpartum depressi on in the literature investigating the impact of postpartum depression on child development have included the EPDS, the CES-D, the BDI, and the D SSI. Further, studie s employing diagnostic criteria to classify women as depressed in the postnatal period ha ve utilized several interviews including the SCID, the SPI, and the SADS. These interviews have been used in conjunction with either the DSM criteria or the RDC criteria for diagnosis. Furthermore, the time period in which researchers consider to represent postpartum depression greatly varies across studi es. For example, some studies strictly use symptoms during the first six to 12 m onths following delivery, whereas others include episodes of depression throughout the child’s life to represent risks associated with postpartum depression. In addition, quite a few studies employed lifetime indices of depression including maternal episodes of depression prior to pregnancy and birth. Additional methodological limitations of existing literature investigating the impact of postpartum depression on child social competence and problem behaviors include wide variation in study designs, samp le populations, and inclusion of contextual risk factors to control for confounding effects. Study designs employed in investigations have included both cross-sectional and longitudinal designs. Although longitudinal designs are preferable given the desire to establish causality, thes e studies have often employed retrospective inquiry of depressi ve symptoms given long intervals between assessment periods requiring valid recall of symptomatology. Furthermore, a wide variety of samples have been researched including clinical, co mmunity, and high-risk multi-problem samples. It is critical to review and digest research findings in light of the

PAGE 112

90 sample population given that not all studies uniformly control for confounding effects of psychosocial stressors and demographic charact eristics that may place some children at increased risk of experiencing negative de velopmental outcomes. In fact not many studies have included contextual risk factors during data anal ysis to conserve statistical power given small sample sizes. Perhaps the greatest limitation in th e current literature exploring the relationship between postpartum depression and child outcomes is the failure to fully consider the heterogeneity of women’s e xperiences with depr ession. Although more studies have investigated de pression characteristics such as chronicity, severity of symptoms, and timing of symptom onset on be havioral difficulties and child social competence, these studies are quite limited due to unsophisticated analytic techniques employed. Although noteworthy that a considerab le number of studies have investigated distinct features of postpartum depression on adverse outcomes in children, these studies are still restrictive in terms of elucidating cl ear mechanisms of infl uence given they often truncate longitudinal data into a single vari able for analysis purposes and neglect to consider the experience of depression along a ch ild’s developmental trajectory, whether it be chronicity, severity, or timing of onset. For example, many studies constructed a chronicity variable by classifying women into categories based on the number of episodes of depression they had experienced during the study period (e.g., never depressed, sometimes depressed, chronically depressed). This type of classification not only disregards the potential impact of sequential episodes of chronic de pression (women experi encing two episodes of depression two years apart may be classifi ed the same as women experiencing back to

PAGE 113

91 back episodes across assessment periods); but it also ignores the experience of chronicity along the child’s developmental trajectory (a classi fication of chronica lly depressed at 36 months is likely to have a more severe imp act on child outcome than a classification of chronically depressed at six months). Another example is the single study that investigated the im pact of postpartum depression severity on child outcome. The st udy explored the infl uence of depression severity on child problem behaviors at ag e five revealing a significant association between the worst episode of de pression experienced from pregnancy to five years of age and total problem behaviors. Despite the availability of data on symptom severity at multiple time points within the five-year follow-up period, the highest symptom level regardless of time in which the mother experienced it, was used to predict the child’s behavior at age five. This operationalizati on of severity is limited given it does not consider the differential impact severity of postpartum depression symptoms may have at various time points along the chil d’s developmental trajectory. For example, it is likely that a child exposed to severe symptoms ear lier in the postpartum period (one, six, or 12 months) would have increased risk for behavi oral difficulties than a child exposed to severe symptoms later in the postpar tum period (15, 24, or 36 months). Similarly, studies exploring the impact of timing of symptom onset have been limited largely due to the potential for maternal reporting bias in the context of maternal depression, and the relatively few time point s in which depression was measured. For example, one study found that women who experienced depression in 1991 only had higher odds ratios for children with problem behaviors than women who had experienced depression in 1988 only. It appeared that late r onset depression posed the greatest risk to

PAGE 114

92 child problem behaviors, however behavioral difficulties were assessed via maternal reports at the same home visit as maternal depression. It is cr itical for future research to control for or remove from analysis women who are concurrently experiencing depression while providing a ma ternal report on some aspect of development to reduce the potential for reporting bias in the context of maternal depression. Another study found significantly higher levels of problem behaviors among children of mothers who had experienced depressi on at either six months or five years as compared to during pregnancy or immediat ely postpartum. Alt hough interesting, this finding does not elucidate much regardi ng sensitive periods along the child’s developmental trajectory in which onset of depressive sy mptoms may pose the greatest risk due to sporadic assessmen ts of postpartum depression with lengthy intervals during the postnatal period. One last limitation of current research is the relatively lit tle exploration of mediating and moderating influences between the impact of postpartum depression on child social competence and display of pr oblem behaviors. To fully develop the prevention implications of promoting child development in the context of postpartum depression, it is critical to understand the indi vidual, social and envi ronmental factors that influence the likelihood that exposure will lead to negative outcomes; as well as understanding the proximal consequences of ex posure that result in increased risk for negative outcome. Conclusions In conclusion, postpartum depression has been accepted as having serious implications for the mother’s ability to func tion effectively in her role as mother and

PAGE 115

93 caregiver posing a threat to the developing child. Despite a considerable amount of research exploring the effects of postpartu m depression on child social competence and display of problem behaviors including so cial competence and display of problem behaviors, much of the literature has faile d to fully consider the heterogeneity of women’s experiences with postpartum depre ssion and how the nature and course of depressive episodes may differentially impact the likelihood of negative outcomes in children. Even among longitudinal studies that have investigated more finite features of depression including chronicit y, severity, and timing of sy mptom onset, the ways in which variables were operationalized and treated during analysis greatly undermined their ability to eluc idate causal mechanisms of risk along the child’s developmental trajectory. It is also critical for future re search endeavors to mo re thoroughly test the influences of mediating and moderating factor s to inform prevention efforts directed to promote the healthy development of childre n in the context of maternal depression. Mechanisms of Influence As stated above, future research explor ing the impact of postpartum depression on child social competence and display of probl em behaviors needs to more thoroughly test the influences of mediating and moderating f actors to inform prevention implications of social development in the context of maternal depression. Using attachment theory as a conceptual framework, it is hypothesized that postpartum depression impacts long-term child social competence and display of probl em behaviors through a decreased ability of the mother to recognize, inte rpret, and respond to infant signals. Decreased maternal sensitivity in turn is expected to impact the child’s internal working model of self, other, and the relationship between self and other, influencing expecta tions and affective

PAGE 116

94 responses in future interpersonal relati onships, thus impacting long-term social development. Exposure to nonmaternal care is hypothesized to mode rate the relationship between maternal depression and child so cial competence and display of problem behaviors by influencing the formation and/or translation of internal working models in future interpersonal relationships. Figure A below presents the conceptual framework for testing the effect of matern al depression on child social competence and display of problem behaviors. The empirical eviden ce for each of these pathways will follow.

PAGE 117

95 Figure A Conceptual Framework for Testing the Effect of Maternal Depression on Child Social Competence and Display of Problem Be haviors with Mediating and Moderating Influences Effects of Postpartum Depre ssion on Maternal Sensitivity As previously discussed, maternal sensitivity refers to maternal behaviors that are contingent on infant behavi ors, timely, and appropriate (A insworth et al., 1974; Maccoby & Martin, 1983; Smither & Pederson, 1988). Ainsworth described the sensitive mother as one who meshes her initiations with her in fant’s actions and behaviors, capitalizing on the infant’s sense of timing and interest s to achieve a cooperative and harmonious interaction (Ainsworth, 1969). A sensitive mother is capable of perceiving things from the child’s point of view, regards the child as a separate person, and respects the child’s Maternal Depression Exposure Timing of Initial Onset Severity Chronicity Child Social Competence and Display of Problem Behaviors at First Grade Moderating Variables Maternal Sensitivity Exposure to Nonmaternal Care Maternal Sensitivity

PAGE 118

96 activities and thus avoids interruption (Ainswor th, Bell, & Stayton, 1971). Further, sensitive mothers must also learn to adapt to their infant’s behavioral changes and signals of autonomy as the child transitions to t oddlerhood (LeCuyer-Maus, 2000). As empirical evidence suggests, a mother experiencing depression may find it extremely difficult to parent with high levels of se nsitivity (Klaus et al., 1995). In fact depression in general, for both men and women, is associated with a range of difficulties with interpersonal communi cation (Murray et al., 1999). Depressed mothers are thought to intera ct less sensitively with their infants because they are preoccupied with their own negative cogniti ons and feelings (Pyszczynski & Greenberg, 1987), interfering with their ab ility to notice and respond con tingently to infant cues. Depressed mothers, particularly those also f aced with social adversity, have been shown to be more withdrawn, inconsistent, intr usive, hostile, under stimulating, and over stimulating; comprising their ability to provide responsive, sensitive, and nurturing environments for their infants to optimally grow and develop (Cummings & Cicchetti, 1990; Field, Healy, Goldstein, & Guther tz, 1990; Lyons-Ruth, Zoll, Connell, & Grunebaum, 1986). Cox, Puckering, Pound, and Mills (Cox, Puckering, Pounds, & Mills, 1987) found that depressed mothers of two-year-old children were less responsive to the child’s overtures and cues and more likely to respond with control when the child was distressed. Other studies have shown that depression influenced a mother’s ability to accurately interpret the needs of her infant (Main, Kaplan, & Cassidy, 1985; ZahnWaxler, Iannotti, Cummings, & Denham, 1990). Mothers unable to appropriately read infants’ cues were more likely to try to soci alize with their infants when the infant was hungry, play with the infant when the child wa s tired, and feed the child when he or she

PAGE 119

97 was attempting to initiate social interac tion (Ainsworth et al., 1974). Incongruent interactions such as these convey to the chil d that his or her attachment behaviors are functionally ineffective impacting the ch ild’s appraisal pro cesses underlying the attachment behavioral system, which will eventually become internalized as mental representations guiding future so cial interactions. Further, depressed mothers and their toddlers have been found to exhibit less mutu ally positive affect, more mutually negative affect, as well as less interactive coordi nation as compared to nondepressed mothertoddler dyads (Jameson et al., 1997 ; Radke-Yarrow et al., 1985). Chronicity of depressive symptoms has b een documented to particularly impair a mother’s ability to nurture her child in a sensitive and responsive manner. Campbell, Cohn, and Meyers (Campbell, Cohn, & Meye rs, 1995) found that women who were chronically depressed from the postpartum pe riod through six months were less sensitive and engaged during feeding and play episodes at four months, and less positive with their infants during face-to-face inte ractions at two, four, and five months; as compared to mothers whose depression had remitted by six months. Another study found that chronically depressed mothers were less sensitiv e and available to their infants, as well as more negative during play than women who had reported intermittent depression (Frankel & Harmon, 1996). Yet still the NI CHD Early Child Care Research Network, utilizing the same data set as the present dissertation, f ound that women who reported no symptoms of depression during the first three year’s of the child’s life were the most sensitive, followed next by mothers reporti ng intermittent depression, with chronically depressed mothers having the lowest levels of maternal sensitivity (NICHD Early Child Care Research Network, 1999b). When comparing changes in maternal sensitivity from

PAGE 120

98 15 to 24 months, the researchers also found that the chronically depressed group of women was the only group to show a systema tic decline, suggesting that developmental changes in children’s behavior at 24 months (increased mobili ty, verbal ability, bids for autonomy) may be particularly challenging fo r mothers experiencing chronic depression. Research exploring moderators of the a ssociation between postpartum depression and maternal sensitivity revealed significan t effects of maternal perceived acceptance by each parent during childhood, as well as spous al aggression. Among depressed mothers, those who reported high levels of percei ved acceptance by her own parents during childhood, as well as those who reported low levels of spousal aggression, were less likely to display insensitive maternal behavi ors as compared to depressed mothers who reported high levels of perceived parental rejection during chil dhood and high levels of spousal aggression (Crockenberg & Leerkes, 2003). In conclu sion, strong empirical evidence supports a relationship between mothers’ experiences with postpartum depression and decreased levels of maternal sensitivity. Effect of Maternal Sensitivity on Child Development Most studies exploring the influen ce of maternal sensitivity on child development consider its impact on child attach ment security to the primary caregiver, as measured by the Strange Situation procedure. As previously presented, Ainsworth and her colleagues were the first to examine such an association in her seminal study conducted with 26 mother-infant dyads from Baltimore, concluding that sensitive responsiveness to infant signa ls and communications was the most important aspect of maternal behavior associated with attachme nt security (Ainsworth, Blehar, Waters, & Wall, 1978b). Critics argued that Ainsworth ove r generalized the resu lts from her small

PAGE 121

99 sample (Gerweitz & Boyd, 1977; Lam b, Thompson, Gardner, & Charnov, 1985b) sparking controversy in the fi eld. In a meta-analytic study attempting to sort through issues of association, Goldsmith and Alansky (Goldsmith & Alansky, 1987) concluded that many studies did in fact replicate th e predictive power of maternal sensitivity, however noted that the actual size of the effect was much smaller than once believed. Arguments ensued as many researchers c ontinued to support Bo wlby’s theoretical emphasis and Ainsworth’s empirical findings on maternal sensitivity as a crucial antecedent of attachment security clai ming methodological weaknesses of studies included in the meta-analysis as the cu lprit behind a low effect size. A second meta-analytic review (DeWolf f & van Ijzendoorn, 1997) including 66 studies exploring the relations hip between maternal sensitiv ity and attachment security concluded that a mother’s ability to respond ap propriately and promptly to infant signals was in fact a moderate predic tor of attachment security. Although the researchers noted that sensitivity could not be considered the exclusive and most important factor in the development of attachment, they did empha size a direct and fairly strong causal relationship. Other parental behaviors found to be related to attachment security were emotional warmth and support, positive regard for the child, as well as synchrony and mutuality in the interaction. They also determined that de spite initial expectations, the duration of home observation in ascertaining maternal sensi tivity (long or short) did not impact the magnitude of the association, nor did setting of observation (home versus laboratory). Other studies published afte r the aforementioned second meta-analytic review on the association between maternal sensitivity and attachment security in toddlers have

PAGE 122

100 found a positive, direct relationship (Koren -Karie, Oppenheim, Dolev, Sher, & EtzionCarasso, 2002; Pederson, Gleason, Moran, & Bento, 1998). The deba te regarding its importance as an antecedent of attachment security continues however as no study has been able to replicate the magnitude of ef fect originally reported by Ainsworth. In response, Ainsworth and other proponents of the cr itical role of matern al sensitivity in the development of attachment security and othe r developmental outcomes in children assert that the difficulty providing empirical conf irmation may be related to the challenge involved in employing a valid a nd comprehensive description of mother-child interaction, rather than indicating the theo retical unimportance of the co nstruct (Ainsworth & Marvin, 1995; Pederson & Moran, 1995). In summary, it is fairly accepted that maternal sensitivity including emotional warmth, re sponsiveness to stress, moderate and appropriate stimulation, as well as interactio nal synchrony is associated with attachment security at 12 to 15 months in a normal setting (Belsky, Rosenberger, & Crnic, 1995). Another subset of research exploring the impact of maternal sensitivity on child development investigates its association with cognitive development including language skills and social competence. One study f ound positive associations between maternal sensitivity at six, 13-, and 24-months a nd child IQ and language competence at 24 months (Olson, Bates, & Bayles, 1984). Mate rnal sensitivity during infancy was also found to be related to child vocabulary at 12 months of age (Ruddy & Bornstein, 1982), as well as child IQ at age six (Coates & Lewi s, 1984). In a more recent study, maternal sensitivity averaged over a six to 36-month pe riod was found to be a powerful predictor of child school readiness a nd verbal comprehension scores at age three (NICHD Early Child Care Research Network, 1999b). It is suggested that sensit ive mothers stimulate

PAGE 123

101 their children more often than insensitive moth ers, and engage in more verbal interaction promoting cognitive and language development. Maternal sensitivity during infancy has also been associated with increased social competence of toddlers (Clarke-Stewart, Va nderstoep, & Killian, 1979) These children have been shown to be more sociable, self-conf ident, curious, and more able to tolerate frustration (Matas, Arend, & Sroufe, 1979; Vandell, Owen, Wilson, & Henderson, 1988). Children of insensitive parents display more problematic behaviors and have lower levels of reported cooperation (Kopp, 1992; NICHD Early Child Care Research Network, 1999b). In addition to having a direct effect on ch ild development, maternal sensitivity has also been shown to mediat e the impact of postpartum depression on cognitive and language development (NICHD Early Child Care Research Network, 1999b). Conclusions from such research deduce that low levels of sensitivity demonstrated by depressed women may be more directly re sponsible for poorer cognitive and language development in children than the presence of depressed mood alone. Maternal sensitivity was also shown to moderate or buffer the re lationship between ma ternal depression and child verbal comprehension at age three. Children whose mother reported being depressed some of the time performed bett er on expressive language when they had a more sensitive mother, as compared to those with a less sensitive mother. The potential mediating and moderating influences of mate rnal sensitivity on the association between maternal depression and child social compet ence and display of problem behaviors at grade one is important to el ucidate to guide future inte rventions targeting depressed mothers.

PAGE 124

102 Further, there is a substan tial body of empirical literature exploring the impact of attachment security on long-term outcomes in cluding social and cognitive development. The research in this area will be reviewed in this section given maternal sensitivity is a positive predictor of attachment security, whic h in turn is thought to be the earliest assessment of a child’s internal working model of self and other within the social world, which is hypothesized to be the mechanism of influence linking maternal sensitivity and long term child development. Attachment security has been linked with cognitive and social development in children. Ainsworth and Bell (Ainsworth & Bell, 1970; Ainsworth & Bell, 1974) found a signification correlation between child attach ment security and intelligence as measured by intelligence quotient (IQ) scores among 11-month-old infants. Although Matas (Matas et al., 1979) documented a higher average IQ at 23 months among children classified as secure at 18 m onths, as compared to those classified as insecure, the difference was not statistically significant (Matas et al., 1979). In a study of two and three year old children, child attachment s ecurity was positively associated with peer responsiveness and reciprocal interaction, and negatively associated with unsociable behaviors including crying, physic al aggression, and verbal th reats (Lieberman, 1977). Attachment security was also found to be rela ted to display of problem behaviors at age four and five with more children classified as insecurely attached at 12 and 18 moths of age displaying problem behaviors (Erickson, Sroufe, & Egeland, 1985).

PAGE 125

103 Moderating Influence of Expos ure to Nonmaternal Care Women in the Workforce. The U.S. Bureau of the Census (U.S. Bu reau of the Census, 1992) reported that by the early 1990’s, more than one half of moth ers of infants under the age of one were in the labor force. As the number of two income families continues to increase, along with the changing roles of women and men in family life, more children in the United States are being cared for by individuals outside the home. It has been estimated that between 1975 and 1990, the number of licensed childcare centers grew 300%, and the number of children in nonmaternal childcare in creased fourfold (Fellmeth, 1999). A number of factors have been associated with the increase in number of children in nonmaternal care including welfare policie s, financial constraints, and for some families a perceived need to enhance the edu cational environment fo r their child (Bailey, 2000). Perhaps the greatest influence on the ri se in number of children in nonmaternal care during the 1990’s, particularly among low income children, was the enactment of the Personal Responsibility and Work Opportuni ty Reconciliation Ac t, also known as welfare to work, in 1996 (Collins & Lundgren -Gaveras, 1999). Welfare to work legislation increased employment requirement s for welfare recipients including mothers of young children creating a vast need for mo thers to find childcare arrangements while maintaining employment (Collins & Lundgr en-Gaveras, 1999; Handler, 2004; Handler, Zimbeck, Chavkin, & Adams, 2003; Howes, 2000). Between 1996 and 1998, it has been estimated that an additional 741,000 never-marri ed single mothers entered the labor force as a result of welfare reform (Lerman & Ratcliffe, 2001).

PAGE 126

104 Regardless of reason, exposure to nonm aternal care has become a normative experience for American children. By 1997, 79% of children under the age of three years regularly spent time in non-parental care, wi th 39% of these children in care for 35 or more hours per week (Cappizzano & Adam s, 2000). Among kindergarten students enrolled in 1998, 81% had previous exposure to nonmaternal care pr ior to school entry (West, Denton, & Germino-Hausken, 2000). The changing context of the early caregi ving environment as well as the increased number of infants in nonmaternal care have raised a number of concerns among parents, researchers, and policymakers about the potenti al risks of exposure to nonmaternal care to the social and emotional development of children, particularly the mother-child attachment relationship. Concern is part icularly aroused for mothers with limited financial resources, such as women affected by welfare reform policy, who are less likely than more affluent mothers to secure high quality childcare that is more adept at facilitating optimal social development (Howes, 2000). In this study however, I plan to explore the potential benefit of exposure to nonmaternal care, in the context of postpartum depression. The literature supporti ng a potential buffering influence will be summarized below. Hierarchy of Attachment. Perhaps the most misunderstood construct in attachment theory is that of monotropy (Lieberman & Zeanah, 1995), the prem ise that infants tend to have a primary attachment figure, usually the mother (Bowlby, 1969). Bowlby was influenced by the work of Lorenz who documented that young goslings become imprinted onto a single

PAGE 127

105 moving object during a sensitive period within th e first few days of life (Lorenz, 1952). Although Bowlby believed that human infants do in fact have a primary caretaker, he did not believe that infants only become attach ed to one person. Instead he asserted the routine attachment of infants to multiple indi viduals, in which an internal hierarchy, or order of preference, is establ ished. Within this hierarchy, a child may first turn to the preferred or primary caretaker upon activati on of the attachment behavioral system, relying on secondary and tertia ry attachment figures when necessary. The internal hierarchy of attachment figures explains why some children cease crying sooner when soothed by their mother as compared to their father, or prefer to be put to bed by one parent to the other. Such a preference is not indicative of greater love or affection. In fact, Lieberman and Zeanah (Lieberman & Z eanah, 1995) suggest that children may have several attachment hierarchies fulfilling distin ct needs. For example, a child may prefer to play with his or her father displaying a greater range of positive affect and exploratory behaviors, but still turn to the mother when feeling in need of comfort or protection. Indeed, several studies have investig ated the development of attachment relationships between children and alternat ive childcare providers all supporting the notion that attachment relationships in child care settings develop similarly to motherchild attachment relationships. In one study of center-based care, Ra ikes (Raikes, 1993) found that attachment security of toddlers increased as childr en spent more time with the providers. It was observed th at as time with the child increased, childcare providers became more sensitive and responsive to chil d cues. Further, Barnas and Cummings (Barnas & Cummings, 1997) compared children’ s attachment behaviors with long-term staff members, classified as being at the cente r for three or more months, with short-term

PAGE 128

106 staff members. Results showed that children directed more attachment behaviors to longterm staff as compared to short-term sta ff, and long-term staff providers were more successful in soothing distressed children than short-term staff providers. Further, research suggests that childre n with a prior history of relationship difficulties can form secure attachment relationships with new caregivers, given sensitive and responsive care. One study (Howes & Sega l, 1993) observed toddlers that had been removed from their homes due to maternal abuse or neglect and placed in high-quality shelter care. After two months nearly half (47%) of the children had developed secure attachment relationships with their caregive r, as assessed via the Attachment Q-Sort. Furthermore, children who had remained in the shelter for longer periods of time were found to be more secure than children with shorter stays. Another study (Howes & Ritchie, 1998) investigated changes in children’s attachment organization with teachers af ter being removed from community childcare environments due to problem behaviors, a nd placed in a therapeutic preschool. Child attachment security was assessed using the A ttachment Q-Sort two and six months after placement in the therapeutic school. Results documented that attachment security increased over time, suggesting that childre n reorganized their attachments to teachers following repeated positive interactions (Howes, 1999). In summary, Bowlby asserted the routine attachment of infants to multiple individuals. Although he was pr imarily referring to infant attachment to both mother and father figures, empirical support has documen ted that children expos ed to nonmaternal care develop attachment relationships with car etakers in much the same way as children develop attachment relationships to parents. As time with a given caregiver increases,

PAGE 129

107 children’s experiences of interacting with the caregiver become cognitively organized and internalized. Although internal working models of the mother-child attachment relationship, once established, might serve as a template to guide future social interactions and relationships with other in dividuals including alternative caregivers, children can form multiple attachment rela tionships with multiple caregivers, with varying levels of security (Zeanah, Boris, & Scheeringa, 1997). Children with prior relationship difficulties, when exposed to alte rnative settings with sensitive caregivers, appear to either reorganize their attachment representations when they encounter sensitive and responsive caregiv ers, or they construct inde pendent relationships based on new experiences with new providers. Thus, in the context of mate rnal depression where maternal sensitivity is expected to be low, exposure to nonmaternal care is hypothesized to buffer the negative impact of depression on child social competence and display of problem behaviors.

PAGE 130

108 Summary In this study, the following hypotheses will be tested. 1. Maternal depression is asso ciated with child social competence and display of problem behaviors at first grade. 1.1. Maternal depression, in general, during the first three years postpartum is negatively associated with child social competence and display of problem be haviors at first grade. 1.2. Timing of initial onset of mate rnal depression, during the first three years postpartum, is associat ed with child social competence and display of problem behaviors at first grade, with earlier onset episodes having the greatest negative impact. 1.3. Severity of maternal depression along the trajectory of initial onset is associated with child so cial competence and display of problem behaviors at first grade, with severity of symptoms having the greatest negative imp act with earlier onset episodes. 1.4. Chronicity of maternal depression is associated with child social competence and display of problem behaviors at first grade, with chronicity of symptoms having th e greatest negative impact at later time points. 2. The relationship between maternal depr ession and child social competence and display of problem be haviors at first grade is mediated by maternal sensitivity.

PAGE 131

109 3. The effect of maternal depression on child social competence and display of problem behaviors at first grade is st ronger among mothers who rate low in maternal sensitivity. 4. The effect of maternal depression on child social competence and display of problem behaviors at first grade is weaker among children exposed to nonmaternal care. The proposed study aims to advance the cu rrent knowledge base by addressing many of the limitations in the current literature in the field of postpartum depression and child social competence and display of problem be haviors by addressing the following points. 1. Consideration of women’s heterogene ity of depression experiences using statistical modeling that capit alizes on longitudinal data to clarify the impact of timing of initial onset, symptom severit y, and depression chronicity along the child’s developmental trajectory. 2. Statistical adjustment for a multitude of sociodemographic variables including child sex, child birth order, maternal ag e, maternal race, maternal ethnicity, maternal education, presence of husband or partner in the home, family receipt of public assistance (including food stamps, ai m to families with dependent children, or WIC), and maternal ra tes of breastfeeding at one month postpartum. 3. Exploration of the mediating and modera ting effect of maternal sensitivity. 4. Exploration of the moderating effect of exposure to nonmaternal care. 5. Exclusion of women reporting elevated le vels of depression at first grade to reduce potential for ma ternal reporting bias.

PAGE 132

110 Chapter Three: Methodology This chapter describes the methods that will be used to conduct this study. The chapter is divided into seven sections: (1) purpose of study; (2) research hypotheses; (3) overview of study design; (4) overview of NICHD SECC data set; (5) NICHD SECC population and sample; (6) NICHD SECC instru mentation and study variable selection; (7) NICHD SECC data collection methods; (8) da ta scoring; (9) data analysis; and (10) design limitations. Purpose of Study The purpose of this study is to invest igate the associations between maternal depression during the first three years pos tpartum and child social competence and display of problem behaviors at first grade. The impact of several characteristics of maternal depression will be examined including general exposure, timing of initial onset, severity of symptoms, and chronicity of sy mptoms. This study will also explore the mediating and moderating influences of mate rnal sensitivity, as well as the moderating influence of exposure to nonmaternal care. Research Hypotheses This study tests the following hypotheses: 1. Maternal depression is asso ciated with child social competence and display of problem behaviors at first grade.

PAGE 133

111 1.1. Maternal depression, in ge neral, during the first three years postpartum is negatively associated with child social competence and display of problem behaviors at first grade. 1.2. Timing of initial onset of maternal depression, during the first three years postpartum, is associated with chil d social competence and display of problem behaviors at first grade, wi th earlier onset episodes having the greatest negative impact. 1.3. Severity of maternal depression alo ng the trajectory of initial onset is associated with child social competence and display of problem behaviors at first grade, with severity of symp toms having the greatest negative impact with earlier onset episodes. 1.4. Chronicity of maternal depression is associated with child social competence and display of problem behaviors at first grade, with chronicity of symptoms having the gr eatest negative impact at later time points. 2. The relationship between maternal de pression and child social competence and display of problem be haviors at first grade is mediated by maternal sensitivity. 3. The effect of maternal depression on child social competence and display of problem behaviors at first grade is st ronger among mothers who rate low in maternal sensitivity.

PAGE 134

112 4. The effect of maternal depression on child social competence and display of problem behaviors at first grade is weaker among children exposed to nonmaternal care. Overview of Study Design This study is based upon secondary analysis of data from the National Institute of Child Health and Human Development St udy of Early Child Care (NICHD SECC), a multi-site, prospective, longitudinal study design ed to explore the effects of exposure to nonmaternal care on child developmental ou tcomes. During Phase I of the study, 1,364 families with full-term healthy newborns fr om ten sites across the nation were enrolled, of which 1,223 continued to participate in Phas e II. Data for this study will consist of infants and their families who participated in both Phase I and Phase II of the NICHD SECC, and had complete contextual data availabl e for all variables of interest. Further, mothers that report elevated symptoms of depr ession at first grade will be excluded from analyses to decrease potentia l maternal reporting bias of child social competence and display of problem behaviors. A combinati on of ordinary least square (OLS) regression and logistic regression models will be built to explore the impact of maternal depression on child social competence and display of problem behaviors at first grade. Overview of NICHD Study of Early Child Care The NICHD Study of Early Child Care (SECC) is a multi-site, prospective, three-phase longitudinal study designed to e xplore the effects of exposure to nonmaternal care during infancy, early childhood, and middle childhood on child developmental outcomes. The NICHD director initiated th e study in 1989 in response to the growing number of children receiving care from i ndividuals other than their mothers (NICHD

PAGE 135

113 Early Child Care Research Network, 1994). The primary goal of the study was to examine variations in early child care experi ences of infants and toddlers, and to study the effects of these variations on concurrent and long-term child development, including social, emotional, language, cognitive, and physical development. To accomplish study objectives, a multi-cente r research network was formed to design and implement a common research prot ocol for each participating site. Ten locations near a major university were chosen including Lit tle Rock, Arkansas (University of Arkansas), Irvine, California (University of California), Lawrence, Kansas (University of Kansas), Boston, Massachus etts (University of New Hampshire), Philadelphia, Pennsylvania (Temple Universi ty), Pittsburgh, Pennsylvania (Penn State University), Charlottesville, Virginia (Uni versity of Virginia), Morganton, North Carolina (Western Carolina Center), Seattl e, Washington (University of Washington), and Madison, Wisconsin (Univers ity of Wisconsin). The pr incipal investigators from each site represent a wide range of theoretica l and methodological perspectives related to childcare research, thus providing a rich base of knowledge and expertise. During Phase I of the study (1991-1995) a cohort of 1,364 children and their families were enrolled when the children were one month of age. This cohort of children was studied intensively until the children were 36 months in age. During Phase II of the study (1996-1999), 1,223 children and their families continued to par ticipate and were followed through first grade to investigate the impact of nonmaternal care on long-term child development. Phase III of the study is currently underway and will follow up with over 1,100children and families who participat ed in Phase I through sixth grade.

PAGE 136

114 The study design was guided using an ecological developmental framework taking into consideration a multitude of factors including family and home environments, childcare environments, and individual diffe rences among children. Multiple measures of a single construct are often included in the SECC given that many aspects of child behavior and development under review do not have a universal, well-accepted instrument of measurement. Further, multip le methods are employed to collect empirical data including trained observat ions, in-depth interviews, que stionnaires, and standardized testing. During Phase I, data were collected on an age-based chronological schedule at five major assessment periods including one month (enrollment), six months, 15 months, 24 months, and 36 months. One-month enrollmen t data were collected in a visit to the child’s home. At six months, data were coll ected in the child’s home and at the child’s primary child care arrangement. At 15, 24, a nd 36 months, data were collected in the child’s home, childcare arrangement, and in the laboratory. Data were collected via telephone interview at three-month intervals in be tween major data collection assessments. During Phase II, data were collected during three major assessment periods including 54 months, kindergart en, and first grade. At 54 months, data were collected in the child’s home, childcare arrangement, and in the laboratory. At kindergarten, questionnaires were sent to the child’s fam ily at home, and to the child’s teacher at school. At first grade, data were collected in the child’s home, at school, in the laboratory, and in the child’ s after-school care arrangement Phone interviews were conducted at 42, 46, 50, 60, and 66 months.

PAGE 137

115 In summary, the NICHD SECC study design was envisioned to be a benchmark in modern social science researc h. With a greater purpose of providing empirical data to formulate and inform social policy related to infant care, the SECC employed much effort to ensure study integrity including: (1) ad equate funding to allow the procurement of a large sample across multiple sites; (2) time for previously independent researchers to meld ideas into a cohesive unitary design; (3) adoption of an ecological developmental framework to guide the study; (4) careful traini ng of all data collect ors across sites; (5) use of significant incentives and family fr iendly assessment practices to minimize attrition; and (6) ensured support for a pe riod long enough to allow documentation of the interplay of contextual forces in helping to shape the development of children (NICHD Early Child Care Research Network, 1994). NICHD SECC Population and Sample Target Population and Sampling Frame The ten study sites selected to participate in the NICHD SECC were selected based on competing scientific merit rather than probabilistic sampling considerations (NICHD Early Child Care Research Networ k, 1991c). At each site, participating hospitals from which participants would be sampled were chosen based on factors such as location within a given city, previous working relationship with the principal investigator, nature of patient load, and avai lability. The sample was thus not intended to be nationally representative. Instead, the target population for generalization of findings is defined as the catchment of all babies (the index subjects) and their related families born in participating hospitals during the re cruitment phase of th e study, January through November 1991. The sampling frame for th e study was 8,986 women giving birth during

PAGE 138

116 selected 24-hour sampling periods, which were established on a site-by-site basis to ensure an unbiased and fully representative sample at each site (NICHD Early Child Care Research Network, 1991c). Sampling Plan The study employed a three-stage sampli ng plan through which a total of 1,364 families with full-term healthy newborns we re enrolled when the children were one month of age. During the first stage of r ecruitment, mothers were contacted in the hospital shortly after their infant’s birth and were screened for eligibility and invited to participate in the study, if elig ible. Participants were excl uded from the study if: (1) the mother was under the age of 18; (2) the mother was not proficient in English; (3) the family planned to move outside the catchment area within the year; (4) infants of multiple births, and those with medical complicati ons; (5) mothers with medical problems or acknowledged substance abuse, or who were placing their infants up for adoption; (6) mothers who lived a considerable distance from the site or who were enrolled in another study; and (7) mothers who lived in a ne ighborhood deemed by po lice too unsafe for home visitation; (NICHD Early Child Care Research Network, 2000). Another 1.5% refused to be interviewed in the hospital, and 3.4% of women requested not to be contacted after returning home from the hos pital. A total of 5,416 (60.3%) mothers met eligibility criteria and agreed to be contacted for furthe r participation in the study. This phase of recruitment provided the resear ch team insight into the demographic characteristics of the catchment populati on to guide sampling and recruitment during phase two.

PAGE 139

117 The second stage of recruitment occurre d during a phone interview conducted by site staff two weeks after infant delivery. Ca ll lists of families eligible at stage one were generated by a central data center, and fam ilies were contacted to recruit two to three families in the study per week at each site (NICHD Early Child Care Research Network, 2003). The data center monitored specific char acteristics of eligible families and made adjustments to increase the opportunity for ad equate representation of various groups (at least 10% single parent house holds; at least 10% mothers with less than a high school education; and at least 10% ethnic minority mothers (Research Triangle Institute, 1999a). The data center also ensured that recruited families included mothers who planned to work or go to school full time (60%) or part time (20%) in the child’s first year, as well as mothers who planned to stay at home with the child (20%). Furt her exclusion criteria applied during the second stage of recruitm ent included the elimination of infants who remained in the hospital for more than seve n days postpartum, families that planned to move outside the catchment area within the ne xt three years, families that could not be reached after three attempts, and those who refused to pa rticipate. Phone interviews were conducted with 3,015 (55.7%) families, of which 1,526 (50.6%) were eligible for the onemonth interview. During the third stage of recruitment, s ite staff attempted to contact the 1,526 families who remained eligible for study par ticipation after stage two of the sampling plan. Families were officially enrolled upon successful completion of all data collection through the one-month interview. The fi nal sample for the NICHD SECC was 1,364 families. A summary of the sampling process is provided as a flow chart in Figure B.

PAGE 140

118 Figure B The NICHD SECC Sampling Process Two-Week Phone Interview Not Conducted N=2401 (44.3%) One-Month Interview Not Scheduled N=1489 (49.4%) One-Month Interview Not Completed N=162 (10.6%) One-Month Interview Completed N=1364 (89.4%) One-Month Interview Scheduled N=1526 (50.6%) Two-Week Phone Interview Conducted N=3015 (55.7%) Eligible for Two-Week Phone Interview N=5416 (60.3%) Not Eligible for Two-Week Phone Interview N=3570 (39.7%) Hospital Screening N=8986 NICHD Sample: Phase I Based on hospital screening recruitment data, 53% of mothers enrolled during Phase I (n=1,364) planned to work or go to sc hool full time during the child’s first year, 23% planned to work or go to school part tim e, and 24% planned to stay home with the child (Appelbaum, 1993). Hospital screening recruitment data also revealed that participants represented a wi de range of socioeconomic and sociocultural backgrounds (11% low education, 24% ethnic-minorities as observed by the hospital interviewer); and both two-parent families (86%) and single-parent families (14%) were included. Recruited families did not differ significantly from other families eligible to participate in the study on major demographic characteris tics (maternal education, percentage in

PAGE 141

119 different ethnic groups, and pr esence of a husband or partne r in the home) ; except that mothers in the study had a 4% higher rate of intention to be empl oyed during the child’s first year as compared to nonparticipating mothers (NICHD Early Child Care Research Network, 2001). Although the sample was not de signed to be nationally representative, a comparison of the final sample with others living in the same census tracts and nationally revealed that the SECC sample was similar to both comparison groups related to household income and ethnicity (NICHD Earl y Child Care Research Network, 2000). Table 2 describes selected demographic char acteristics of the enrolled SECC sample based on hospital screen ing recruitment data.

PAGE 142

120 Table 2 Selected Characteristics of NICHD SECC Sample from Hospital Screening Recruitment Data Characteristic N Percent Educational level – Mother --<12th grade 152 11.1 High school/GED 328 24.0 Some college 389 28.5 BA level work 324 23.8 Postgraduate work 171 12.5 Ethnicity of mother --White 1035 75.9 Black 179 13.1 Hispanic 82 6.0 Other 68 5.0 Partner at home --No 194 14.2 Yes 1170 85.8 Mother’s plan for hours/week at work or school during child’s first year --<10 hours 329 24.1 10-29 hours 307 22.5 30+ hours 727 53.3 Don’t know 1 0.1 Note : Due to rounding, some percentages may not total 100. Limitations of Study Sample Results are only generalization to the population of infants and their related families born in participating hospitals with in the ten sites during 1991, given use of nonprobabilistic methods for sample determination. Results cannot be ge neralized to infants born to mothers under the age 18, those not flue nt in English, and those with medical conditions including substance abuse. Results also cannot be generali zed to infants that were part of a multiple birth, and those that experienced medical complications.

PAGE 143

121 Restrictions in the sample to families who resided in neighborhoods considered safe for home visitation, based on police reports, also may impact the generalizability of findings. Further, during Phase II of the study, 141 (10.6%) families were lost to attrition. Reasons for non-continuation in th e study included: child or family moved away (31.21%); child illness or death ( 2.84%); study takes too much time (17.73%); family missed two major assessments (20.57%); personal reasons (20.57%); and safety or other (7.1%). Comparison of family characterist ics of those that remained enrolled in the study during Phase II versus those that did not continue to participat e were assessed using chi-square tests of independence and one-w ay ANOVAs. Results revealed that nonwhite mothers, mothers with less education, mo thers with no husband or partner living at home at one-month postpartum, families on publ ic assistance, mothers not breastfeeding at one-month postpartum, and younger mothers we re more likely to drop out of the study. No differences were detected by maternal ethnicity, child sex, presence of maternal depression at one month, or maternal sensit ivity at six months postpartum. Table 3 summarizes findings.

PAGE 144

122 Table 3 Comparison of Phase II Participants Versus Drop Outs Chi-Square Tests of Independence % Drop Out ChiSquare(df) Pattern of Finding Maternal Race White 9.23 Black/Afro-American 16.67 Other 12.70 9.40(2)* Non-white mothers were more likely to drop out of the study. Maternal Education High School or Less 13.85 13.93(3)* Some College 10.99 Bachelor’s Degree 6.69 Postgraduate Work 6.06 Mothers will less education were more likely to drop out of the study. Partner/Husband at Home No 15.66 Yes 9.43 7.07(1)* Mothers with no husband or partner living at home at one month were more likely to drop out of the study. Receipt of Public Assistance No 8.31 Yes 19.0 26.03(1)* Families on public assistance were more likely to drop out of the study. Maternal Ethnicity Non-Hispanic 10.05 2.53(1) Not Significant Hispanic 16.39 Child Sex Male 10.21 Female 10.46 0.02(1) Not Significant Breastfeeding at One Month No 15.50 29.66(1)* Yes 6.44 Mothers not breastfeeding at one month were more likely to drop out of the study. Depressed at One Month No 9.86 0.72(1) Not Significant Yes 11.46 Continued on the next page

PAGE 145

123 Table 3 (Continued) One-Way ANOVAs Mean (SD) F (df) Pattern of Finding Maternal Age Drop Out 26.61 (5.63) 11.23(1)* Non-Drop Out 28.28 (5.61) Younger mothers were more likely to drop out of the study. Maternal Sensitivity at Six Months Drop Out 3.10 (0.61) 0.15(1) Non-Drop Out 3.07 (0.59) Not Significant Indicates significance at p<.05. The study sample will further be limited. Only infants and their families that had complete contextual data on all variables of interest, and those whose mothers did not have elevated depression scores at grade one will be included in analyses. Infants and families with incomplete contextual data, or with mothers with elevated depression scores at grade one, may be stat istically different. NICHD SECC Instrumentation and Study Variable Selection The NICHD SECC is a comprehensive study implementing 67 instruments in Phase I and 78 instruments during Phase II for data collection purposes. Only instrumentation and variables pertinent to th is study will be reviewed below, organized by construct within the conceptual framew ork for testing the e ffect of maternal depression on child social competence and disp lay of problem behaviors at first grade. Refer to Figure A, page 100, for a revi ew of the study’s conceptual framework. Sociodemographic Control Variables Sociodemographic data were assessed using the One Month Home Interview (NICHD Early Child Care Research Networ k, 1991a), a standardized interview guide

PAGE 146

124 developed for the NICHD SECC. Data collect ed included information on the family and child including child sex, birth order, mo ther’s age and ethnicity, prevalence of breastfeeding, and informati on on the child’s father. Sociodemographic variables selected to statistically cont rol for potential confounding infl uences for this study include child sex, child birth order, maternal age, maternal race, maternal ethnicity, maternal education, breastfeeding at one month, presence of husband or partner in the home, and whether or not the family is on public assistance (including food stamps, aid to families with dependent children, or WIC). The sociodemographic control variables to be included in this study, along with respons e options are listed in Table 4.

PAGE 147

125 Table 4 Sociodemographic Control Variables Selected for Study Inclusion Variable Data Source Item(s) Response Coding Child Sex One Month Home Interview What is the sex of your baby? 1=Male 2=Female Child Birth Order One Month Home Interview Is BABY your first child? How many other children have you had? Continuous variable Maternal Age One Month Home Interview Let me verify your age. My records show you’re __. Is that right? Continuous variable Maternal Race One Month Home Interview How would you describe yourself? 1=White 2=Black or Afro-American 3=Other 1=American Indian, Eskimo, Aleut 2=Asian or Pacific Islander 5=Other (Specify) Maternal Ethnicity One Month Home Interview How would you describe yourself? Are you of Hispanic origin? 0=No 1=Yes Continued on the next page Note: For purposes of this study, the “oth er” maternal race category has been modified from its original form to include: (1) American Indian, Eskimo, Aleut; (2) Asi an or Pacific Islander; and (3) Other.

PAGE 148

126 Table 4 (Continued) Variable Data Source Item(s) Response Coding Maternal Education One Month Home Interview How much school have you completed? 1=High school graduate/GED or less 2=Some college but no degree, AA degree or vocational school 3=Bachelor’s degree from college of university 4=Graduate level work Breastfeeding One Month Home Interview Are you breastfeeding BABY now? 1=Yes 0=No Presence of husband/partner in the home One Month Home Interview Can you tell me who lives here in your home with you, their relationship to you, and their ages? 0=No 1=Yes Public assistance One Month Home Interview Is your family receiving income from any of the sources listed below (food stamps, AFDC, or WIC? 0=No public assistance 1=Public assistance Note: For purposes of this study, the “g raduate level work” maternal education categor y has been modified from its original fo rm to include: (1) some graduate work or a master’s degree; (2) law degree; and (3) more than one master’s degree or a doctoral degre e (M.D., Ph.D., Ed.D., etc.).

PAGE 149

127 Maternal Depression Maternal depression is the major construc t of interest in the study. In addition to exploring the impact of exposure to mate rnal depression, in general, on child social competence and display of problem behaviors, characteristics of depressive episodes will also be considered. Specifically, the impact of timing of initial onset of depression, the severity of symptoms, and the chronicity of symptoms, will be explored in relation to child social competence and display of problem behaviors at first grade. Maternal depression was assessed in the NICHD SECC using the Center for Epidemiologic Studies Depression Scale (C ES-D) (Radloff, 1977), a well-validated 20item self-report scale intended to measure symptoms of depression in non-clinical populations. Respondents are aske d to circle the answer that comes closest to describing how often they have felt a particular way dur ing the past week. Sixteen of the items express negative experiences and four items express positive experiences to avoid tendencies toward response set (see Table 5 fo r a complete listing of CES-D items). The CES-D is not designed to diagnose depressi on but instead to asse ss the prevalence of depressive symptomatology in the general population.

PAGE 150

128 Table 5 Items Comprising the CES-D Scale Item I was bothered by things that usually don’t bother me. I felt that everything I did was an effort. I felt that I was just as good as other people. I had trouble keeping my mind on what I was doing. I felt sad. I felt fearful. I felt lonely. I had crying spells. I talked less than usual. My sleep was restless. I enjoyed life. I felt that I could not shake off the blues even with the help from my family/ friends. I thought my life had been a failure. I was happy. I could not get “going.” I felt hopeful about the future. People were unfriendly to me. I did not feel like eating; my appetite was poor. I felt depressed. I felt that people dislike me. Response options include 0=rarely or none of the time (less than once a week); 1=some or a little of the time (one to two days a week); 2=occasionally or a moderate amount of time (three to four days a week); or 3=most or all of the time (five to seven days a week). A total score is computed by summing each of the 20 items after reverse coding the four positively worded items. The potential range of scores is zero to 60 with values greater than or equal to 16 indicative of potential depression. Scale development included item selecti on from a pool of previously validated depression assessment tools including the B eck’s Depression Inve ntory (Beck et al., 1961), the Zung Self-Rating Depression Scale (Zung, 1965), and the Raskin Depression

PAGE 151

129 Rating Scale (Raskin, Schulterbrandt, Reati ng, & Rice, 1967). Pretests were conducted among small convenience samples to improve item clarity and acceptability. Field tests conducted by CE S-D researchers included four samples: 2,846 adults who participated in an onehour interview including the CES-D (Q1); 1,089 adults who participated in a shortened version of the Q1 interview and participated in test-retest procedures (Q2); 1,552 adults from Q1 that we re reinterviewed at a later time interval (Q3); and 105 psychiatric patien ts (Q4) (Radloff, 1977). Th ese field test samples were used to generate both reliability and validity estimates. Although factor analysis conducted on field test samples identified four factors with loadings above .40 including depressed a ffect, positive affect, somatic and retarded activity, and interpersonal pr oblems (Derogatis & Lynn, 1999; Radloff, 1977), internal consistency estimates were high resulting in the recommendation that the total aggregate score be utilized. Coefficien t alpha values in the genera l population were .85 for Q1 whites, .85 for Q2 whites, and .84 for Q3 whites, and the value was .90 in the clinical sample (Radloff, 1977). Test-retest reliability estimates were exp ected to be moderate given that the CESD scale was designed to measure current le vel of depression, which is expected to fluctuate across time. Test-r etest assessments were conduc ted using two methods (selfreport mail-back and one-on-one interview) a nd across two, four, six, eight, and 12 week intervals. Correlations ranged from .51 to .67 for the mail-back method (n=419) and from .32 to .54 for the one-on-one inte rview method (n=1441) (Radloff, 1977). Scale validation was demons trated by testing the instrument’s ability to discriminate among clinical and non-clinical adults. Results revealed much higher CES-

PAGE 152

130 D scores for psychiatric inpatien ts as compared to non-clinical adults. Of the psychiatric inpatients, 70% scored at or above the cutoff value of 16, whereas only 21% of the nonclinical sample scored at or above the threshold level (Radloff, 1977). Convergent validity was also assessed by correlating test sc ores from the CES-D with other measures of depressive symptoms including the Rock liff Depression Rating Scale (Rockliff, 1971), the Raskin Depression Rating Scale (Raskin et al., 1967), and the Hamilton Rating Scale (Hamilton, 1960), among psychiatric patients. Correlations ranged from .44 to .54 at time of admission (Radloff, 1977). The home interviews conducted with fiel d test samples Q1, Q2 and Q3 also included an array of additional measures su ch as depressed mood and well-being. Across all samples, the pattern of correlations between the CES-D and these other measures provided evidence of discriminant validity. Co rrelations were much higher between the CES-D and other indices of depressed mood or psychologi cal symptoms (.37-.63) as compared to correlations between the CES-D and indices intended to measure positive aspects of functioning or othe r variables such as medicati on usage or physical illness (.25 .32) (Radloff, 1977). Other validati on studies using the CES-D have been conducted showing further evidence of constr uct validity and can be reviewed elsewhere (Derogatis, Lipman, & Covi 1973; Weissman, Sholomaskas, Pottenger, Prusoff, & Locke, 1977). In this study, composite CES-D scores at one month, six months, 15 months, 24 months and 36 months will be included as predictor variables. Maternal depression symptomatology at first grade will be considered an excl usion variable. Women scoring at or above threshold levels of depression symptoms at first grade will be removed from

PAGE 153

131 analyses given that previous research findings have documented inflated maternal ratings of adverse outcomes in children in the context of simultaneous depression. In some analyses, the composite CES-D score will be dichotomized to represent whether or not a woman scored at or above th reshold levels for symptomatic depression. In other analyses, the continuous depression score will be use d. Yet still, to investigate the impact of maternal depression chronicity on child social compet ence and display of problem behaviors at first grade, a series of dummy variables will be created at each time point. The dummy variables will represent women who: (1) are newly depressed; (2) have remitted depression; (3) have recurrent depression, or depression after remittance; and (4) are chronically depressed. For each dummy variable, the reference category will be nondepressed women. Table 6 outlines the maternal depression va riables selected for study inclusion, the data collection site for each time point, and the number of mothers completing the instrument at each assessment.

PAGE 154

132 Table 6 Maternal Depression Variables Selected for Study Inclusion Maternal Depression Data Source # Participants 1 month 6 months 15 months 24 months 36 months 1st grade CES-D CES-D CES-D CES-D CES-D CES-D 1,343 1,265 1,223 1,097 1,181 1,002 Maternal Sensitivity Maternal sensitivity is proposed to me diate the relationship between maternal depression and child social competence and disp lay of problem behaviors at first grade. Maternal sensitivity is also hypothesized to moderate the relationship, with maternal depression having the greatest negative impact on child social competence and display of problem behaviors at first grade among children of mothers rated less sensitive. Maternal sensitivity is best observed in the context of mother-child interaction. No standard measurement instrument of moth er-child interaction was prevalent in the literature, prompting NICHD SECC researcher s to sponsor the development of its own procedure and scaling instrument for admi nistration, the Mother-Child Interaction procedure (NICHD Early Child Care Resear ch Network, 1991b), of which maternal sensitivity is one key component. The Mother-Child Interaction procedure (NICHD Early Child Care Research Network, 1991b) consists of a semi-structured pl ay situation in which maternal and child interactive behaviors are obser ved and rated for quality. Th e procedure was designed to elicit a representation of interactions that may occur between mother and child and to

PAGE 155

133 highlight the mother’s capacity to interact in a sensitive, warm, and stimulating manner with her infant (Research Tr iangle Institute, 1999b). The play situation can occur in either the child’s home or in a laboratory setting, and activities involved in the play situation vary depending on the child’s developmental age. Trained coders view the interaction via ta ped video recordings and rate quality of behaviors on a four-point or seven-point scal e ranging from ‘not at all characteristic’ to ‘highly characteristic’ for the four-point scale, and ‘very low’ to ‘very high’ for the seven-point scale. The mate rnal behaviors rated in th e Mother-Child Interaction procedure to assess maternal sensitivity at six, 15, and 24 months include sensitivity to nondistress, positive regard for the child, a nd intrusiveness. At 36 months, three conceptually similar but more age appropriate maternal behaviors toward the child were observed including supportive presence, resp ect for autonomy, and hostility. Higher scores on the maternal sensitivit y composite represent higher leve ls of the attribute. Brief descriptions of the maternal behaviors indicative of maternal sensitivity are described in Table 7. More detailed markers of mate rnal sensitivity as delineated by NICHD SECC researchers include: (1) acknowledgment of the child’s affec t; (2) contingent vocalizations by the mother/mother conversation that is responsive to the content of the child’s talk and or activity, for three year olds; (3) facilitating th e manipulation of an object of child movement/facilitating, but not over controlling the child’s play with objects or his or her motor activity, for three ye ar olds; (4) appropriate attention focusing; (5) evidence of good timing paced to the child’ s interest and arousal level; (6) slowing the pace when the child appears overstimulated or tired/changing the pace when the child appears understimulated, overexc ited, or tired, for three year olds; (7) picking up on the

PAGE 156

134 child’s interest in toys or games; (8) shared positive affect; (9) encouragement and praise of the child’s efforts; (10) providing an a ppropriate level of stimulation when needed; (11) sitting on the floor or a lo w seat, at the child’s level, to interact; and (12) timely discipline that matches the nature of the violation under consideration and the child’s ability to understand and benefit from whatever reprimand is offered, for three year olds (NICHD Early Child Care Research Network, 1992, 1993). Table 7 Brief Description of Maternal Sensitivity Behaviors Assessed Via the Mother-Child Interaction Procedure Behavior Brief Description of Behavior Sensitivity to non-distress Extent to which mo ther-child interacti on is characterized by prompt and appropriate responses to the child’s social gestures, expression s, and signals, and is generally child-centered. Intrusiveness Degree to which the mother imposes her agenda on the child as opposed to interacti ng in a way that provides a sense of control to the child. Positive regard for the child Quality and qua ntity of expressions to the child that connote the mother’s positive feelings toward the child. Supportive presence Extent to which mother expresses positive regard and emotional support to the child. May include acknowledging the child’s accomplishments, encouraging the child, or providing reassurance when the child is having difficulty with a task. Respect of autonomy Reflects the degree to which the mother acts in a way that recognizes and respects the validity of the child’s individuality, motives, and perspectives. Hostility Maternal expression of anger, discounting, or rejecting the child. The Mother-Child Interaction procedure was developed in response to a lack of previous standardized measures to assess quality of mother-ch ild interaction. The activities included in the play situations were derived from previous research on parent-

PAGE 157

135 child interaction (Snow, 1989; Vandell, 1979). No further detail is provided on the development of the Mother-Chi ld Interaction procedure. Deborah Vandell and Margaret Owen devel oped the qualitative rating scale to assess observed behaviors in th e procedure (Research Triang le Institute, 1999b) based on previous research on parenting quality and its importance to the development of secure attachment relationships and competence in children (Ainsworth et al., 1978a; Egeland & Farber, 1984; Fish, Stifter, & Belsky, 1993). In developing the qualitative rating for maternal sensitivity, Vandell and Owen supported the observation of maternal sensitivity to nondistress, positive regard for the child, a nd intrusiveness at th e six, 15, and 24 month observations, and supportive presence, respect of child autonomy, and hostility at the 36 month observation. Detailed descriptions of mothers exhibiting each behavior were explicated to facilitate codi ng of observed play interactio ns between mother and child. Estimates of internal consistency reli ability were computed for maternal sensitivity composites at each data colle ction time point and ranged from 0.70 to 0.78 (Appelbaum, Batten, & Wendell, 1994; Bl and, Batten, Appelbaum, & Wendell, 1995, 1996; James, Appelbaum, Batten, & Wendell, 1993) Intercoder reliability at each time point was also assessed by having two coders independently rate levels of maternal sensitivity for 19-20% of the tapes randomly selected at each time point (NICHD Early Child Care Research Network, 1999a, 1999b). Intraclass correlations for maternal sensitivity were .87, .83, .84, and .84 at six, 15, 24, and 36 months, respectively. No evidence of validity was provided. Maternal sensitivity is hypothesized to bot h mediate and moderate the relationship between maternal depression during the firs t three postpartum years and child social

PAGE 158

136 competence and display of problem behavior s at first grade. Maternal sensitivity composites created from observation data from the Mother-Child Interaction procedure at six months, 15 months, 24 months, and 36 mont hs, continuously scored, will each be used in this study. Table 8 outlines the meas urement of maternal sensitivity including time assessment, data source, maternal behaviors rated, and response options.

PAGE 159

137 Table 8 Measurement of Mate rnal Sensitivity Time Assessment Data Source Rated Maternal Behaviors Response Coding 6 Months Mother-Child Interaction Procedure Sensitivity/responsivity to nondistress Positive regard for the child Intrusiveness (reverse scored) 1=Not at all characteristic 2=Minimally characteristic 3=Moderately characteristic 4=Highly characteristic 15 Months Mother-Child Interaction Procedure Sensitivity/responsivity to nondistress Positive regard for the child Intrusiveness (reverse scored) 1=Not at all characteristic 2=Minimally characteristic 3=Moderately characteristic 4=Highly characteristic 24 Months Mother-Child Interaction Procedure Sensitivity/responsivity to nondistress Positive regard for the child Intrusiveness (reverse scored) 1=Not at all characteristic 2=Minimally characteristic 3=Moderately characteristic 4=Highly characteristic 36 Months Mother-Child Interaction Procedure Supportive presence Respect for child autonomy Hostility (reverse scored) 1=Very low 2=Low 3=Moderately low 4=Moderate 5=Moderately high 6=High 7=Very high

PAGE 160

138 Social Competence at First Grade Social competence at first grade is one of two major dependent variables selected for this study. Social competen ce during the early elementary period is necessary to develop and maintain adequate social relationships and is a key component of healthy social development of children. The Social Skills Rating System (SSRS ) (Gresham & Elliott, 1990) is the m easurement tool employed by NICHD SECC researchers to assess levels of social competence among first grade youth. The instrument has been well validated, and allows the assessment of several aspects of social competence, as well as ratings to be conduc ted by multiple individuals including parents and teachers. The Social Skills Rating System (SSR S) (Gresham & Elliott, 1990) is a comprehensive assessment of social skills fo r children ages three to 18 years old. Behaviors indicative of social competence are rated for frequency and importance. Dimensions of social competence assesse d by the SSRS include cooperation, assertion, responsibility, and self-control. NICHD researchers, using existing SSRS items, developed an additional social competence subs cale entitled peer competence. Separate questionnaires are available for preschool (age 3-5), elementary (grades K-6), and secondary (grade 7-12) age youth. The SSRS allows for multi-rater assessments to be completed for each child including a separate ra ting scale for teachers, parents, as well as a self-report form for students (Merrell, 1999c). Each rater form can be used alone or in conjunction with the other rater forms.

PAGE 161

139 The first grade SSRS for mothers consists of 38 items documenting the perceived frequency of target behavior s that influence the development of social competence. Response options for each item are 0=never, 1=sometimes, and 2=very often. The instrument was designed to identify deficits in positive social behaviors grouped under four subscales including cooperati on, assertion, responsibility, and self-control. The first grade SSRS for teachers consists of 30 ite ms documenting the perceived frequency of socially competent behaviors. The social skill items are similar and in some cases identical to those included on the SSRS moth er form, but presented in a classroom context. Response options are the same. The di mensions of social sk ills that are assessed with the teacher form include cooperation, assertion, self-cont rol, and peer competence. Responsibility is not assessed with the teacher form. The SSRS items were developed based on empirical research from child development, clinical psychology, educa tional psychology, and special education literatures (Gresham & Elliott, 1990). An initial item pool of 100 was reduced to 52 behaviors following field tests (Clark, Gresham, & Elliott, 1985). A national pilot test was conducted during 1987 to finalize items, refi ne administration protocols, recruit for a national standardization sample, and assess psychometric pr operties of the instrument. During 1988, the refined SSRS was standardi zed on a national sample of 4,170 children using self-ratings as well as ratings of children made by parents/guardians (n=1,0273) and teachers (n=259) (Gresham & Elliott, 1990). Standardization procedures included conve rting scale raw scores to standard scores with a mean of 100 and standard deviation of 15 based on percentile ranks (Gresham & Elliott, 1990; Merrell, 1999c). St andard scores were stratified by child’s age

PAGE 162

140 group, sex, and handicap status. Cutoff points of plus and mi nus one standard deviation were selected to establish estimates of behavior level categories. A behavior level categorization of “fewer” indicates that a pa rticular child scored below one standard deviation of the mean score, a categorization of “a verage” indicates that the child scored within plus or minus one standard deviati on of the mean, and a categorization of “more” indicates that a child scored above one standa rd deviation of the mean. These behavioral levels were developed by SSRS researchers to easily enable professionals to classify children suspected to have significan t social behavior problems. Reliability assessments conducted by th e American Guidance Service for the SSRS have included internal cons istency reliability and test-r etest reliability (Gresham & Elliott, 1990). Estimates of coefficient alpha reliabilities (Cronbach, 1951) for the elementary level SSRS as reported by the American Guidance Service (Gresham & Elliott, 1990) are summarized in Table 9. Ov erall, estimates of internal consistency reliability indicate a relatively high degree of scale homogeneity. For the elementary form of the SSRS, the American Guidance Se rvice asked parents (n=45) and teachers (n=288 for the social skills subscale) to ra te the same children 4 weeks after the first rating to assess stability reliability. Results indicated a high level of stability reliability and are summarized in Table 8 (Gresham & Elliott, 1990).

PAGE 163

141 Table 9 Coefficient Alpha and Stability Reliability Coefficients for the Elementary Level SSRS Parent/Mother Form Teacher Form Social Competence Subscale Internal Consistency Reliability Stability Reliability Internal Consistency Reliability Stability Reliability Cooperation Assertion Responsibility Self-Control Total Scale .77 .74 .65 .80 .87 .81 .77 .84 .77 .87 .92 .86 -.91 .94 .88 .75 -.80 .85 Validity assessments conducted and repor ted by the American Guidance Service (Gresham & Elliott, 1990) included estimates of content, criterion, and construct validity. Content validation efforts included a broa d survey of empirical literature on the assessment and training of social skills in children and adoles cents to guide item development to ensure adequate representa tion of the domains of interest (social competence, problem behaviors, and academic competence). Further validation efforts included selecting a panel of qualified e xperts to review, comment, and ultimately nominate a pool of items to include in field and pilot tests. Criterion-related validity was established for the elementary levels of both the parent and teacher forms of the SSRS using co nvergent validity procedures against the Child Behavior Checklist-Parent Report Fo rm (CBCL) (Achenbach & Edelbrock, 1983), the Social Behavior Assessment (SBS) (Ste phens, 1978) and the Harter Teacher Rating Scale (TRS) (Harter, 1985). Correlations between the social competence subscales of the SSRS and the social competence subscale of the CBCL ranged from .37 .51, with the total social competence scale of the SSRS correlating .58 with the social competence

PAGE 164

142 subscale of the CBCL. The total SSRS social competence subscale co rrelated -.68 with the total SBS social skills subscale (low scor es on the SSRS indicate social incompetence whereas high scores on the SBS i ndicate deficits in social skil ls). Further, results from convergent validity procedure indi cated that students rated as well adjusted on the Harter TRS tended to have well-developed social sk ills as measured by the SSRS (Gresham & Elliott, 1990). SSRS researchers report establishing furt her evidence for construct validity including (1) factor analysis to assess the test’s factorial validity; (2) determining if changes in test scores refl ect expected developmental changes and sex differences in social competence; and (3) group separation or comparing groups th eoretically expected to differ on the test. Results from these te sts all support the cons truct validity of the SSRS and can be reviewed in detail elsewhere (Gresham & Elliott, 1990). This study will investigate the impact of depression on both maternal report and teacher report of child social competence at first grade. Further, for both maternal and teacher reports of child social competen ce, the impact of depression on the total composite will be explored, as well as for each subscale including cooperation, assertion, responsibility (mother form only), self-contro l, and peer competence. The level of measurement for these variables is continuous. Raw scores will be used in analyses for each SSRS subscale, and standardized scores will be used for both total composites (mother and teacher reported). Furthermore, each interval-level standardized SSRS total composite score will be dichotomized to re present youth with average or greater than average social competence versus those with less than average, or at-risk, levels of social

PAGE 165

143 competence. The items used to measure each subscale, for both mother and teacher forms of the SSRS, along with corresponding re sponse options, are listed in Table 10.

PAGE 166

144 Table 10 Measurement of Social Competence Variable Data Source Item Response Coding Uses free time in an acceptable way. Keeps room clean and neat without being reminded. Congratulates family members on accomplishments. Puts away toys or other household property. Volunteers to help family members with tasks. Helps you with household tasks without being asked. Attempts household tasks before asking for your help. Mother-Rated Cooperation 1st Grade SSRS Mother Form Gives compliments to friends or other children in the family. Completes household task s within a reasonable time. Uses time appropriately while waiting for your help with homework or some other task. 0=Never 1=Sometimes 2=Very Often Teacher-Rated Cooperation 1st Grade SSRS Teacher Form Uses free time in an acceptable way. Finishes class assignments within time limits. Uses time appropriately while waiting for help. Produces correct schoolwork. Follows your directions. Puts work materials or school property away. Ignores peer distracti ons when doing classwork. Keeps desk clean and neat without being reminded. Attends to your instructions. Easily makes transition from one classroom activity to another. 0=Never 1=Sometimes 2=Very Often Continued on the next page

PAGE 167

145 Table 10 (Continued) Variable Data Source Item Response Coding Mother-Reported Assertion 1st Grade SSRS Mother Form Joins group activities without being told. Invites others to your home. Makes friends easily. Shows interest in a variety of things. Is liked by others. Starts conversations rather than waiting for others to talk first. Is self-confident in social situ ations such as parties or group outings. Accepts friends’ ideas for playing. Easily changes from on e activity to another. Reports accidents to appropriate persons. 0=Never 1=Sometimes 2=Very Often Teacher-Reported Assertion 1st Grade SSRS Teacher Form Introduces himself/herself to new people without being told. Appropriately questions rules that may be unfair. Says nice things about himsel f/herself when appropriate. Invites others to join in activities. Makes friends easily. Initiates conversations with peers. Appropriately tells you when he or she thinks you have treated him or her unfairly. Gives compliments to peers. Volunteers to help peers with classroom tasks. Joins ongoing activity or group w ithout being told to do so. 0=Never 1=Sometimes 2=Very Often Continued on the next page

PAGE 168

146 Table 10 (Continued) Variable Data Source Item Response Coding Mother-Reported Responsibility 1st Grade SSRS Mother Form Introduces herself or himself to new people without being told. Asks sales clerks for information or assistance. Attends to speakers at meetings su ch as in church or youth groups. Politely refuses unreasonable requests from others. Answers the phone appropriately. Appropriately questions househol d rules that may be unfair. Asks permission before using anot her family member’s property. Requests permission before leaving the house. Acknowledges compliments or praise from friends. Reports accidents to appropriate persons. 0=Never 1=Sometimes 2=Very Often Mother-Reported Self-Control 1st Grade SSRS Mother Form Speaks in an appropriate tone of voice at home. Responds appropriately when h it or pushed by other children. Politely refuses unreasonable requests from others. Avoids situations that are likely to result in trouble. Receives criticism well. Controls temper when ar guing with other children. Ends disagreements with you calmly. Controls temper in conf lict situations with you. Responds appropriately to teasing fr om friends or relatives of his or her own age. Cooperates with family members without being asked to do so. 0=Never 1=Sometimes 2=Very Often Continued on the next page

PAGE 169

147 Table 10 (Continued) Variable Data Source Item Response Coding Teacher-Reported Self-Control 1st Grade SSRS Teach Form Controls temper in conflict situations with peers. Compromises in conflict situations by changing own ideas to reach agreement. Responds appropriately to peer pressure. Responds appropriately to teasing by peers. Controls temper in conflict situations with adults. Receives criticism well. Accepts peers’ ideas for group activities. Cooperates with peers without prompting. Responds appropriately when pus hed or hit by other children. Gets along with people who are different. 0=Never 1=Sometimes 2=Very Often Mother-Reported Peer Competence 1st Grade SSRS Mother Form Joins group activities without being told. Responds appropriately when h it or pushed by other children. Makes friends easily. Controls temper when ar guing with other children. Is liked by others. Gives compliments to friends or other children in the family. Is self-confident in soci al situations such as parties or group outings. Responds appropriately to teasing from friends or relatives of his or her own age. Accepts friends’ ideas for playing. Acknowledge compliments or praise from friends. 0=Never 1=Sometimes 2=Very Often Continued on the next page

PAGE 170

148 Table 10 (Continued) Variable Data Source Item Response Coding Teacher-Reported Peer Competence 1st Grade SSRS Teacher Form Controls temper in conflict situations with peers. Responds appropriately to peer pressure. Makes friends easily. Responds appropriately to teasing by peers. Initiates conversations with peers. Accepts peers’ ideas for group activities. Gives compliments to peers. Cooperates with peers without prompting. Joins ongoing activity or group w ithout being told to do so. Responds appropriately when pus hed or hit by other children. 0=Never 1=Sometimes 2=Very Often Mother-Reported Social Skills Total Raw Score 1st Grade SSRS Mother Form --Mother-Reported Social Skills Total Standardized Score 1st Grade SSRS Mother Form --Teacher-Reported Social Skills Total Raw Score 1st Grade SSRS Teacher Form --Teacher-Reported Social Skills Total Standardized Score 1st Grade SSRS Teacher Form --

PAGE 171

149 Display of Problem Behaviors Display of problem behaviors at firs t grade is the second major dependent variable selected for this study. Display of problem beha viors in children is a sign of adverse social development in creasing risk for mental health disorders and antisocial behavior in early adulthood. The Child Be havior Checklist for children ages 4-18 (CBCL/4-18) (Achenbach & Edelbrock, 1983) was selected as the measurement instrument for display of problem behaviors, given that it is the most widely accepted and used screening instrument av ailable for tracking the emerge nce of behavior problems in children. The Child Behavior Checklist for children ages 4-18 (CBCL/4-18) (Achenbach & Edelbrock, 1983) is one of the most widely researched parent rating scales of child problem behaviors with over 800 published st udies using it as the primary measure (Merrell, 1999d). The CBCL includes 118 be havioral and emotional problems of children. Parents are asked to determine how well that item describes their child presently or within the last six months. Re sponse options include 0=not true (as far as you know); 1=somewhat or sometimes true; and 2=very true or often true. Broad band dimensions of problem behavior s called syndrome scales are assessed including withdrawn, somatic complaints, anxious/depresse d, social problems, thought problems, attention problems, delinquent beha vior, and aggressive behavior. Table 11 provides a brief description of item conten t for each broad band syndrome scale as adapted from Merrell (1999). Narrow band di mensions of problem behaviors are also assessed. An internalizing problem behavi or composite scale is comprised of the withdrawn, somatic complaints, and anxi ous/depressed syndrome scales. An

PAGE 172

150 externalizing problem behavior composite scale is comprised of the delinquent and aggressive behaviors. A tota l problem behavior scale is co mprised of all eight syndrome scales. Thus a hierarchy of scores is computed for the CBCL/4-18 including from the lowest to highest level: br oad band syndrome scales, narrow band composite scales, and the total problems scale. Table 11 Brief Description of Item Content by Syndrome Scale of the CBCL/4-18 Syndrome Scale Brief Description of Item Content Withdrawn Likes to be alone, withdraw n and isolated, doesn’t get involved with other people. Somatic Complaints Reports feeling dizzy, feels overt ired, reports physical problems without known medical cause. Anxious/Depressed Unhappy, sad, or depressed aff ect, frequent crying, verbal report of suicidal ideation, nervous tension. Social Problems Does not get along well with other children, is not well liked by other children. Thought Problems Hears sounds or voices that ar en’t there, acts confused or seems to be in a fog, sees things that aren’t there. Attention Problems Impulsive behavior (acts without thinking), difficulty concentrating, fidgety or re stless behavior, inattention. Delinquent Behavior Keeps company with children who ge t into trouble, runs away from home, school truancy, lying and cheating. Aggressive Behavior Arguing, cruel behavior toward persons and animals, bullying or meanness to others, physically fighting. The CBCL/4-18 was an extension and revisi on of the Behavior Problem Checklist (Achenbach, 1966; Achenbach & Lewis, 1971) that was developed from a thorough review of the clinical and research literat ures, case histories of over 1,000 psychiatric patients, and consultation with clinical and developmental psychologists, child psychiatrists, and psychiatri c social workers (Achenbach, 1991). Clinical samples of male and female youth at ages 4-5, 6-11, a nd 12-18 were used to conduct principal components analyses leading to the developm ent and refinement of the eight broad band

PAGE 173

151 syndrome scales and two narrow band composite scales (internalizi ng and externalizing behaviors) (Achenbach, 1994). The CBCL/4 -18 was tested with a nationwide nonclinical normative sample of 2,368 youth who ha d not received mental health, substance use, or special education se rvices for major behavioral, emotional, or developmental problems in the previous 12 months (Achenb ach & Rescorla, 2001). Additional clinical cases from mental health settings were also included to establish clin ical cutoff criterion. Normalized t-scores were construc ted separately by sex and age group. Internal consistency, stability, and interr ater reliability assessments have been conducted on the CBCL/4-18 with strong suppor t for the instrument’s reliability. Estimations of reliab ility remain high after stratifica tion by sex and age group. Specific results can be reviewed elsewhere (Achenbach, 1991). Content validity was established utilizing an extensive spectrum of sources during item development to ensure thorough repres entation of the cons truct domain. As discussed previously, items were developed based on the clinical and research literatures, conversations with parents and mental health professionals regarding problems behaviors perceived to be a concern, as well as cons ultation with clinical and developmental psychologists, child psychiatrist s, and psychiatric social work ers. These individuals were perceived to be both clinical and practical experts related to child problem behaviors. Construct validity of the CBCL/4-18 has b een demonstrated via several methods including accurate distincti on among clinical and non-clinical samples. Item analyses during instrument development revealed that clinically referred children obtained significantly higher scores on the problem be havior syndrome and composite scales as compared to demographically similar non-clin ical children. The significant associations

PAGE 174

152 between the CBCL/4-18 classifications of devi ant behavior and referral status supported that the measure in fact is related to the i ndependently established me ntal health concerns that led to referral (Ache nbach, 1991), attesting to the construct validity of the instrument. Evidence of convergent valid ity against the Quay-Peterson Revised Behavior Problem Checklist (Quay & Peterson, 1983) and the Conners Parent Rating Scale (Conners, 1973) further substant iated the construct validity of the CBCL/4-18. Data were collected from parents of a sample of 60 clin ically referred 6-11 year old children from various outpatient settings (Achenbach, 1991). Parents completed the three instruments and Pearson correlations between the CBCL/418 and the two additional measures were computed. The CBCL/4-18 total problem score was corrected .81 with the total problems score of the Revised Behavi or Problem Checklist, and .82 with the Connors Parent Questionnaire (Achenbach, 1991; Merrell, 1999d). CBCL/4-18 data for total problem behavi ors, internalizing behaviors, and externalizing behaviors will serve as the major dependent variables assessing display of problem behaviors. No t eacher data were collected. Measurement of total problem behavi ors, internalizi ng behaviors, and externalizing behaviors resulted in standard ized, continuous scores, which will be used during analyses. Further, guidance wa s provided in instrument documentation (Achenbach, 1991) as to how to classify deviant and non-deviant groups based on standardized t-scores. Thus, t-scores below the recommended value of 60 will be considered non-deviant, and those at or a bove 60 will categorized as deviant (borderline

PAGE 175

153 clinical range or clinically significant). Table 12 summarizes the items of the CBCL/418 by composite and syndrome scales, al ong with corresponding response options.

PAGE 176

154 Table 12 Measurement of Problem Behaviors Variable by Composite and Syndrome Data Source Item Response Option Total Problem Behaviors T-Score 1st Grade CBCL/4-18 --Internalizing T-Score 1st Grade CBCL/4-18 --Withdrawn T-Score 1st Grade CBCL/4-18 Would rather be alon e than with other. Refuses to talk. Secretive, keeps things to self. Shy or timid. Stares blankly. Sulks a lot. Under active, slow moving, or lacks energy. Unhappy, sad, or depressed. Withdrawn, doesn’t get involved with others. 0=not true 1=somewhat or sometimes true 2=very true or often true Continued on the next page

PAGE 177

155 Table 12 (Continued) Variable by Composite and Syndrome Data Source Item Response Option Somatic Complaints TScore 1st Grade CBCL/4-18 Feels dizzy. Overtired. Aches or pains (not stomachaches or headaches). Headaches. Nausea, feels sick. Problems with eyes (not if corrected by glasses). Rashes or other skin problems. Stomachaches or cramps. Vomiting, throwing up. 0=not true 1=somewhat or sometimes true 2=very true or often true Anxious/Depressed TScore 1st Grade CBCL/4-18 Complains of loneliness. Cries a lot. Fears he/she might think or do something bad. Feels he/she has to be perfect. Feels or complains that no one loves him/her. Feels others are out to get him/her. Feels worthless or inferior. Nervous, high strung, or tense. Too fearful or anxious. Feels too guilty. Self-conscious or easily embarrassed. Suspicious. Unhappy, sad, or depressed. Worries. 0=not true 1=somewhat or sometimes true 2=very true or often true Continued on the next page

PAGE 178

156 Table 12 (Continued) Variable by Composite and Syndrome Data Source Item Response Option Externalizing T-Score 1st Grade CBCL/4-18 --Delinquent Behavior TScore 1st Grade CBCL/4-18 Doesn’t seem to feel guilty after misbehaving. Hangs around with others who get in trouble. Lying or cheating. Prefers being with older kids. Runs away from home. Sets fires. Steals at home. Steals outside the home. Swearing or obscene language. Thinks about sex too much. Truancy, skips school Uses alcohol or drugs for non-medical purposes (describe). Vandalism 0=not true 1=somewhat or sometimes true 2=very true or often true Continued on the next page

PAGE 179

157 Table 12 (Continued) Variable by Composite and Syndrome Data Source Item Response Option Aggressive Behavior T-Score 1st Grade CBCL/4-18 Argues a lot. Bragging, boasting. Cruelty, bullying, or meanness to others. Demands a lot of attention. Destroys his/her own things. Destroys things belonging to his/her family or others. Disobedient at home. Disobedient at school. Easily jealous. Gets in many fights. Physically attacks people. Screams a lot. Showing off or clowning. Stubborn, sullen, or irritable. Sudden changes in mood or feelings. Talks too much. Teases a lot. Temper tantrums or hot temper. Threatens people. Unusually loud. 0=not true 1=somewhat or sometimes true 2=very true or often true Continued on the next page

PAGE 180

158 Table 12 (Continued) Variable by Composite and Syndrome Data Source Item Response Option Other Syndrome Scales ---Social Problems TScore 1st Grade CBCL/4-18 Acts too young for his/her age. Clings to adults or too dependent. Doesn’t get along with other kids. Gets teased a lot. Not liked by other kids. Overweight. Poorly coordinated or clumsy. Prefers being with younger kids. 0=not true 1=somewhat or sometimes true 2=very true or often true Thought Problems TScore 1st Grade CBCL/4-18 Can’t get mind off certain thoughts/obsessions. Hears sounds or voices that aren’t there. Repeats certain acts ove r and over; compulsions. Sees things that aren’t there. Stares blankly. Strange behavior. Strange ideas. 0=not true 1=somewhat or sometimes true 2=very true or often true Continued on the next page

PAGE 181

159 Table 12 (Continued) Variable by Composite and Syndrome Data Source Item Response Option Attention Problems TScore 1st Grade CBCL/4-18 Acts too young for his/her age. Can’t concentrate, can’t pay attention. Can’t sit still, res tless, or hyperactive. Confused or seems to be in a fog. Daydreams or gets lost in his/her thoughts. Impulsive or acts without thinking. Nervous, high strung, or tense. Nervous movements or twitching. Poor school work. Poorly coordinated or clumsy. Stares blankly. 0=not true 1=somewhat or sometimes true 2=very true or often true

PAGE 182

160 Exposure to Nonmaternal Care Exposure to nonmaternal care is hypothes ized to moderate the relationship between maternal depression and child so cial competence and display of problem behaviors at first grade, with depression having the least negative impact on child development among children exposed to nonmat ernal care. Data on characteristics of nonmaternal care were collected using standa rdized interview guides developed for the NICHD SECC. At each home in terview assessment time point (one month, six months, 15 months, 24 months, and 36 months), moth ers were asked about current childcare arrangements including type of care. Response options fo r the child’s primary care at each time point included: (1) mother; (2) fa ther/partner; (3) grandparent, in-home; (4) grandparent, other home; (5) other relative, in -home; (6) other relative, other home; (7) non-relative, in-home; (8) non-relative, other home; (9) child care center; and (10) other. For this study, at each assessment point response options for nonmaternal care including father/partner, gra ndparent, other relative, non-rela tive, child care center, and other, will be collapsed to represent affirma tive exposure to nonmaternal care, a value of one. Mothers who report being the child’s primary caretaker will be coded as zero, representing no exposure to nonmaternal care. Next, a series of new variables will be created to represent ‘ever e xposure’ to nonmaternal care, one for six months, one for 15 months, one for 24 months, and one for 36 mont hs. These will be the variables used during analyses to explore the potential m oderating influence of exposure to nonmaternal care. Further, an additional variable will be created at each time assessment collapsing the non centerbased caretaking environments. Th is variable will consist of three levels:

PAGE 183

161 (1) no exposure to nonmaternal care (mother is primary caretaker); (2) exposure to non center-based care; and (3) exposur e to center-based care. Sim ilar to the process described above, at each time assessment (one, six, 15, 24, and 36 months), a new variable will be created to represent type of ‘ever exposure’ including non center-based and center-based care. The variable will be dummy coded w ith no exposure to nonmaternal care as the reference category. These variables will be us ed during analyses to explore the potential moderating influence of exposure to center-based nonmaternal care. NICHD SECC Data Co llection Methods During Phase I, NICHD data collecti on procedures followed an age-based chronological schedule at five major assessment periods including one month, six months, 15 months, 24 months, and 36 months. Data collected at one month postpartum were gathered during an interview conducted in the child’s home by trained personnel. Sociodemographic and exposure to nonmaterna l care data were coll ected using the One Month Home Interview standardized guide. In addition, ratings of maternal depression were assessed at one-month via self-report using the CES-D. Although all 20 items of the original CES-D were included in the assessment, questions were reordered. At six months, exposure to nonmaternal ca re, maternal depression, and maternal sensitivity were assessed in the child’s hom e. The Six Month Home Interview guide (parallel to the One Month Home Interview) was used to collect data on the infant’s exposure to and type of childcare arrangement and the CES-D was again administered as a self-report scale to determine le vels of maternal depression. The Mother-Child Interaction procedure was administered to assess display of maternal sensitivity during a 15-minute play se ssion that was observed by the interviewer

PAGE 184

162 and videotaped to allow for subsequent rating by trained coders blind to other information on participating families. The pl ay situation consisted of two episodes that were implemented in a fixed order (NICHD Ea rly Child Care Research Network, 1991b). During the first seven to eight minutes, mother s were encouraged to spend time in a fixed location with their infants and asked to play with their infants using any toy or object available in the home or none at all. Duri ng the second seven to eight minutes of the procedure, mothers were given a standard set of toys they could use in play with their infants. Examples of toys included a stuffe d animal, a twin rattle a ball, and a book with shapes and faces. During the 15-month data collection time point, exposure to nonmaternal care, maternal depression, and maternal sensitivity were once again ascert ained in the child’s home. Exposure to nonmaternal care was assess ed using the Home Interview (parallel to both the One Month Home Interview and the Six Month Home Interview) administered by trained staff, and maternal depression was measured using the self-report CES-D. For maternal sensitivity, the observation procedure differed from the protocol implemented at six months, and followed a “three boxes” procedure in which mothers were asked to show their child ren age-appropriate toys in th ree containers in a set order (NICHD Early Child Care Research Networ k, 2001). The home visitor instructed the mother to have her child spend time with each of the three toys in the labeled bags in sequential order during the next 15 minutes. Toys were selected that would: (1) be interesting to children at that particular age; (2) foster diffe rent types of activities; and (3) have the potential of being used differently by children and their mothers. Mothers were not instructed to interact with their child, nor were they advised as how to break up the 15

PAGE 185

163 minutes among the three toys. Ambiguous inst ructions were purposefully provided to assess: (1) how sensitive mothers were to thei r child’s interests and needs without clear instructions from the home visitor; (2) if mo thers appeared either uninvolved or intrusive in the face of ambiguity; and (3) how mothers handled the transition from one toy to another. At 24 and 36 months, data were collected both in the child’s home and in laboratory settings. Trained staff personnel once again asse ssed exposure to nonmaternal care via administration of the Home Intervie w. Maternal depression was measured via self-report using the CES-D. Maternal se nsitivity was observed following the same “three boxes” procedure descri bed in the 15-month assessment, however observation of mother-child interaction at 24 and 36 months occurred in the laboratory. During Phase II, data include d in this study were collect ed at first grade in the laboratory and in the child’s school. In the laboratory, following a greeting and brief explanation of the visit’s activities, mothers were provided with various questionnaires to complete. Maternal depression was assessed using the CES-D, and maternal report of child social competence was measured using a modified version of the elementary level SSRS. Modifications to the original SSRS included not rating behaviors for importance, only frequency. Child display of problem beha viors was also assessed at first grade in the laboratory via maternal completion of the CBCL/4-18. Teacher reports of child social competence were collected at the child’s schoo l. Teachers were asked to complete the elementary level SSRS. As with the mother form of the elementary SSRS, socially competent behaviors were only rated for frequency, not importance.

PAGE 186

164 Data Scoring Scoring for the CES-D at all assessment points was completed by first reverse coding the four positively worded items on the scale and summing all 20 items to create a total score for maternal depression where ther e were no missing values. For individuals with four or fewer missing items on the CESD scale, composite scores were created using proportional weighting techniques. Fo r proportional weighti ng, a composite score was imputed by first dividing the number of items completed by an individual by the total number of items in the scale, then mu ltiplying the product by the obtained summation score. Scores potentially ranged from zero to 60 with higher scores representing higher levels of depressive symptomatology. Scores greater than or equal to 16 were considered to be of clinical significance. Videotaped interactions from the Mothe r-Child Interaction Procedure from all assessment points were sent to the Timberla nd Foundation to be coded by trained raters using the qualitative scales to rate maternal and child interactive behaviors (NICHD Early Child Care Research Network, 1992). At six, 15, and 24 months, three four-point maternal behavior ratings (sensitivity to nondi stress, positive regard, and intrusiveness, reverse scored) were summed to create a ma ternal sensitivity composite (NICHD Early Child Care Research Network, 2001). At 36 mo nths, three seven-point maternal behavior ratings, age-appropriate vers ions of those ascertained at six, 15, and 24 months (supportive presence, respect for autonomy, and hostility, re verse scored) were summed to create a maternal sensitivity composite. For child social competence, scoring fo r each SSRS subscale (for both mother and teacher reports) was conducted using the ASSISTTM software purchased from the

PAGE 187

165 American Guidance Service (Circle Pines, MN). The software considered the amount of missing data for each respondent and only comput ed subscale composite scores for those individuals missing responses for 2 or fewer ite ms for each subscale. If an individual did have missing values on 2 or fewer items, a 1 was inputted for that item prior to summation. For each subscale, higher scores represented a stronger affinity to demonstrate the characteristic of interest (c ooperative skills, assertiv eness, responsibility, self-control, or peer compet ence) as perceived by the mo ther or teacher. For both maternal and teacher reports of child social competence, a social skills total raw score was computed by summing the raw scores for each of the items in the instrument. Higher values indi cated greater youth displays of socially acceptable learned behaviors, as perceived by the mother or teacher. Standardized scores were also computed using the raw scores with a scor e above 100 indicating th at the raw score was above the mean score of similar students in which the instrument was standardized. Child display of problem behaviors, as assessed via maternal completion of the CBCL/4-18, was scored using the Cross-Inform ant Program of the Achenbach System of Empirically Based Assessment (ASEBA) Windows software for the CBCL/4-18 (University Medical Educati on Associate, Inc., Burlingt on, VT). Scoring was only completed for tests with no more than 20 missing items on the instrument excluding 4 items that asked parents to write-in ‘othe r’ responses. The Cross-Informant Program computed standardized t-scores for each syndrome scale by first assigning each raw score to the midpoint of the percent ile of the cumulative frequency distribution that it spanned. Midpoint percentiles were used rather than cumulative percentiles to provide a smoother, less skewed basis for t-score computation. After obtaining the midpoint percentiles,

PAGE 188

166 normalized t-scores were assigned enabling s cales to be directly compared given a common metric. The standardized t-scores for each syndrome scale potentially ranged from 50 to 100 with higher scores indicati ng a stronger affinity to demonstrate that particular behavior. For each of the composite scales (internali zing behaviors, externalizing behaviors, and total problem behaviors), summing the items comprising each scale generated raw scores which were then were converted to standardized t-scores in the same process described above. The standardized t-scor es for internalizing problems behaviors potentially ranged from 31-100, 30 to 100 for ex ternalizing behaviors, and 23 to 100 for total problem behaviors. Higher scores we re indicative of a stronger affinity to demonstrate that particular behavior. For the internalizing, exte rnalizing, and tota l problem behavior composites, tcores from 60 to 63 are considered to be in the borderline clinical range, t-scores greater than 63 are considered to be of clinical signi ficance, and t-scores under 60 are considered to be within normal range (Achenbach, 1999) For categorical discrimination between deviant and non-deviant groups related to pr oblem behaviors, t-scores below 60 are recommended (Achenbach, 1991). Data Analysis Several data analytic techniques we re considered for the present study including survival analysis and hierarchical linear modeling (HLM). The multiple observations of maternal depression during th e first three postpartum years as predictor variables complicates the inclusion of each assessment of maternal depression in a single model due to issues of multicollinearity. Although survival analysis allows for time-

PAGE 189

167 varying covariates to be incl uded in regression models, the dependent variables for this study are not appropriate for event history analysis. Furthermore, HLM is most often applied to situations where repeated measures exist for the dependent variable. Although grow th factors (change over time) can be used to predict developmental outcomes, interpreta tion of findings is statistically complex and results may not be practical for public heal th purposes. Given that this study hopes to elucidate the contexts in which maternal depr ession poses the greatest risk to child social competence and display of problem behaviors, interpretability and practicality of results is critical. These values underl ie the decision for the proposed data analytic plan. This plan includes a combination of univariate, bi variate, and multivariate analyses. The SAS software program version 8.01 (SAS Institut e Inc, Cary, NC) will be used in data analyses. A description of the univariate, bivariate, and multivariate analyses follows. Univariate and Bivariate Analyses First, characteristics of families incl uded in the study sample will be compared to those who were excluded from the study sa mple based on unavailability of complete contextual data or exposure to mate rnal depression at first grade. Second, frequency distributions of every cate gorical variable to be included in the study will be computed to assess response di stributions. For continuous variables, means and standard deviations will be generated. Through these steps, response values that are out of range or non-meaningful can be identified and recoded. For categorical variables, response options may be collapsed further based on response dist ributions. Response frequencies for study sample sociodemographi c variables will be generated. Further, tests of distribution normality for each conti nuous dependent variable will be conducted.

PAGE 190

168 Next, bivariate analyses will be conducted to test the association between each of the sociodemographic control variables (child sex, child birth order, maternal age, maternal race, maternal ethnicity, maternal education, presence of husband or partner in the home, whether or not the family is on public assistance, and breastfeeding at one month postpartum) and (1) maternal depression; (2) maternal sensitivity; (3) child social competence and display of problem behavior s at first grade; and (4) exposure to nonmaternal care. Then, Pearson correlation matrices will be computed for maternal depression and maternal sensitivity across data collection time points (one month, six months, 15 months, 24 months, and 36 months) to assess association over time. Pearson correlation matrices will also be computed to assess (1) the associations between maternal depression and maternal sensitivity across data collection time points; (2) the associations between maternal depressi on over time and each of the continuous dependent variables; and (3) the associations between maternal sens itivity over time and each of the continuous dependent variables. Bi variate associations will also be explored between (1) maternal depression and exposur e to nonmaternal care; and (2) maternal sensitivity and exposure to nonmaternal care. Chi-squares will be computed for analyses investigating an association between two categorical variables, and independent samples t-tests or ANOVAS will be computed for analyses investigating an association be tween a categorical independent variable and a continuous dependent variable. Last, a Pearson correlation coefficient will be conducted to assess the association between maternal and teacher reports of social competence in children at first grade.

PAGE 191

169 Multivariate Analyses A combination of ordinary least s quare (OLS) regression and logistic regression models will be built to explore the impact of maternal depression on child social competence and display of problem beha viors at first grade. OLS regression will be used when the dependent variable is c ontinuously scored (total social competence, social competence subscale scores, total beha vior problems, total internalizing problems, and total externalizing problems); and logi stic regression will be used when the dependent variable is dichotomous (less than average versus average or greater than average social competence; and deviant versus non-deviant display of total problem behaviors). A specific order of steps will be followed when building regression models (either OLS or logistic). First, the sociodemographi c control variables will be entered as a block to account for as much variance in the depende nt variable as possibl e, followed next by the predictor variable (maternal depression). If testing a mediating influence, as with hypothesis two, the mediating vari able will next be entered into the model. After all variables of interest have been entered into the regression m odel, a series of diagnostics will be conducted including tests for outliers, leverage, influence, and collinearity. The last step in model building w ill be the testing of moderating effects, if appropriate given the hypothesis. Each interacti on term to be tested (mater nal sensitivity and exposure to nonmaternal care) will be entered into the full model individually. In OLS regression, the model R-square, or percentage of variance in the dependent variable that is acc ounted for by the independent and control variables, will be used to assess the overall ability of the independent and control variables to predict each

PAGE 192

170 dependent variable. The signi ficance of maternal depres sion (controlling for all other variables in the model) in predicting child social competence at first grade will be assessed by the p-value associated with it s beta weight, or standardized multiple regression coefficient. A p-value less than 0.05 will be used as the criterion for judging statistical significance and rejec ting the null hypothesis that th e regression coefficient is equal to zero. For logistic regression, the overall model Wald F will be used to assess the model’s goodness of fit and to test the null hypothesis that none of the independent variables are related linearly to the log odds of the depe ndent variable. The Cox and Snell R-square, similar to the model R-square statistic computed in OLS regression, will also be calculated to approxi mate the strength of associa tion between the predictor and criterion variables. The si gnificance of maternal depres sion (controlling for all other variables in the model) in predicting child social competence at first grade will be determined by its associated odds ratio and 95% confidence interval. A p-value less than 0.05 will be used to judge statistical signifi cance and reject the null hypothesis of no association. Given that a major aim of this study is to explicate what aspects of depression pose the greatest risk to child social compet ence and display of probl em behaviors at first grade, considering timing of initial onset, se verity of symptoms along the trajectory of initial onset, and chronicity of symptoms a series of models will be tested and unstandardized regression coefficients compared, for hypotheses 1.2 through 1.4. For example, to test the hypothesis that timing of initial onset of maternal depression is associated with child social competence and di splay of problem behaviors at first grade,

PAGE 193

171 with earlier onset episod es having the greatest negative impact, the following series of models will be computed, where y equals the dependent variable and x1 through x36 represent maternal depression at various tim e points (one month, six months, 15 months, 24 months, and 36 months), dichotomously scored. y=x1 y=x6 (among those with no previ ous exposure to depression) y=x15 (among those with no previous exposure to depression) y=x24 (among those with no previous exposure to depression) y=x36 (among those with no previous exposure to depression) Significance of maternal depression will be assessed for each model, and in cases where maternal depression is a significant predictor of child social competence and display of problem behaviors at first grade, unstandardized regression coefficients will be compared to determine which time point ha d the greatest magnitude of influence. Similarly, to test the hypot hesis that severity of ma ternal depression along the trajectory of initial onset is associated w ith child social competence and display of problem behaviors at first grade, with severi ty of symptoms having the greatest negative impact with earlier onset episodes, the fo llowing series of mode ls will be tested. y=x1 y=x6 (among those with no previ ous exposure to depression) y=x15 (among those with no previous exposure to depression) y=x24 (among those with no previous exposure to depression) y=x36 (among those with no previous exposure to depression) This time, maternal depression will be scor ed continuously. Unst andardized regression coefficients for each assessment of maternal depression found to be significant will be compared to ascertain which time point had the greatest magnitude of influence. Last, to test the hypothesis that chronicity of maternal depression is associated with child social competence and display of problem behaviors at first grade, with

PAGE 194

172 chronicity of symptoms having the greatest impa ct at later time point s, another series of models will be built. For this series of regr ession models, at each assessment time point (one month, six months, 15 months, 24 months and 36 months), the maternal depression predictor variable will be a sequence of dummy variables comp aring women who are newly depressed (dep_new), those with remitted depression (dep_remitted), those with recurrent depression or depression after remittance (dep_recurrence), and those with chronic depression (dep_chronic). The refe rence category for these dummy variables will be never depressed women. Time 1: y=dep_new Time 2: y=dep_new dep_remitted dep_chronic Time 3: y=dep_new dep_remitted dep_recurrence dep_chronic Time 4: y=dep_new dep_remitted dep_recurrence dep_chronic Time 5: y=dep_new dep_remitted dep_recurrence dep_chronic Design Limitations This study has several design limitations. Fi rst, the study is longitudinal in nature thus data were collected in two phases. From Phase I to Phase II of the study, 144 (10.6%) families were lost to at trition. Given that the mothers of infants lost to attrition were more likely to be young, non-white, less educated, living with no husband or partner at home, and receiving public assistance, ge neralization of study results are limited. Further, data for this study will only cons ist of infants and their families that had complete contextual data on all variables of interest. Infants and families with incomplete contextual data may statistically differ from those with complete data related to the independent and depende nt variables in the study. Since this study involves secondary data analysis, the selection of measurement instruments to assess the major variables in the study may not have matched the primary

PAGE 195

173 choice of the researcher. In some instan ces, NICHD SECC documentation provided little rationale as to why a given instrument was selected. Further, although the CES-D is a widely accepted and employed measurement of depressive symptomatology among non-clinical samples, it is not a diagnostic tool for maternal depression. Women scoring at or above suggested cut points indicative of depression, may not in fact be experiencing de pression. It must be emphasized that the measurement instrument for maternal depres sion in this study only captures depressive symptomatology. The assessment tool for maternal sensitivity was created for implementation in the NICHD SECC, with little documentation of psychometric properties including reliability or validity. Videotaped pl ay situations between mother and child were sent to independent reviewers for coding. No discussion of efforts ta ken to enhance or empirical evidence of interrater reliability is provided. Although assured that information collect ed during study participation would be confidential, data collection of ten occurred in the context of a home interview, with the home visitor having full knowledge of the part icipant’s identity and living situation. Many items ascertained during the home visit were of a sensitive nature including experiences of depression. This scenario may have biased maternal responses. In addition, maternal behaviors may have been altered as a result of data collection procedures. In the case of the Mother-Child Interaction pr ocedure, mothers were fully aware that they were being observed in inte raction with their child. This may have promoted social desirability bias and decrea sed sample variation related to maternal sensitivity.

PAGE 196

174 Phase I of the NICHD SECC began in 1991 when infants were one month of age. Infants born today may be di fferent from infants born in 1991, potentially limiting the generalization of results.

PAGE 197

175 Chapter Four: Results Study Sample A total of 679 children and their families met inclusion criteria for this study, which included: (1) participation in both Ph ase I and Phase II of the NICHD SECC; (2) complete data on all variables of interest – sociodemographic va riables, maternal depression, maternal sensitivity, social compet ence at first grade, display of problem behaviors at first grade, and exposure to nonmaternal care; and (3) non-presence of elevated levels of maternal depression sy mptomatology at first grade. Table 13 below summarizes the impact of each study inclusi on criteria on the final study sample size, highlighting both the number and percent of pa rticipants lost across each application of inclusion criteria, as well as the number and percent of participants lost from the total 1364 infants and families enrolled in the NICHD SECC during Phase I. Table 13 Impact of Study Inclusion Crite ria on Final Study Sample Size Inclusion Criteria Beginning Sample Size Ending Sample Size N (%) Lost N (%) Lost from Phase I 1 1364 1223 141 (10.34%) 141 (10.34%) 2 1223 819 404 (33.03%) 545 (39.96%) 3 819 679 140 (17.09%) 685 (50.22%) As already discussed in chapter two, pages 125-127, families that participated in both Phase I and Phase II of the NICHD SECC significantly differed from families lost to attrition. Results from bi variate analyses revealed that non-white mothers, mothers with less education, mothers w ith no husband or partner li ving at home at one month

PAGE 198

176 postpartum, families on public assistance at one month postpartum, mothers not breastfeeding at one month postpartum, and younger mothers, were more likely to drop out of the study during Phase II. Further, of the 1223 families participa ting in both Phase I and Phase II of the NICHD SECC, only 819 (66.97%) had complete data on all variables of interest for the present study, the majority ( 88.37%) of which were missing data on maternal report of child social competence at first grade, ma ternal report of child display of problem behaviors at first grade, or maternal depression at one mont h postpartum – all critical variables for this study. Biva riate analyses conducted to understand the differences between study participants with complete data versus those children and families excluded due to incomplete data revealed that non-participants were more likely to have non-White mothers, younger mothers, mother s with less educati on, mothers with no husband or partner living at home, mothers not breastfeeding at one month postpartum, mothers experiencing higher levels of de pressive symptomato logy at one month postpartum, and to be on public assistance. No significant differences were detected by child sex or maternal ethnicity. Table 14 summarizes findings.

PAGE 199

177 Table 14 Comparison of Study Participants with Co mplete Data Versus Excluded Children and Families with Incomplete Data Chi-Square Tests of Independence % Excluded ChiSquare(df) Pattern of Finding Child Sex Male 34.76 1.76(1) Female 31.19 Not Significant Maternal Race White 30.01 Black/Afro-American 50.34 Other 43.64 26.67(2)* Non-white mothers were more likely to have incomplete data and be excluded from the study sample. Maternal Ethnicity Non-Hispanic 32.68 1.60(1) Hispanic 41.18 Not Significant Maternal Education High School or Less 45.23 Some College 35.06 Bachelor’s Degree 20.75 Postgraduate Work 22.04 53.66(3)* Mothers will less education were more likely to have incomplete data and be excluded from the study sample. Partner/Husband at Home No 53.89 Yes 29.73 38.04(1)* Mothers with no husband or partner living at home at one month were more likely to have incomplete data and be excluded from the study sample. Receipt of Public Assistance No 30.05 Yes 47.60 24.03(1)* Families on public assistance were more likely to have incomplete data and be excluded from the study sample. Breastfeeding at One Month No 43.15 38.57(1)* Yes 26.13 Mothers not breastfeeding at one month were more likely to have incomplete data and be excluded from the study sample. Continued on the next page

PAGE 200

178 Table 14 (Continued) One-Way ANOVAs Mean (SD) F (df) Pattern of Finding Maternal Age Incomplete Data 26.47 (5.70) 66.33(1)* Complete Data 29.18 (5.35) Younger mothers were more likely to have incomplete data and be excluded from the study sample. Maternal Depression at One Month Incomplete Data 12.44 (9.94) 10.31(1)* Complete Data 10.68 (8.53) Mothers with higher levels of depression symptomatology at one month postpartum were more likely to have incomplete data and be excluded from the study sample. Indicates statistical significance at p<.05. In addition, of the 819 familie s participating in both Ph ase I and Phase II of the NICHD SECC and having complete data on a ll variables on intere st, 679 (82.90%) had non-depressed mothers at first grade. Biva riate analyses conduc ted to understand the differences between study participants with non-depressed mothers at grade one versus those excluded due to maternal ratings of elevated depressive symptoms at grade one revealed that those excluded from the study sample were more likely to have mothers with less education, mothers w ith no husband or partner li ving at home, mothers not breastfeeding at one month postpartum, mothers with elevated levels of depression at one month postpartum, and to be on public assi stance. No significant differences were detected by child sex, maternal race, matern al ethnicity, or maternal age. Table 15 summarizes these findings.

PAGE 201

179 Table 15 Comparison of Study Participants with N on-Depressed Mothers at Grade One Versus Excluded Children and Families with Depressed Mothers at Grade One Chi-Square Tests of Independence % Excluded ChiSquare(df) Pattern of Finding Child Sex Male 17.68 0.20(1) Female 8.18 Not Significant Maternal Race White 16.62 Black/Afro-American 23.61 Other 12.90 2.66(2) Not Significant Maternal Ethnicity Non-Hispanic 16.60 3.66(1) Hispanic 30.00 Not Significant Maternal Education High School or Less 27.36 Some College 19.39 Bachelor’s Degree 11.43 31.33(3)* Postgraduate Work 6.90 Mothers will less education were more likely to have elevated levels of depressive symptoms at grade one and be excluded from the study sample. Partner/Husband at Home No 35.06 Yes 15.23 19.37(1)* Mothers with no husband or partner living at home at one month were more likely to have elevated levels of depressive symptoms at grade one and be excluded from the study sample. Receipt of Public Assistance No 15.21 Yes 29.36 13.34(1)* Families on public assistance were more likely to have mothers with elevated levels of depressive symptoms at grade one and be excluded from the study sample. Continued on the next page

PAGE 202

180 Table 15 (Continued) Chi-Square Tests of Independence % Excluded ChiSquare(df) Pattern of Finding Breastfeeding at One Month No 22.70 Yes 14.15 9.52(1)* Mothers not breastfeeding at one month were more likely to have elevated levels of depression at grade one and be excluded from the study sample. Maternal Depression at One Month No 11.29 64.50(1)* Yes 36.32 Mothers with elevated levels of depression at one month postpartum were more likely to have elevated levels of depression at grade one and be excluded from the study sample. One-Way ANOVAs Mean (SD) F (df) Pattern of Finding Maternal Age No Depression at Grade 1 29.31 (5.18) 2.43(1) Depression at Grade 1 28.54 (6.08) Not Significant *Indicates significance at p<.05. Univariate Analyses Of the 679 children who participated in th e study, 50.07% were male. Nearly half (45.21%) of participants were the first child born into a fa mily, another 39.76% being the second born child. The majority of moth ers were White (87.92%), non-Hispanic (96.91%), and living with a husband or part ner at one month postpartum (92.64%). Nearly one-third (31.22%) of mothers ha d completed some college, earned an Associate’s degree or comple ted vocational school, 27.39% had completed a Bachelor’s degree, 21.50% had completed high school/G ED or less, and 19.88% had completed some type of postgraduate work. Most families (88.66%) were not receiving public assistance, and over half (67.89%) of mothers co ntinued to breastfeed their infants at one

PAGE 203

181 month postpartum. Table 16 summarizes the sociodemographic characteristics of the study sample. In addition, the mean ag e of mothers was 29.31 years (standard deviation=5.18). Table 16 Sociodemographic Characteristics of Study Sample (n=679) Characteristic N Percent Child Sex Male 340 50.07 Female 339 49.93 Child Birth Order 1 307 45.21 2 270 39.76 3 73 10.75 4 or more 29 4.27 Maternal Race White 597 87.92 Black 55 8.10 Other 27 3.98 Maternal Ethnicity Non-Hispanic 658 96.91 Hispanic 21 3.09 Presence of Husband/Partner in Home No 50 7.36 Yes 629 92.64 Maternal Education High School/GED or less 146 21.50 Some college 212 31.22 Bachelor’s degree 186 27.39 Postgraduate work 135 19.88 Family Receipt Public Assistance No 602 88.66 Yes 77 11.34 Mother Breastfeeding at One Month No 218 32.11 Yes 461 67.89

PAGE 204

182 Table 17 summarizes the means, standa rd deviations, and ranges for each continuous dependent variable under investigation. Summary statistics are presented for each raw subscale social competence score (from both mothers and teachers), standardized total social competence scor es (from both mothers and teachers), and standardized problem behavior scores for in ternalizing problem behaviors, externalizing problem behaviors, and to tal problem behaviors. Table 17 Means, Standard Deviations, and Ranges for Dependent Variables Characteristic Mean Standard Deviation Range Social Competence – Maternal Report Cooperation 12.89 3.01 3-20 Assertion 17.37 2.26 8-20 Responsibility 13.97 2.64 5-20 Self-Control 13.24 3.30 4-20 Peer Competence 15.89 2.58 6-20 Total Standardized Score 106.33 14.50 67-130 Social Competence – Teacher Report Cooperation 15.93 3.84 2-20 Assertion 13.56 3.73 0-20 Self Control 15.42 3.54 3-20 Peer Competence 15.55 3.47 4-20 Total Standardized Score 104.62 13.22 64-130 Display of Problem Behaviors Internalizing Problem Behaviors 47.36 8.51 33-73 Externalizing Problem Behaviors 47.56 9.40 30-78 Total Problem Behaviors 46.75 9.24 24-74 Table 18 summarizes the categorical classi fications of child social competence and display of problem behavi ors among study participants. First, total standardized

PAGE 205

183 social competence scores (for both mother s and teachers) were dichotomized into behavioral level categories using reco mmended cut points established by SSRS researchers. Children scoring below 85 were cl assified as having less than average social competence, and children scoring between 85-115, or above 115 were classified as having average or greater than average displa y of social competence (Gresham & Elliott, 1990) Among study participants, mothers ra ted 47 (6.92%) children as having less than average social competence, and teachers ra ted 47 (6.92%) children as having less than average social competence. Of children rated as having less than average social competence by either their mother or t eacher (n=85), 9 (10.59%) were rated by both mothers and teachers as having less than aver age social competence. The remaining 76 (89.41%) children were only rated as having less than average social competence by a single individual (either mo ther or teacher). Further, children were classified in to deviant and non-deviant groups based on standardized scores of total problem behavi ors (including internalizing, externalizing and other problem behaviors), as reported by mothers. A recommended cut value of 60 was used to dichotomize children into subgroups (Achenbach, 19 91). Children scoring below 60 were classified as displaying non-deviant or normal levels of problem behaviors. Children scoring at or above 60 were classifi ed as displaying borde rline or clinically significant levels of problem behaviors. Among study participants, mothers rated 57 (8.39%) children as displaying borderline or clinically significant levels of problem behaviors.

PAGE 206

184 Table 18 Categorical Classification s of Child Social Competence and Display of Problem Behaviors N % Social Competence – Maternal Report Less than Average 47 6.92 Average or Greater 632 93.08 Social Competence – Teacher Report Less than Average 47 6.92 Average of Greater 632 93.08 Display of Problem Behaviors Normal/Non-Deviant 622 91.61 Borderline or Clinical ly Significant/Deviant 57 8.39 Frequencies of the categorical maternal de pression variables, a nd means, standard deviations, and ranges for the continuous ma ternal depression vari ables are provided in Table 19. During the first three years postp artum, over one-third (37.41%) of the study sample had mothers who reported elevated leve ls of depressive symptoms at some point, with the highest prevalence of depression o ccurring at one month postpartum (17.82%). Rates of self-reported depression declined du ring both the sixand 15-month assessments to 12.81% and 10.75% respectively, increasing to 11.93% at the 24-month assessment. Maternal depression declined to 11.19% at the 36-month assessment.

PAGE 207

185 Table 19 Univariate Statistics for Maternal Depression Variables Study Sample (N=697) N (%) Mean (SD) Range Maternal Depression Depression at One Month -9.48 (7.79) 0-44 Yes 121 (17.82) --Depression at Six Months -7.66 (7.15) 0-44 Yes 87 (12.81) --Depression at 15 Months -7.42 (6.70) 0-44 Yes 73 (10.75) --Depression at 24 Months -7.66 (6.81) 0-50 Yes 81 (11.93) --Depression at 36 Months -7.46 (6.82) 0-41 Yes 76 (11.19) --Ever Depressed ---Yes 254 (37.41) --Means, standard deviations, and ranges fo r levels of maternal sensitivity at six, 15, 24, and 36 months, are provided in Table 20. At six, 15, and 24 months postpartum, mean levels of maternal sensitivity as asse ssed via direct observa tion of mother-child interaction were 3.17, 3.22, and 3.23, demonstrating th at on average, maternal displays of responsivity to child nondistress, positive regard for the child, and unintrusiveness were moderate. A different rating s cale was employed at 36 months resulting in a mean level of maternal sensitivity of 5.92, indicating m oderately high to high levels of observed maternal sensitivity, operationalized as mate rnal supportive presence, maternal respect for child autonomy, and non-hostili ty toward to the child.

PAGE 208

186 Table 20 Univariate Statistics for Maternal Sensitivity Variables Study Sample (N=697) Mean (SD) Range Maternal Sensitivity (MS) MS at Six Months 3.17 (0.56) 1.00-4.00 MS at 15 Months 3.22 (0.50) 1.33-4.00 MS at 24 Months 3.23 (0.54) 1.33-4.00 MS at 36 Months 5.92 (0.77) 2.33-7.00 Frequencies of exposure to nonmaternal ca re at one, six, 15, 24, and 36 months, as well as the percent of infants ever expos ed to nonmaternal care at each time point are provided in Table 21. At one month postpartu m, the majority (93.23%) of infants were cared for by their mothers, with only 6.77% of infants exposed to nonmaternal care. The percent of children in nonmat ernal care dramatically incr eased to 67.60% by six months postpartum, and continued to rise at 15 months, 24 months, and 36 months to 73.34%, 75.70%, and 83.80% respectively. Similarly, the percent of infants ever exposed to nonmaternal care noticeably increased between one and six months postpartum from 6.77% to 68.92%. Ever exposur e to nonmaternal care contin ued to rise steadily and reaching 91.90% by 36 months postpartum.

PAGE 209

187 Table 21 Univariate Statistics for Exposur e to Nonmaternal Care Variables Study Sample (N=697) N (%) Exposure to Nonmaternal Care Ever Exposed to Nonmaternal Care Exposure to Nonmaternal Care Exposure at One Month Yes 46 (6.77) 46 (6.77) Exposure at Six Months Yes 459 (67.60) 468 (68.92) Exposure at 15 Months Yes 498 (73.34) 547 (80.56) Exposure at 24 Months Yes 514 (75.50) 581 (85.57) Exposure at 36 Months Yes 569 (83.80) 624 (91.90) Frequencies of exposure to center-based nonmaternal care at one, six, 15, 24, and 36 months, as well as the percent of infants ever exposed to center-based nonmaternal care at each time point are provided in Table 22. As with exposure to nonmaternal care in general, exposure to center-based nonma ternal care increased over time; with the largest escalations occurri ng between one and six months postpartum (from 0.29% to 10.31%) and between 24 and 36 months postpar tum (19.59% to 32.11%). Similarly, the percent of infants ever exposed to center-b ased nonmaternal care increased over time with over one-third (36.82%) of infants exposed at some point to center-based nonmaternal care by 36 months postpartum.

PAGE 210

188 Table 22 Univariate Statistics for Exposure to Ce nter-Based Nonmaternal Care Variables Study Sample (N=697) N (%) Exposure to Center-Based Nonmaternal Care Ever Exposed to Center-Based Nonmaternal Care Exposure to Center-Based Nonmaternal Care Exposure at One Month Maternal Care Only 633 (93.23) 633 (93.23) Non-Center-Based Care 44 (6.48) 44 (6.48) Center-Based Care 2 (0.29) 2 (0.29) Exposure at Six Months Maternal Care Only 220 (32.40) 211 (31.08) Non-Center-Based Care 389 (57.29) 396 (58.32) Center-Based Care 70 (10.31) 72 (10.60) Exposure at 15 Months Maternal Care Only 181 (26.66) 132 (19.44) Non-Center-Based Care 410 (60.38) 445 (65.54) Center-Based Care 88 (12.96) 102 (15.02) Exposure at 24 Months Maternal Care Only 165 (24.30) 98 (14.43) Non-Center-Based Care 381 (56.11) 422 (62.15) Center-Based Care 133 (19.59) 159 (23.42) Exposure at 36 Months Maternal Care Only 110 (16.20) 55 (8.10) Non-Center-Based Care 351 (51.69) 374 (55.08) Center-Based Care 218 (32.11) 250 (36.82)

PAGE 211

189 Bivariate Analyses Chi-square tests were used to estimat e associations between sociodemographic control variables (all but maternal age) and each dichotomously scored maternal depression variable including ever depresse d, depressed at one month, depressed at six months, depressed at 15 months depressed at 24 months, a nd depressed at 36 months. Results are summarized in Tables 23-28. In ge neral, family receipt of public assistance was the only sociodemographic control variable significantly associated with maternal depression at all time points, with families on public assistance at one month postpartum being more likely to have mothers who had ev er been depressed, were depressed at one month postpartum, and were depressed at six, 15, 24, and 36 months postpartum. Furthermore, maternal education, presence of partner or husband at home at one month postpartum, and breastfeeding at one month postpartum were a ll statistically associated with maternal depression ever during the firs t 36 months postpartum, as well as at one, six, 15, and 24 months postpartum. Specifical ly, mothers with less education, mothers with no husband or partner li ving at home, and mothers not breastfeeding were more likely to have ever been depressed, as well as depressed at one, six, 15, and 24 months postpartum. Black/Afro-American mothers were more likely to have ever been depressed, depressed at one month postpartum, as well as depressed at 24 months postpartum. Child birth order was also asso ciated with maternal depression at one and six months, with mothers with more childre n being more likely to exhibit elevated symptoms of depression. Neither child se x nor maternal ethnicity were found to be significantly associated with maternal depression.

PAGE 212

190 Table 23 Chi-Square Results of Ever Depressed by Sociodemographic Control Variables Ever Depressed % Yes ChiSquare(df) Pattern of Finding Child Sex Male 35.88 0.68(1) Female 38.94 Not Significant Child Birth Order 1 35.50 4.55(3) 2 38.15 3 35.62 4 or More 55.17 Not Significant Maternal Race White 35.34 15.27(2) Black/Afro-American 61.82 Other 33.33 Black/Afro-American mothers were more likely to have ever been depressed. Maternal Ethnicity Non-Hispanic 37.23 0.27(1) Hispanic 42.86 Not Significant Maternal Education High School or Less 48.63 16.90(3)* Some College 41.04 Bachelor’s Degree 29.03 Postgraduate Work 31.11 Mothers will less education were more likely to have ever been depressed. Partner/Husband at Home No 64.00 16.30(1)* Yes 35.29 Mothers with no husband or partner living at home at one month were more likely to have ever been depressed. Receipt of Public Assistance No 34.39 20.71(1)* Yes 61.04 Families on public assistance at one month were more likely to have mothers who had ever been depressed. Breastfeeding at One Month No 51.38 26.76(1)* Yes 30.80 Mothers not breastfeeding at one month were more likely to have ever been depressed. *Indicates statistical significance at p<.05.

PAGE 213

191 Table 24 Chi Square Results of Maternal Depr ession at One Month Postpartum by Sociodemographic Control Variables Depressed at One Month Postpartum % Yes ChiSquare(df) Pattern of Finding Child Sex Male 19.41 1.18(1) Female 16.22 Not Significant Child Birth Order 1 18.57 9.79(3) 2 14.81 3 17.81 4 or More 37.93 Mothers with more children were more likely to be depressed at one month. Maternal Race White 16.42 9.16(2) Black/Afro-American 32.73 Other 18.52 Black/Afro-American mothers were more likely to be depressed at one month. Maternal Ethnicity Non-Hispanic 17.93 Hispanic 14.29 Fisher’s Exact Test Not Significant Maternal Education High School or Less 23.29 4.95(3) Some College 18.40 Bachelor’s Degree 15.59 Postgraduate Work 14.07 Not Significant Partner/Husband at Home No 36.00 12.18(1)* Yes 16.38 Mothers with no husband or partner living at home at one month were more likely to be depressed at one month. Receipt of Public Assistance No 15.61 17.64(1)* Yes 35.06 Families on public assistance at one month were more likely to have mothers who were depressed at one month. Breastfeeding at One Month No 27.98 22.64(1)* Yes 13.02 Mothers not breastfeeding at one month were more likely to be depressed at one month. *Indicates statistical significance at p<.05.

PAGE 214

192 Table 25 Chi-Square Results of Maternal Depre ssion at Six Months Postpartum by Sociodemographic Control Variables Depressed at Six Months Postpartum % Yes ChiSquare(df) Pattern of Finding Child Sex Male 10.88 2.27(1) Female 14.75 Not Significant Child Birth Order 1 9.12 20.65(3) 2 13.70 3 15.07 4 or More 37.93 Mothers with more children were more likely to be depressed at six months. Maternal Race White 11.39 Black/Afro-American 29.09 Other 11.11 Fisher’s Exact Test Black/Afro-American mothers were more likely to be depressed at six months. Maternal Ethnicity Non-Hispanic 12.77 Hispanic 14.29 Fisher’s Exact Test Not Significant Maternal Education High School or Less 20.55 11.79(3) Some College 12.26 Bachelor’s Degree 11.29 Postgraduate Work 7.41 Mothers with less education were more likely to be depressed at six months. Partner/Husband at Home No 28.00 11.14(1)* Yes 11.61 Mothers with no husband or partner living at home at one month were more likely to be depressed at six months. Receipt of Public Assistance No 10.96 16.25(1)* Yes 27.27 Families on public assistance at one month were more likely to have mothers who were depressed at six months. Breastfeeding at One Month No 17.89 7.40(1)* Yes 10.41 Mothers not breastfeeding at one month were more likely to be depressed at six months. *Indicates statistical significance at p<.05.

PAGE 215

193 Table 26 Chi-Square Results of Maternal Depre ssion at 15 Months Postpartum by Sociodemographic Control Variables Depressed at 15 Months Postpartum % Yes ChiSquare(df) Pattern of Finding Child Sex Male 8.82 2.63(1) Female 12.68 Not Significant Child Birth Order 1 9.45 3.72(3) 2 10.74 3 12.33 4 or More 20.69 Not Significant Maternal Race White 10.55 Black/Afro-American 14.55 Other 7.41 Fisher’s Exact Test Not Significant Maternal Ethnicity Non-Hispanic 10.64 Hispanic 14.29 Fisher’s Exact Test Not Significant Maternal Education High School or Less 17.81 12.32(3) Some College 11.32 Bachelor’s Degree 8.06 Postgraduate Work 5.93 Mothers with less education were more likely to be depressed at 15 months. Partner/Husband at Home No 26.00 13.08(1)* Yes 9.54 Mothers with no husband or partner living at home at one month were more likely to be depressed at 15 months. Receipt of Public Assistance No 8.64 24.71(1)* Yes 27.27 Families on public assistance at one month were more likely to have mothers who were depressed at 15 months. Breastfeeding at One Month No 16.97 12.95(1)* Yes 7.81 Mothers not breastfeeding at one month were more likely to be depressed at 15 months. *Indicates statistical significance at p<.05.

PAGE 216

194 Table 27 Chi-Square Results of Maternal Depre ssion at 24 Months Postpartum by Sociodemographic Control Variables Depressed at 24 Months Postpartum % Yes ChiSquare(df) Pattern of Finding Child Sex Male 11.76 0.02(1) Female 12.09 Not Significant Child Birth Order 1 12.05 2 12.59 3 8.22 4 or More 13.79 Fisher’s Exact Test Not Significant Maternal Race White 10.72 Black/Afro-American 23.64 Other 14.81 Fisher’s Exact Test Black/Afro-American mothers were more likely to be depressed at 24 months. Maternal Ethnicity Non-Hispanic 11.70 Hispanic 19.05 Fisher’s Exact Test Not Significant Maternal Education High School or Less 21.23 26.30(3) Some College 15.09 Bachelor’s Degree 5.91 Postgraduate Work 5.19 Mothers with less education were more likely to be depressed at 24 months. Partner/Husband at Home No 26.00 10.17(1)* Yes 10.81 Mothers with no husband or partner living at home at one month were more likely to be depressed at 24 months. Receipt of Public Assistance No 10.47 10.83(1)* Yes 23.38 Families on public assistance at one month were more likely to have mothers who were depressed at 24 months. Breastfeeding at One Month No 16.97 7.77(1)* Yes 9.54 Mothers not breastfeeding at one month were more likely to be depressed at 24 months. *Indicates statistical significance at p<.05.

PAGE 217

195 Table 28 Chi-Square Results of Maternal Depre ssion at 36 Months Postpartum by Sociodemographic Control Variables Depressed at 36 Months Postpartum % Yes ChiSquare(df) Pattern of Finding Child Sex Male 9.41 2.17(1) Female 12.98 Not Significant Child Birth Order 1 12.05 2 11.48 3 8.22 4 or More 6.90 Fisher’s Exact Test Not Significant Maternal Race White 11.22 Black/Afro-American 14.55 Other 3.70 Fisher’s Exact Test Not Significant Maternal Ethnicity Non-Hispanic 11.09 Hispanic 14.29 Fisher’s Exact Test Not Significant Maternal Education High School or Less 15.07 3.13(3) Some College 10.85 Bachelor’s Degree 10.22 Postgraduate Work 8.89 Not Significant Partner/Husband at Home No 16.00 1.25(1) Yes 10.81 Not Significant Receipt of Public Assistance No 10.13 6.00(1)* Yes 19.48 Families on public assistance at one month were more likely to have mothers who were depressed at 36 months. Breastfeeding at One Month No 11.93 0.17(1) Yes 10.85 Not Significant *Indicates statistical significance at p<.05.

PAGE 218

196 To test the associations between mate rnal age and each dichotomously scored maternal depression variable, one-way ANOVAs were computed. Results are summarized in Table 29 below. Maternal age was statistically associated with maternal depression at each time point including one, six, 15, 24, and 36 months postpartum, with younger mothers having higher levels of self-re ported depressive symptoms. Mothers who had ever been depressed during the first 36 months postpartum were also found to be significantly younger in age. Table 29 ANOVA Results of Maternal Depression Scores Over Time by Maternal Age Mean (SD) F Ever Depressed No 29.73 (4.79) 7.63* Yes 28.60 (5.71) Depressed at One Month No 29.51 (4.99) 4.94* Yes 28.36 (5.90) Depressed at Six Months No 29.50 (5.05) 6.54* Yes 27.99 (5.82) Depressed at 15 Months No 29.46 (5.04) 5.04* Yes 28.03 (6.09) Depressed at 24 Months No 29.64 (5.02) 20.89* Yes 26.88 (5.70) Depressed at 36 Months No 29.47 (5.17) 5.51* Yes 28.00 (5.03) *Indicates statistical significance at p<.05.

PAGE 219

197 One-way ANOVAs were computed to identify associ ations between sociodemographic control variables (all but maternal age) and each continuously scored maternal depression variable including de pression at one month, depression at six months, depression at 15 months depression at 24 months, a nd depression at 36 months. Results are summarized in Table 30. Significant effects on maternal depression across one, six, 15, 24, and 36 months postpartum were detected for maternal race, maternal education, presence of husband or partner in the home, family receipt of public assistance, and maternal breastfeeding. Mo re specifically, Black or Afro-American mothers, mothers with less education, mother s with no husband or partner living at home at one month postpartum, mothers receiving pu blic assistance at one month postpartum, and mothers not breastfeeding at one month pos tpartum, were found to have significantly higher levels of maternal depression. Child sex had a significant effect on maternal depression scores at six and 15 months, with mothers of females reporting a higher level of depressive symptoms. Child birth order wa s also found to have a significant impact on maternal depression, but only at one and six mo nths postpartum. Mothers with four or more children exhibited signifi cantly higher levels of matern al depression at one and six months postpartum. There were no signifi cant differences detected in maternal depression scores between Hispan ic and non-Hispanic mothers.

PAGE 220

198 Table 30 ANOVA Results of Maternal Depr ession Scores Over Time by Soc iodemographic Control Variables Maternal Depression One Month Six Months 15 Months 24 Months 36 Months Mean (SD) F Mean (SD) F Mean (SD) F Mean (SD) F Mean (SD) F Child Sex Male 9.62 (8.10) 0.24 6.99 (6.39) 5.95* 7.03 (6.64) 2.27 7.18 (6.90) 3.49 6.93 (6.34) 4.12* Female 9.33 (7.48) 8.33 (7.79) 7.80 (6.74) 8.15 (6.79) 7.99 (7.24) Child Birth Order 1 9.77 (7.96) 3.78* 7.10 (6.42) 5.75* 7.41 (6.46) 1.32 7.71 (6.73) 0.57 7.59 (6.32) 0.40 2 8.82 (7.28) 7.67 (7.34) 7.22 (6.60) 7.41 (6.40) 7.48 (7.24) 3 8.97 (7.69) 7.91 (6.67) 7.22 (7.34) 7.82 (8.31) 6.67 (6.79) 4 or More 13.72 (9.64) 12.79 (11.17) 9.79 (8.22) 9.10 (7.26) 7.86 (6.12) Continued on the next page

PAGE 221

199 Table 30 (Continued) Maternal Depression One Month Six Months 15 Months 24 Months 36 Months Mean (SD) F Mean (SD) F Mean (SD) F Mean (SD) F Mean (SD) F Maternal Race White 9.14 (7.68) 6.01* 7.22 (6.84) 12.67* 7.22 (6.77) 3.02* 7.20 (6.42) 15.16* 7.26 (6.95) 4.19* Black/Afro-American 12.90 (8.77) 12.21 (9.18) 9.53 (6.46) 12.34 (8.79) 9.97 (5.92) Other 9.96 (6.78) 8.00 (6.13) 7.56 (4.67) 8.48 (7.23) 6.70 (4.40) Maternal Ethnicity Non-Hispanic 9.48 (7.84) 0.00 7.66 (7.21) 0.00 7.38 (6.73) 0.67 7.61 (6.76) 1.30 7.42 (6.83) 0.73 Hispanic 9.52 (6.55) 7.67 (5.10) 8.59 (5.46 9.33 (8.08) 8.71 (6.39) Maternal Education High School or Less 11.73 (8.70) 5.87* 10.18 (8.47) 9.68* 9.59 (7.14) 9.53* 10.43 (8.66) 15.54* 9.54 (7.56) 6.76* Some College 9.40 (7.39) 7.69 (7.08) 7.83 (6.92) 8.22 (6.77) 7.43 (6.31) Bachelor’s Degree 8.59 (7.38) 6.88 (6.67) 6.12 (6.36) 6.07 (5.45) 6.73 (6.55) Postgraduate Work 8.39 (7.52) 5.96 (5.47) 6.20 (5.61) 5.99 (5.02) 6.27 (6.69) Continued on the next page

PAGE 222

200 Table 30 (Continued) Maternal Depression One Month Six Months 15 Months 24 Months 36 Months Mean (SD) F Mean (SD) F Mean (SD) F Mean (SD) F Mean (SD) F Partner/Husband at Home No 12.83 (10.12) 10.14* 11.70 (8.87) 17.65* 10.90 (8.24) 14.93* 12.97 (9.49) 34.39* 10.96 (7.30) 14.48* Yes 9.21 (7.53) 7.34 (6.90) 7.14 (6.49) 7.24 (6.37) 7.18 (6.71) Receipt of Public Assistance No 8.80 (7.19) 43.02* 7.12 (6.75) 31.94* 6.95 (6.39) 26.36* 7.15 (6.29) 32.34* 7.03 (6.63) 21.85* Yes 14.80 (10.02) 11.90 (8.66) 11.04 (7.92) 11.73 (9.03) 10.83 (7.40) Breastfeeding at One Month No 11.88 (8.78) 31.98* 9.29 (7.87) 17.01* 8.35 (7.26) 6.35* 9.13 (7.95) 15.20* 8.53 (7.17) 7.99* Yes 8.34 (7.01) 6.89 (6.66) 6.97 (6.38) 6.97 (6.08) 6.96 (6.60) *Indicates statistical significance at p<.05.

PAGE 223

201 Pearson correlation coefficients were computed to assess the relationships between maternal age and maternal depressi on scores at one, six, 15, 24, and 36 months postpartum. Correlations ranged from –0.14 to –0.21. All correlations were negative indicating that as maternal age increased, levels of depression decreased. Although correlations were significant, associations were weak with maternal age having the greatest association w ith maternal depression scores at 24 months postpartum (r=-0.21). One-way ANOVAs were computed to identify associ ations between sociodemographic control variables (all but ma ternal age) and maternal sensitivity over time including maternal sensitivity at six, 15, 24, and 36 months postpartum. Results are summarized in Table 31. Significant effect s on maternal sensitivity across six, 15, 24, and 36 months were detected for maternal race, maternal education, presence of husband or partner in the home, family receipt of pub lic assistance, and maternal breastfeeding. More specifically, Black or Afro-America n mothers, mothers with less education, mothers with no husband or partner living at home at one month postpartum, mothers receiving public assistance at one month pos tpartum, and mothers not breastfeeding at one month postpartum, were found to have significantly lower levels of maternal sensitivity at six, 15, 24, and 36 months pos tpartum. Child sex was significantly associated with maternal sensitivity at 24 months only, with mothers of males scoring lower in observed levels of maternal sensitivit y. There were no signi ficant differences in levels of maternal sensitivity at six, 1 5, 24, and 36 months between Hispanic and nonHispanic mothers, nor were differences detected by child birth order.

PAGE 224

202 Table 31 ANOVA Results of Maternal Sens itivity Scores Over Time by Sociodemographic Control Variables Maternal Sensitivity Six Months 15 Months 24 Months 36 Months Mean (SD)F Mean (SD)F Mean (SD)F Mean (SD)F Child Sex Male 3.18 (0.58)0.06 3.20 (0.50)1.19 3.19 (0.56)3.92* 5.89 (0.78)1.06 Female 3.17 (0.54) 3.24 (0.50) 3.27 (0.51) 5.95 (0.76) Child Birth Order 1 3.17 (0.60)1.06 3.21 (0.52)2.78 3.21 (0.52)0.81 5.94 (0.77)0.73 2 3.20 (0.51) 3.27 (0.45) 3.17 (0.55) 5.94 (0.74) 3 3.14 (0.52) 3.16 (0.53) 3.19 (0.54) 5.85 (0.81) 4 or More 3.01 (0.66) 3.03 (0.53) 3.16 (0.57) 5.76 (0.94) Maternal Race White 3.21 (0.53)17.03* 3.29 (0.44)53.51* 3.27 (0.51)19.74* 5.98 (0.77)21.97* Black/Afro-American 2.80 (0.63) 2.62 (0.66) 2.82 (0.65) 5.29 (0/91) Other 2.94 (0.70) 3.06 (0.51) 3.12 (0.45) 5.81 (0.72) Maternal Ethnicity Non-Hispanic 3.18 (0.56)2.83 3.23 (0.50)1.08 3.23 (0.54)0.12 5.92 (0.77)0.33 Hispanic 2.97 (0.57) 3.11 (0.38) 3.27 (0.44) 5.83 (0.73) Maternal Education High School or Less 2.84 (0.61)31.43* 2.95 (0.55)25.45* 2.97 (0.58) 19.34* 5.43 (0.88)35.06 Some College 3.15 (0.53) 3.21 (0.51) 3.21 (0.54) 5.90 (0.77) Bachelor’s Degree 3.28 (0.50) 3.32 (0.43) 3.32 (0.46) 6.12 (0.60) Postgraduate Work 3.41 (0.45) 3.40 (0.37) 3.40 (0.45) 6.21 (0.55) Continued on the next page

PAGE 225

203 Table 31 (Continued) Maternal Sensitivity Six Months 15 Months 24 Months 36 Months Mean (SD)F Mean (SD)F Mean (SD)F Mean (SD)F Partner/Husband at Home No 2.86 (0.66) 16.99* 2.79 (0.63)43.64* 2.93 (0.63)16.95* 5.45 (1.07)20.51* Yes 3.20 (0.54) 3.26 (0.47) 3.25 (0.52) 5.96 (0.73) Receipt of Public Assistance No 3.22 (0.53)42.04* 3.28 (0.45)79.22* 3.27 (0.52)38.45* 6.00 (0.71)57.34* Yes 2.79 (0.66) 2.77 (0.62) 2.88 (0.56) 5.32 (0.93) Breastfeeding at One Month No 2.97 (0.60)42.59* 3.03 (0.58)53.62* 3.06 (0.61)33.39* 5.67 (0.84)35.47* Yes 3.26 (0.51) 3.32 (0.43) 3.31 (0.48) 6.04 (0.71) *Indicates statistical significance at p<.05.

PAGE 226

204 Pearson correlation coefficients were computed to assess the relationships between maternal age and maternal sens itivity at one, six, 15, 24, and 36 months postpartum. Correlations ranged from 0.17 to 0.29. All correlations were positive indicating that as maternal age increased, leve ls of maternal sensitivity also increased. Although correlations were significant, associ ations were low with maternal age having the greatest association with maternal sensitivity at 36 mont hs postpartum (r=0.29). Chi-square tests were used to identif y associations between sociodemographic control variables (all but maternal age) and each dichotomously scored dependent variable including maternal report of child social competence, teacher report of child social competence, and mother report of child display of problem behaviors. Results are summarized in Tables 32-34. In general, both mother and teacher re ports of child social competence at first grade were associated with child sex, maternal race, maternal education, and family receipt of public assi stance. Both mother s and teachers rated female children as having less than average social competence at first grade. Furthermore, children of Black/Afro-American mothers, mothers with less education, and children from families receiving public assi stance at one month postpartum were more likely to be rated by both mothers and teach ers are having less than average social competence. In addition, mothers were more likely to rate their children as having less than average social competence if (1) they had more than one child; (2) they had no husband or partner living at home at one month postpartum; and (3) they were not breastfeeding at one month postpartum. Ma ternal ethnicity was not associated with either maternal or teacher reports of ch ild social competence at first grade.

PAGE 227

205 Related to maternal reports of child disp lay of problem behaviors at first grade, the only sociodemographic cont rol variable found to be st atistically associated was family receipt of public assistance at one month postpartum. Families on public assistance were more likely to have mothers who rated their childre n as having borderline or clinically significant/deviant levels of problem behaviors.

PAGE 228

206 Table 32 Maternal Report of Child Social Competen ce by Sociodemographic Control Variables Less than Average Social Competence % Yes ChiSquare(df) Pattern of Finding Child Sex Male 4.71 5.19(1) Female 9.14 Female children were more likely to be rated by mothers as having less than average social competence. Child Birth Order 1 3.58 2 9.26 3 6.85 4 or More 20.69 16.13(3) Children lower in birth order were more likely to be rated by mothers as having less than average social competence. Maternal Race White 5.36 Black/Afro-American 21.82 Other 11.11 Fisher’s Exact Test Black/Afro-American mothers were more likely to rate their children as having less than average social competence. Maternal Ethnicity Non-Hispanic 6.84 Hispanic 9.52 Fisher’s Exact Test Not Significant Maternal Education High School or Less 12.33 Some College 9.43 Bachelor’s Degree 3.23 Postgraduate Work 2.22 Fisher’s Exact Test Mothers with less education were more likely to rate their children as having less than average social competence. Partner/Husband at Home No 16.00 Yes 6.20 6.90(1) Mothers with no husband or partner living at home at one month were more likely to rate their children as having less than average social competence. Receipt of Public Assistance No 5.32 21.26(1)* Yes 19.48 Families on public assistance at one month were more likely to have mothers who rated their children as having less than average social competence. Breastfeeding at One Month No 13.76 23.31(1) Yes 3.69 Mothers not breast feeding at one month were more likely to rate their children as having less than average social competence. *Indicates statistical significance at p<.05.

PAGE 229

207 Table 33 Teacher Report of Child Social Competen ce by Sociodemographic Control Variables Less than Average Social Competence % Yes ChiSquare(df) Pattern of Finding Child Sex Male 5.00 3.90(1) Female 8.85 Female children were more likely to be rated by teachers as having less than average social competence. Child Birth Order 1 6.84 2 5.93 3 8.22 4 or More 13.79 Fisher’s Exact Test Not Significant Maternal Race White 5.53 Black/Afro-American 18.18 Other 14.81 Fisher’s Exact Test Teachers of students with nonWhite mothers were more likely to rate their students as having less than average social competence. Maternal Ethnicity Non-Hispanic 6.69 Hispanic 14.29 Fisher’s Exact Test Not Significant Maternal Education High School or Less 15.07 Some College 7.08 Bachelor’s Degree 3.76 Postgraduate Work 2.22 Fisher’s Exact Test Teachers of students with mothers having less education were more likely to rate their students as having less than average social competence. Partner/Husband at Home No 12.00 2.16(1) Yes 6.52 Not Significant Receipt of Public Assistance No 5.98 7.31(1)* Yes 14.29 Teachers of students on public assistance at one month were more likely to rate their students as having less than average social competence. Breastfeeding at One Month No 7.34 0.09(1) Yes 6.72 Not Significant *Indicates statistical significance at p<.05.

PAGE 230

208 Table 34 Mother Report of Child Problem Behavio rs by Sociodemographic Control Variables Borderline or Clinically Si gnificant/Deviant Display of Problem Behaviors % Yes ChiSquare(df) Pattern of Finding Child Sex Male 8.82 0.16(1) Female 7.96 Not Significant Child Birth Order 1 9.12 2 7.78 3 8.22 4 or More 6.90 Fisher’s Exact Test Not Significant Maternal Race White 7.87 Black/Afro-American 14.55 Other 7.41 Fisher’s Exact Test Not Significant Maternal Ethnicity Non-Hispanic Hispanic Unable to compute due to cell size of zero Maternal Education High School or Less 10.96 3.65(3) Some College 7.55 Bachelor’s Degree 9.68 Postgraduate Work 5.19 Not Significant Partner/Husband at Home No 12.00 0.91(1) Yes 8.l1 Not Significant Receipt of Public Assistance No 6.81 17.32(1)* Yes 20.78 Families on public assistance at one month were more likely to have mothers who rated their children as having borderline or clinically significant/deviant levels of problem behaviors. Breastfeeding at One Month No 8.26 0.01(1) Yes 8.46 Not Significant *Indicates statistical significance at p<.05.

PAGE 231

209 To test the associations between mate rnal age and each dichotomously scored dependent variable including maternal report of child social competence, teacher report of child social competence, and mother report of child display of problem behaviors, oneway ANOVAs were computed. Results are su mmarized in Table 35 below. Maternal age was statistically associated with both mo ther and teacher reports of child social competence at first grade. More specifical ly, children rated by mothers and teachers as having less than average social competen ce had statistically younger mothers, as compared to children with average or greater than average social competence. No association was detected between maternal ag e and maternal report of child display of problem behaviors at first grade. Table 35 ANOVA Results of Child Social Competen ce and Display of Problem Behaviors by Maternal Age Mean (SD) F Maternal Report of Social Competence Average or Greater 29.53 (5.11) 17.49* Less than Average 26.30 (5.27) Teacher Report of Social Competence Average or Greater 29.44 (5.13) 5.85* Less than Average 27.55 (5.59) Maternal Report of Problem Behaviors Normal/Non-Deviant 29.40 (5.04) 2.21 Borderline or Clinically Significant/Deviant 28.33 (6.47) *Indicates statistical significance at p<.05. One-way ANOVAs were computed to identify associ ations between sociodemographic control variables (all but mate rnal age) and (1) maternal report of child

PAGE 232

210 social competence; (2) teacher report of child social competen ce; and (3) maternal report of child display of problem behaviors, at firs t grade. Results are summarized in Table 36. Both mother and teacher reports of child soci al competence at first grade were associated with maternal race, maternal education, pres ence of husband or partner in the home, and family receipt of public assistance. More sp ecifically, both mothers and teachers rated children of (1) Black/Afro-American mothers; (2) mothers with less education; (3) mothers with no husband or partner living at home at one month postpartum; and (4) families receiving public assistance at one month postpartum, as exhibiting lower levels of social competence at first grade. Furtherm ore, mothers were more likely to rate female children, non-first born children, and children not breastfee d at one month postpartum, as having lower levels of social competence at first grade. No significant association between maternal ethnicity and child social competence at first grade was detected. Related to maternal reports of child disp lay of problem behaviors at first grade, maternal education and family receipt of public assistance were found to have statistically significant associations. Mo thers with less education and those receiving public assistance at one month postpartum rated their children as having higher displays of total problem behaviors. No associations were detected for child sex, child birth order, maternal race, maternal ethnicity, presence of husband or partner at home at one month postpartum, or maternal breastf eeding at one month postpartum.

PAGE 233

211 Table 36 ANOVA Results of Child Social Compet ence and Display of Problem Behaviors by Sociodemographic Control Variables Maternal Report of Social Competence Teacher Report of Social Competence Maternal Report of Problem Behaviors Mean (SD) F Mean (SD) F Mean (SD) F Child Sex Male 107.84 (13.55) 7.42* 105.17 (12.04) 1.16 46.93 (9.11) 0.25 Female 104.82 (15.27) 104.07 (14.29) 46.58 (9.37) Child Birth Order 1 108.50 (14.01) 4.49* 104.37 (13.19) 1.43 47.19 (9.14) 1.60 2 104.65 (14.49) 105.41 (12.99) 46.94 (9.18) 3 105.00 (14.98) 104.49 (13.03) 44.79 (9.69) 4 or More 102.34 (16.00) 100.24 (15.57) 45.31 (9.41) Maternal Race White 107.28 (14.17) 11.43* 105.20 (12.93) 7.75* 46.54 (9.15) 1.55 Black/Afro-American 98.33 (15.70) 97.96 (13.52) 47.85 (10.50) Other 101.59 (13.80) 105.30 (15.40) 49.26 (8.18) Maternal Ethnicity Non-Hispanic 106.41 (14.51) 0.60 104.71 (13.13) 0.95 46.72 (9.26) 0.31 Hispanic 103.90 (14.55) 101.86 (15.88) 47.86 (8.68) Maternal Education High School or Less 101.99 (14.55) 9.78* 100.77 (14.44) 10.03* 48.55 (9.39) 3.15* Some College 104.89 (15.33) 103.14 (12.98) 46.88 (9.46) Bachelor’s Degree 109.37 (12.96) 107.16 (12.61) 46.22 (9.29) Postgraduate Work 109.10 (13.73) 107.63 (11.62) 45.34 (8.40) Continued on the next page

PAGE 234

212 Table 36 (Continued) Maternal Report of Social Competence Teacher Report of Social Competence Maternal Report of Problem Behaviors Mean (SD) F Mean (SD) F Mean (SD) F Partner/Husband at Home No 101.18 (15.01) 6.86* 98.74 (11.02) 10.85* 48.44 (9.47) 1.80 Yes 106.74 (14.40) 105.09 (13.27) 46.62 (9.21) Receipt of Public Assistance No 107.10 (14.07) 15.46* 105.33 (13.01) 15.46* 46.39 (8.85) 8.32* Yes 100.27 (16.40) 99.10 (13.57) 49.60 (11.50) Breastfeeding at One Month No 103.09 (15.83) 16.40* 103.98 (13.21) 0.75 46.73 (9.44) 0.00 Yes 107.86 (13.59) 104.92 (13.22) 46.76 (9.15) *Indicates statistical significance at p<.05.

PAGE 235

213 Pearson correlation coefficients were computed to assess the relationships between maternal age and (1) maternal report of child social competence; (2) teacher report of child social competence; and (3) ma ternal report of child display of problem behaviors, at first grade. Ma ternal age was positively associ ated with both maternal and teacher reports of child social competence at first grade (r=0.11 and r=0.12, respectively). As maternal age increased, mother and teacher reported child social competence increased. Maternal age was negatively co rrelated with child display of problem behaviors at first grade (r=-0.13) As maternal age increase d, maternal report of child display of problem behaviors decreased. Although correlations were significant, associations were weak. Chi-square tests were used to identif y associations between sociodemographic control variables (all but matern al age) and (1) ever exposure to nonmaternal care at each time point; and (2) ever exposure to center-b ased nonmaternal care at each time point. Results for ever exposure to nonmaternal care and ever exposure to center-based nonmaternal care were similar, and are pres ented in Tables 37-46. Child birth order, maternal education, and family receipt of pub lic assistance were a ll three statistically associated with ever exposure to nonmatern al care and ever exposure to center-based nonmaternal care at six, 15, 24, and 36 months postpartum. Children higher in birth order and children of mothers with more educ ation were more likely to have ever been exposed to nonmaternal care as well as to have ever been exposed to center-based nonmaternal care at six, 15, 24, and 36 months postpartum. In addition, families on public assistance at one month postpartum were less likely to have children who were ever exposed to nonmaternal care, as well as to have children ever exposed to center-

PAGE 236

214 based nonmaternal care, at six, 15, 24, and 36 months postpartum. For both ever exposure to nonmaternal care and ever expos ure to center-based nonmaternal care, no associations were detected with child sex, mate rnal race, maternal ethnicity, presence of partner or husband living at home at one mont h postpartum, nor mate rnal breastfeeding at one month postpartum.

PAGE 237

215 Table 37 Ever Exposed to Nonmaternal Care at One Month by Sociodemographic Control Variables Ever Exposed to Nonmaternal Care at One Month % Yes ChiSquare(df) Pattern of Finding Child Sex Male 6.47 0.10(1) Female 7.08 Not Significant Child Birth Order 1 7.17 2 5.19 3 8.22 4 or More 13.79 Fisher’s Exact Test Not Significant Maternal Race White 6.87 Black/Afro-American 7.27 Other 3.70 Fisher’s Exact Test Not Significant Maternal Ethnicity Non-Hispanic Hispanic Unable to compute due to cell size of zero Maternal Education High School or Less 5.48 4.38(3) Some College 4.72 Bachelor’s Degree 9.68 Postgraduate Work 7.41 Not Significant Partner/Husband at Home No 6.00 0.05(1) Yes 6.84 Not Significant Receipt of Public Assistance No 6.81 0.01(1) Yes 6.49 Not Significant Breastfeeding at One Month No 5.96 0.33(1) Yes 7.16 Not Significant *Indicates statistical significance at p<.05.

PAGE 238

216 Table 38 Ever Exposed to Nonmaternal Care at Six Months by Sociodemographic Control Variables Ever Exposed to Nonmaternal Care at Six Months % Yes ChiSquare(df) Pattern of Finding Child Sex Male 66.47 1.92(1) Female 71.39 Not Significant Child Birth Order 1 78.18 2 66.30 3 47.95 4 or More 48.28 33.91(3) Children higher in birth order were more likely to have ever been exposed to nonmaternal care at six months. Maternal Race White 69.01 Black/Afro-American 61.82 Other 81.48 3.29(2) Not Significant Maternal Ethnicity Non-Hispanic 68.54 Hispanic 80.95 Fisher’s Exact Test Not Significant Maternal Education High School or Less 54.79 21.72(3) Some College 70.28 Bachelor’s Degree 70.43 Postgraduate Work 80.00 Mothers with more education were more likely to have children who were ever exposed to nonmaternal care at six months. Partner/Husband at Home No 74.00 0.65(1) Yes 68.52 Not Significant Receipt of Public Assistance No 71.93 22.34(1) Yes 45.45 Families on public assistance at one month were less likely to have children who were ever exposed to nonmaternal care at six months. Breastfeeding at One Month No 65.60 1.66(1) Yes 70.50 Not Significant *Indicates statistical significance at p<.05.

PAGE 239

217 Table 39 Ever Exposed to Nonmaternal Care at 15 Months by Sociodemographic Control Variables Ever Exposed to Nonmaternal Care at 15 Months % Yes ChiSquare(df) Pattern of Finding Child Sex Male 80.88 0.04(1) Female 80.24 Not Significant Child Birth Order 1 89.25 2 77.78 3 60.27 4 or More 65.52 39.51(3) Children higher in birth order were more likely to have ever been exposed to nonmaternal care at 15 months. Maternal Race White 80.57 Black/Afro-American 74.55 Other 92.59 Fisher’s Exact Test Not Significant Maternal Ethnicity Non-Hispanic 80.24 Hispanic 90.48 Fisher’s Exact Test Not Significant Maternal Education High School or Less 65.75 29.44(3) Some College 81.60 Bachelor’s Degree 84.41 Postgraduate Work 89.63 Mothers with more education were more likely to have children who were ever exposed to nonmaternal care at 15 months. Partner/Husband at Home No 82.00 0.07(1) Yes 80.45 Not Significant Receipt of Public Assistance No 83.22 24.04(1) Yes 59.74 Families on public assistance at one month were less likely to have children who were ever exposed to nonmaternal care at 15 months. Breastfeeding at One Month No 77.06 2.51(1) Yes 82.21 Not Significant *Indicates statistical significance at p<.05.

PAGE 240

218 Table 40 Ever Exposed to Nonmaternal Care at 24 Months by Sociodemographic Control Variables Ever Exposed to Nonmaternal Care at 24 Months % Yes ChiSquare(df) Pattern of Finding Child Sex Male 85.59 0.00(1) Female 85.55 Not Significant Child Birth Order 1 93.16 2 82.59 3 69.86 4 or More 72.41 34.90(3) Children higher in birth order were more likely to have ever been exposed to nonmaternal care at 24 months. Maternal Race White 85.26 Black/Afro-American 85.45 Other 92.59 Fisher’s Exact Test Not Significant Maternal Ethnicity Non-Hispanic 85.41 Hispanic 90.48 Fisher’s Exact Test Not Significant Maternal Education High School or Less 72.60 27.91(3) Some College 87.74 Bachelor’s Degree 87.63 Postgraduate Work 93.33 Mothers with more education were more likely to have children who were ever exposed to nonmaternal care at 24 months. Partner/Husband at Home No 88.00 0.26(1) Yes 85.37 Not Significant Receipt of Public Assistance No 87.04 9.37(1) Yes 74.03 Families on public assistance at one month were less likely to have children who were ever exposed to nonmaternal care at 24 months. Breastfeeding at One Month No 83.03 1.67(1) Yes 86.77 Not Significant *Indicates statistical significance at p<.05.

PAGE 241

219 Table 41 Ever Exposed to Nonmaternal Care at 36 Months by Sociodemographic Control Variables Ever Exposed to Nonmaternal Care at 36 Months % Yes ChiSquare(df) Pattern of Finding Child Sex Male 91.47 0.17(1) Female 92.33 Not Significant Child Birth Order 1 96.74 2 90.00 3 80.82 4 or More 86.21 Fisher’s Exact Test Children higher in birth order were more likely to have ever been exposed to nonmaternal care at 36 months. Maternal Race White 91.62 Black/Afro-American 92.73 Other 96.30 Fisher’s Exact Test Not Significant Maternal Ethnicity Non-Hispanic 91.79 Hispanic 95.24 Fisher’s Exact Test Not Significant Maternal Education High School or Less 80.82 Some College 93.87 Bachelor’s Degree 93.55 Postgraduate Work 98.52 Fisher’s Exact Test Mothers with more education were more likely to have children who were ever exposed to nonmaternal care at 36 months. Partner/Husband at Home No 92.00 0.00(1) Yes 91.89 Not Significant Receipt of Public Assistance No 93.19 11.86(1) Yes 81.82 Families on public assistance at one month were less likely to have children who were ever exposed to nonmaternal care at 36 months. Breastfeeding at One Month No 90.37 1.01(1) Yes 92.62 Not Significant *Indicates statistical significance at p<.05.

PAGE 242

220 Table 42 Ever Exposed to Center-Based Nonmaternal Care at One Month by Sociodemographic Control Variables Ever Exposed to Center-Based Nonmaternal Care at One Month % Yes ChiSquare(df) Pattern of Finding Child Sex Male 0.29 Female 0.29 Fisher’s Exact Test Not Significant Child Birth Order 1 2 3 4 or More Unable to compute due to cell size of zero Maternal Race White Black/Afro-American Other Unable to compute due to cell size of zero Maternal Ethnicity Non-Hispanic Hispanic Unable to compute due to cell sizes of zero Maternal Education High School or Less Some College Bachelor’s Degree Postgraduate Work Unable to compute due to cell sizes of zero Partner/Husband at Home No 2.00 Yes 0.16 Fisher’s Exact Test Not Significant Receipt of Public Assistance No 0.17 3.20(2) Yes 1.30 Not Significant Breastfeeding at One Month No Yes Unable to compute due to cell size of zero *Indicates statistical significance at p<.05.

PAGE 243

221 Table 43 Ever Exposed to Center-Based Nonmaternal Care at Six Months by Sociodemographic Control Variables Ever Exposed to Center-Bas ed Nonmaternal Care at Six Months % Yes ChiSquare(df) Pattern of Finding Child Sex Male 11.47 3.09(2) Female 9.73 Not Significant Child Birth Order 1 12.70 2 10.74 3 4.11 4 or More 3.45 Fisher’s Exact Test Children higher in birth order were more likely to have ever been exposed to center-based nonmaternal care at six months. Maternal Race White 11.39 Black/Afro-American 5.45 Other 3.70 Fisher’s Exact Test Not Significant Maternal Ethnicity Non-Hispanic 10.79 Hispanic 4.76 Fisher’s Exact Test Not Significant Maternal Education High School or Less 7.53 Some College 10.38 Bachelor’s Degree 11.29 Postgraduate Work 13.33 22.09(6) Mothers with more education were more likely to have children who were ever exposed to center-based nonmaternal care at six months. Partner/Husband at Home No 6.00 2.39(2) Yes 10.97 Not Significant Receipt of Public Assistance No 11.63 Yes 2.60 Fisher’s Exact Test Families on public assistance at one month were less likely to have children who were ever exposed to center-based nonmaternal care at six months. Breastfeeding at One Month No 10.55 1.74(2) Yes 10.63 Not Significant *Indicates statistical significance at p<.05.

PAGE 244

222 Table 44 Ever Exposed to Center-Based Nonmaternal Care at 15 Months by Sociodemographic Control Variables Ever Exposed to Center-Based Nonmaternal Care at 15 Months % Yes ChiSquare(df) Pattern of Finding Child Sex Male 15.59 0.19(2) Female 14.45 Not Significant Child Birth Order 1 17.26 2 15.93 3 5.48 4 or More 6.90 Fisher’s Exact Test Children higher in birth order were more likely to have ever been exposed to center-based nonmaternal care at 15 months. Maternal Race White 15.91 Black/Afro-American 10.91 Other 3.70 Fisher’s Exact Test Not Significant Maternal Ethnicity Non-Hispanic 15.35 Hispanic 4.76 Fisher’s Exact Test Not Significant Maternal Education High School or Less 11.64 Some College 15.57 Bachelor’s Degree 15.59 Postgraduate Work 17.04 29.52(6) Mothers with more education were more likely to have children who were ever exposed to center-based nonmaternal care at 15 months. Partner/Husband at Home No 16.00 0.09(2) Yes 14.94 Not Significant Receipt of Public Assistance No 15.95 Yes 7.79 24.83(2) Families on public assistance at one month were less likely to have children who were ever exposed to center-based nonmaternal care at 15 months. Breastfeeding at One Month No 14.22 2.51(2) Yes 15.40 Not Significant *Indicates statistical significance at p<.05.

PAGE 245

223 Table 45 Ever Exposed to Center-Based Nonmaternal Care at 24 Months by Sociodemographic Control Variables Ever Exposed to Center-Based Nonmaternal Care at 24 Months % Yes ChiSquare(df) Pattern of Finding Child Sex Male 23.24 0.01(2) Female 23.60 Not Significant Child Birth Order 1 26.71 2 22.59 3 12.33 4 or More 24.14 37.47(6) Children higher in birth order were more likely to have ever been exposed to center-based nonmaternal care at 24 months. Maternal Race White 24.29 Black/Afro-American 16.36 Other 18.52 Fisher’s Exact Test Not Significant Maternal Ethnicity Non-Hispanic 23.86 Hispanic 9.52 Fisher’s Exact Test Not Significant Maternal Education High School or Less 17.81 Some College 24.06 Bachelor’s Degree 23.66 Postgraduate Work 28.15 28.83(6) Mothers with more education were more likely to have children who were ever exposed to center-based nonmaternal care at 24 months. Partner/Husband at Home No 24.00 0.26(2) Yes 23.37 Not Significant Receipt of Public Assistance No 24.09 Yes 18.18 9.59(2) Families on public assistance at one month were less likely to have children who were ever exposed to center-based nonmaternal care at 24 months. Breastfeeding at One Month No 20.64 2.49(2) Yes 24.73 Not Significant *Indicates statistical significance at p<.05.

PAGE 246

224 Table 46 Ever Exposed to Center-Based Nonmaternal Care at 36 Months by Sociodemographic Control Variables Ever Exposed to Center-Based Nonmaternal Care at 36 Months % Yes ChiSquare(df) Pattern of Finding Child Sex Male 36.18 0.24(2) Female 37.46 Not Significant Child Birth Order 1 41.69 2 35.56 3 23.29 4 or More 31.03 28.40(6) Children higher in birth order were more likely to have ever been exposed to center-based nonmaternal care at 36 months. Maternal Race White 37.52 Black/Afro-American 27.27 Other 40.74 Fisher’s Exact Test Not Significant Maternal Ethnicity Non-Hispanic 36.93 Hispanic 33.33 Fisher’s Exact Test Not Significant Maternal Education High School or Less 26.03 Some College 35.85 Bachelor’s Degree 37.63 Postgraduate Work 48.89 42.19(6) Mothers with more education were more likely to have children who were ever exposed to center-based nonmaternal care at 36 months. Partner/Husband at Home No 28.00 1.92(2) Yes 37.52 Not Significant Receipt of Public Assistance No 37.87 Yes 28.57 12.56(2) Families on public assistance at one month were less likely to have children who were ever exposed to center-based nonmaternal care at 36 months. Breastfeeding at One Month No 33.49 2.49(2) Yes 38.39 Not Significant *Indicates statistical significance at p<.05.

PAGE 247

225 To test the associations between ma ternal age and (1) ever exposure to nonmaternal care at each time point; and (2) ever exposure to center-based nonmaternal care at each time point, one-way ANOVAs were computed. Results are summarized in Tables 47-48 below. Table 47 reveals that ma ternal age was not statistically associated with ever exposure to no nmaternal care at one, six, 15, 24, or 36 months. Table 47 ANOVA Results of Ever Exposure to Nonmat ernal Care Over Time by Maternal Age Ever Exposed to Nonmaternal Care Mean (SD) F One Month No 29.31 (5.19) 0.01 Yes 29.24 (5.15) Six Months No 29.02 (5.39) 0.96 Yes 29.44 (5.08) 15 Months No 28.92 (5.49) 0.94 Yes 29.40 (5.10) 24 Months No 29.33 (5.47) 0.00 Yes 29.31 (5.13) 36 Months No 28.89 (5.33) 0.39 Yes 29.35 (5.17) *Indicates statistical significance at p<.05. Table 48 reveals that maternal age was stat istically associated with ever exposure to center-based nonmaternal care at 36 months only. More specifically, children exposed to center-based nonmaternal care at 36 mont hs had older mothers, as compared to children exposed to non-cente r-based nonmaternal care.

PAGE 248

226 Table 48 ANOVA Results of Ever Exposed to Center -Based Nonmaternal Care Over Time by Maternal Age Ever Exposed to Center-Based Nonmaternal Care Mean (SD) F One Month Maternal Care Only 29.31 (5.19) 0.70 Non-Center-Based Care 29.43 (4.96) Center-Based Care 25.00 (9.90) Six Months Maternal Care Only 29.02 (5.39) 1.32 Non-Center-Based Care 29.31 (5.08) Center-Based Care 30.17 (5.04) 15 Months Maternal Care Only 28.92 (5.49) 1.22 Non-Center-Based Care 29.27 (5.12) Center-Based Care 29.97 (5.01) 24 Months Maternal Care Only 29.33 (5.47) 1.63 Non-Center-Based Care 29.07 (5.21) Center-Based Care 29.94 (4.87) 36 Months Maternal Care Only 28.89 (5.33) 3.63* Non-Center-Based Care 28.90 (5.23) Center-Based Care 30.01 (5.01) *Indicates statistical significance at p<.05. Pearson correlation matrices were comput ed to assess the relationships between (1) maternal depression over time; (2) maternal sensitivity over time; and (3) maternal depression and maternal sensit ivity over time. Results ar e summarized in Appendix A. Results indicate a moderate positive correlation between symptoms of maternal depression over time with coefficients ranging from 0.33 to 0.51, with the highest correlation between maternal depression at six and 15 months (r=0.51). Correlations between levels of maternal sensitivity over time were low to mode rate (ranging from 0.22

PAGE 249

227 to 0.38) with slightly higher correlations between later time poi nts in the postpartum period. For example, the correlation coefficien t for maternal sensitivity at six and 15 months, 15 and 24 months, and 24 and 36 m onths were 0.32, 0.35, and 0.38, respectively. Furthermore, correlations between maternal depression and maternal sensitivity were weak at best, ranging from –0.09 to –0.24. Th e strength of the association between maternal depression and maternal sensitivity s lightly increased over time. For example, the correlation coefficients for the associati ons between maternal depression and maternal sensitivity at six, 15, 24, and 36 months were –0.11, -0.15, -0.18, and –0.17, respectively, indicating that as levels of maternal depres sion increased, maternal sensitivity decreased. Next, associations between each of the continuous dependent variables and maternal depression over time were tested vi a Pearson correlation coefficients. Results are summarized in Table 49. Although maternal reports of social competence and display of problem behaviors were significant at p<.05, correlations were weak ranging from – 0.11 to –0.22 for social competence, and 0.09 to 0.23 for display of problem behaviors. As levels of maternal depression increase d, maternal reports of social competence decreased, and maternal reports of problem be haviors increased. Ma ternal depression at 24 months had the strongest correlation (r=-0.22 ) with maternal reports of total social competence at first grade, and maternal de pression at 36 months had the strongest correlation (r=0.23) with matern al reports of total problem behaviors at first grade. Significant correlations between maternal depr ession and teacher reported child social competence at first grade were negligible ranging from –0.08 to –0.11.

PAGE 250

228 Table 49 Pearson Correlation Matrices for Each Continuous De pendent Variable and Maternal Depression Over Time Maternal Depression Dependent Variable 1 Month 6 Months 15 Months 24 Months 36 Months Social Competence – Maternal Report Cooperation -0.11* -0.13* -0.11* -0.14* -0.10* Assertion -0.13* -0.14* -0.13* -0.15* -0.16* Responsibility -0.12* -0.12* -0.16* -0.14* -0.14* Self-Control -0.16* -0.17* -0.19* -0.21* -0.20* Peer Competence -0.16* -0.18* -0.15* -0.20* -0.18* Total Standardized Score -0.17* -0.20* -0.19* -0.22* -0.20 Social Competence – Teacher Report Cooperation -0.07 -0.08* -0.08* -0.03 0.02 Assertion -0.05 -0.04 -0.07 0.00 -0.04 Self Control -0.09* -0.10* -0.05 -0.05 0.00 Peer Competence -0.08* -0.09* -0.08* -0.04 -0.01 Total Standardized Score -0.08* -0.11* -0.10* -0.04 -0.03 Display of Problem Behaviors Internalizing Problem Behaviors 0.10* 0.09* 0.17* 0.13* 0.21* Externalizing Problem Behaviors 0.15* 0.16* 0.17* 0.19* 0.19* Total Problem Behaviors 0.16* 0.15* 0.20* 0.21* 0.23* *Indicates statistical significance at p<.05.

PAGE 251

229 In addition, one-way ANOVAs were com puted to identify the associations between maternal depression over time and (1 ) ever exposure to nonmaternal care over time; and (2) ever exposure to center-based nonmaternal care over time. Results are summarized in Tables 50-51. Table 50 shows th at ever exposure to nonmaternal care at one month was not statistically associated with maternal depression during the first 36 months postpartum. However, ever expos ure to nonmaternal car e at six months, 15 months, 24 months, and 36 months, was related to maternal depression scores at one, six, and 15 months postpartum. Specifically, child ren ever exposed to nonmaternal care at six months had mothers with lower levels of maternal depression at one, six, and 15 months postpartum. The same trend was detected for children ever exposed to nonmaternal care at 15, 24, and 36 months. Similarly, table 51 reveals that ever expos ure to center-based nonmaternal care at one month was not statistically associated with maternal depression during the first 36 months postpartum. At six months, ch ildren ever exposed to non-center-based nonmaternal care had mothers with statistically lower levels of ma ternal depression at one, six, and 15 months postpartum as compared to children exposed to maternal care only. No differences were detected betw een ever exposure to center-based and noncenter-based nonmaternal care at six months. At 15 months, (1 ) children ever exposed to center-based nonmaternal care had mothers with significantly lower levels of maternal depression at 15 months postpartum; and (2 ) children ever expose d to non-center-based nonmaternal care had mothers with significantly lower levels of ma ternal depression at one, six, and 15 months postpartum, as compared to children exposed to maternal care only. Furthermore, at 24 months, both ch ildren ever exposed to center-based

PAGE 252

230 nonmaternal care and non-center-based nonmater nal care had mothers with significantly lower levels of maternal de pression at one, six, and 15 months postpartum, as compared to children exposed to maternal care only. Last, children ever e xposed to center-based nonmaternal care and non-center-based nonmaterna l care at 36 months had mothers with statistically lower levels of maternal depr ession at one and 15 months postpartum.

PAGE 253

231 Table 50 ANOVA Results of Ever Exposed to N onmaternal Care Over Time by Mate rnal Depression Scores Over Time Maternal Depression One Month Six Months 15 Months 24 Months 36 Months Ever Exposed to Nonmaternal Care Mean (SD) F Mean (SD) F Mean (SD) F Mean (SD) F Mean (SD) F One Month No 9.57 (7.91) 1.36 7.65 (7.14) 0.02 7.47 (6.71) 0.72 7.61 (6.75) 0.52 7.43 (6.83) 0.69 Yes 8.18 (7.43 7.80 (7.35) 6.61 (6.51) 8.37 (7.67) 7.85 (6.75) Six Months No 10.82 (8.41) 9.16* 8.66 (7.98) 6.01* 8.73 (7.53) 11.94* 7.79 (6.96) 0.11 7.97 (7.07) 1.72 Yes 8.87 (7.43) 7.21 (6.71) 6.82 (6.21) 7.61 (6.75) 7.23 (6.70) 15 Months No 11.01 (8.73) 6.34* 9.14 (8.43) 7.10* 9.31 (7.39) 13.31* 7.70 (6.50) 0.00 7.75 (7.04) 0.29 Yes 9.11 (7.51) 7.30 (6.77) 6.96 (6.45) 7.66 (6.89) 7.39 (6.77) Continued on the next page

PAGE 254

232 Table 50 (Continued) Maternal Depression One Month Six Months 15 Months 24 Months 36 Months Ever Exposed to Nonmaternal Care Mean (SD) F Mean (SD) F Mean (SD) F Mean (SD) F Mean (SD) F 24 Months ----------No 11.74 (9.35) 9.78* 9.59 (8.49) 8.43* 9.79 (8.12) 14.71* 8.06 (6.60) 0.39 7.69 (7.08) 0.13 Yes 9.10 (7.44) 7.33 (6.86) 7.02 (6.35) 7.60 (6.85) 7.42 (6.78) 36 Months ----------No 12.58 (10.49) 9.58* 9.65 (8.31) 4.66* 10.92 (8.22) 16.77* 8.18 (7.21) 0.34 8.73 (8.15) 2.06 Yes 9.21 (7.46) 7.48 (7.02) 7.11 (6.46) 7.62 (6.78) 7.35 (6.69)

PAGE 255

233 Table 51 ANOVA Results of Ever Exposed to Center -Based Nonmaternal Care Over Time by Maternal Depression Scores Over Time Maternal Depression One Month Six Months 15 Months 24 Months 36 Months Ever Exposed to CenterBased Nonmaternal Care Mean (SD) F Mean (SD) F Mean (SD) F Mean (SD) F Mean (SD) F One Month Maternal Care Only 9.57 (7.91) 2.18 7.65 (7.14) 0.22 7.47 (6.71) 2.01 7.61 (6.75) 0.36 7.43 (6.83) 1.40 Non-Center-Based Care 7.76 (5.35) 7.66 (7.23) 6.22 (5.64) 8.27 (7.50) 7.50 (6.30) Center-Based Care 17.50 (13.44) 11.00 (12.73) 15.00 (19.80) 10.50 (14.85) 15.50 (14.85) Six Months Maternal Care Only 10.82 (8.41) 4.61* 8.66 (7.98) 3.00* 8.73 (7.53) 6.06* 7.79 (6.96) 0.07 7.97 (7.07) 1.18 Non-Center-Based Care 8.83 (7.27) 7.22 (6.74) 6.77 (6.06) 7.58 (6.62) 7.12 (6.56) Center-Based Care 9.10 (8.31) 7.15 (6.55) 7.14 (7.01) 7.75 (7.45) 7.82 (7.42) Continued on the next page

PAGE 256

234 Table 51 (Continued) Maternal Depression One Month Six Months 15 Months 24 Months 36 Months Ever Exposed to CenterBased Nonmaternal Care Mean (SD) F Mean (SD) F Mean (SD) F Mean (SD) F Mean (SD) F 15 Months Maternal Care Only 11.01 (8.73) 3.52* 9.14 (8.43) 3.57* 9.31 (7.39) 6.64* 7.70 (6.50) 0.02 7.75 (7.04) 0.57 Non-Center-Based Care 8.98 (7.40) 7.27 (6.82) 6.96 (6.46) 7.63 (6.77) 7.26 (6.62) Center-Based Care 9.69 (7.99) 7.43 (6.56) 6.95 (6.41) 7.78 (7.41) 7.95 (7.40) 24 Months Maternal Care Only 11.74 (9.35) 5.42* 9.59 (8.49) 4.21* 9.79 (8.12) 7.34* 8.06 (6.60) 0.34 7.69 (7.08) 0.07 Non-Center-Based Care 9.30 (7.51) 7.36 (6.80) 7.01 (6.37) 7.69 (6.82) 7.41 (6.58) Center-Based Care 8.56 (7.25) 7.28 (7.02) 7.02 (6.31) 7.35 (6.94) 7.47 (7.31) 36 Months Maternal Care Only 12.58 (10.49) 5.10* 9.65 (8.31) 2.33 10.92 (8.22) 8.52* 8.18 (7.20) 0.17 8.73 (8.15) 1.04 Non-Center-Based Care 9.40 (7.64) 7.51 (6.87) 7.22 (6.62) 7.64 (6.76) 7.37 (6.62) Center-Based Care 8.91 (7.19) 7.44 (7.25) 6.93 (6.23) 7.60 (6.82) 7.32 (6.81)

PAGE 257

235 Associations between each of the continuous dependent variables and maternal sensitivity over time were also explored via Pearson correlation coefficients. Results are summarized in Table 52. The correlations be tween maternal reports of total social competence and maternal sensitivity increase d over time ranging from 0.10 at one month postpartum to 0.21 at 36 months postpartum. Although significant, these correlations are weak. Teacher reports of total social competence were also significantly correlated with maternal sensitivity at six, 15, 24, and 36 months; nevertheless the associations were weak ranging from 0.09 at six months postp artum to 0.12 at 36 months postpartum. Child display of problem behaviors was negativ ely correlated with maternal sensitivity indicating that as levels of ma ternal sensitivity increased, di splay of problem behaviors at first grade decreased. Total problem behavi ors at first grade were only found to be significantly correlated with levels of mate rnal sensitivity at 24 and 36 months, with coefficients of –0.10 at 24 months and –0.14 at 36 months. Once again, despite significance correlations between total problem behaviors and maternal sensitivity were weak.

PAGE 258

236 Table 52 Pearson Correlation Matrices for Each Continuous De pendent Variable and Matern al Sensitivity Over Time Maternal Sensitivity Dependent Variable 1 Month 6 Months 15 Months 24 Months 36 Months Social Competence – Maternal Report Cooperation -0.06 0.04 0.09* 0.16* Assertion -0.06 0.08* 0.07 0.14* Responsibility -0.11* 0.10* 0.12* 0.14* Self-Control -0.06 0.13* 0.18* 0.20* Peer Competence -0.04 0.07 0.11* 0.14* Total Standardized Score -0.10* 0.11* 0.14* 0.21* Social Competence – Teacher Report Cooperation -0.11* 0.14* 0.13* 0.15* Assertion -0.05 0.04 0.03 0.07 Self Control -0.03 0.07 0.10* 0.12* Peer Competence -0.03 0.06 0.08* 0.11* Total Standardized Score -0.09* 0.10* 0.09* 0.12* Display of Problem Behaviors Internalizing Problem Behaviors --0.03 -0.05 -0.09* -0.19* Externalizing Problem Behaviors -0.00 -0.01 0.00 -0.03 Total Problem Behaviors --0.01 -0.06 -0.10* -0.14* *Indicates statistical significance at p<.05.

PAGE 259

237 In addition, one-way ANOVAs were com puted to identify the associations between maternal sensitivity over time and (1) ever exposure to nonmaternal care over time; and (2) ever exposure to center-based nonmaternal care over time. Results are summarized in Tables 53-54. Table 53 shows th at ever exposure to nonmaternal care at one month and 36 months were not statistica lly associated with maternal sensitivity during the first 36 months postpartum. Howe ver, ever exposure to nonmaternal care at six months was associated with maternal se nsitivity at 15 and 24 months, with children ever exposed at six months having mothers who rated statistically higher in maternal sensitivity. Similar trends were found for ch ildren ever exposed to nonmaternal care at 15 and 24 months. Those children had mothers who rated statistically higher in maternal sensitivity at 15, 24, and 36 months postpartum. Table 54 reveals that ever exposure to center-based nonmaternal care at one month postpartum was associated with maternal sensitivity at six months, with children exposed to non-center-based nonmaternal ca re having mothers who rated higher in maternal sensitivity at six months compared to children exposed to center-based nonmaternal care. At six months, (1) children ever expose d to center-based no nmaternal care had mothers who rated higher in maternal sensit ivity at 15 and 24 months as compared to children in nonmaternal care on ly; and (2) children ever e xposed to non-center-based care had mother who rated higher in maternal sensi tivity at 24 months as compared to children in nonmaternal care only. At 15 months, ch ildren ever exposed to both center-based nonmaternal care and non-center-based nonmater nal care had mothers who rated higher in maternal sensitivity at 15 and 24 months as compared to children in nonmaternal care

PAGE 260

238 only. At 24 months, children ever exposed to center-based nonmaternal care had mothers who rated higher in maternal sensitivity at 15 months as compared to both children in nonmaternal care only, and children in non-cen ter-based nonmaternal care. In addition, children ever exposed to non-center-based nonm aternal care at 24 months had mothers who rated higher in maternal sensitivity at 24 months than children exposed to nonmaternal care only. Last, children ever exposed to both center-based nonmaternal care and non-center-based nonmaternal care at 36 months had mothers who rated higher in maternal sensitivity at 15 as compar ed to children in nonmaternal care only.

PAGE 261

239 Table 53 ANOVA Results of Ever Exposed to Nonmaternal Care Over Time by Maternal Sensitivity Over Time Maternal Sensitivity Six Months 15 Months 24 Months 36 Months Ever Exposed to Nonmaternal Care Mean (SD)F Mean (SD)F Mean (SD)F Mean (SD)F One Month No 3.16 (0.56) 2.84 3.22 (0.49)0.88 3.23 (0.54)0.02 5.93 (0.77)0.63 Yes 3.30 (0.53) 3.29 (0.58) 3.24 (0.47) 5.83 (0.81) Six Months No 3.14 (0.54)0.86 3.15 (0.54)6.83* 3.11 (0.56)15.93* 5.86 (0.82)1.74 Yes 3.18 (0.57) 3.26 (0.48) 3.28 (0.52) 5.95 (0.74) 15 Months No 3.11 (0.54)1.84 3.08 (0.55)13.96* 3.08 (0.57)13.37* 5.80 (0.87)4.18* Yes 3.18 (0.56) 3.26 (0.48) 3.27 (0.52) 5.95 (0.74) 24 Months No 3.07 (0.53)3.35 3.11 (0.48)6.36* 3.11 (0.57)6.23* 5.77 (0.88)4.49* Yes 3.19 (0.56) 3.24 (0.50) 3.25 (0.53) 5.95 (0.75) 36 Months No 3.03 (0.59)3.76 3.08 (0.49)4.58 3.13 (0.59)1.95 5.77 (0.79)2.31 Yes 3.18 (0.56) 3.24 (0.50) 3.24 (0.53) 5.93 (0.77)

PAGE 262

240 Table 54 ANOVA Results of Ever Exposed to Cent er-Based Nonmaternal Care Over Time by Maternal Sensitivity Over Time Maternal Sensitivity Six Months 15 Months 24 Months 36 Months Ever Exposed to CenterBased Nonmaternal Care Mean (SD) F Mean (SD) F Mean (SD) F Mean (SD) F One Month Maternal Care Only 3.16 (0.56) 3.60* 3.22 (0.49) 0.74 3.23 (0.54) 0.03 5.93 (0.77) 0.51 Non-Center-Based Care 3.34 (0.50) 3.30 (0.57) 3.24 (0.48) 5.85 (0.81) Center-Based Care 2.50 (0.71) 3.00 (0.94) 3.17 (0.24) 5.50 (0.71) Six Months Maternal Care Only 3.14 (0.54) 0.46 3.15 (0.54) 5.15* 3.11 (0.56) 8.00* 5.86 (0.82) 0.91 Non-Center-Based Care 3.18 (0.57) 3.24 (0.49) 3.28 (0.51) 5.95 (0.75) Center-Based Care 3.20 (0.58) 3.36 (0.42) 3.30 (0.55) 5.92 (0.71) Continued on the next page

PAGE 263

241 Table 54 (Continued) Maternal Sensitivity Six Months 15 Months 24 Months 36 Months Ever Exposed to CenterBased Nonmaternal Care Mean (SD) F Mean (SD) F Mean (SD) F Mean (SD) F 15 Months Maternal Care Only 3.11 (0.54) 0.98 3.08 (0.55) 8.88* 3.08 (0.57) 6.69* 5.80 (0.87) 2.79 Non-Center-Based Care 3.18 (0.56) 3.24 (0.49) 3.27 (0.50) 5.97 (0.71) Center-Based Care 3.20 (0.58) 3.34 (0.46) 3.26 (0.58) 5.87 (0.85) 24 Months Maternal Care Only 3.07 (0.53) 2.50 3.11 (0.48) 6.87* 3.11 (0.57) 3.15* 5.77 (0.88) 2.27 Non-Center-Based Care 3.17 (0.56) 3.21 (0.51) 3.25 (0.51) 5.95 (0.73) Center-Based Care 3.23 (0.57) 3.33 (0.46) 3.24 (0.57) 5.93 (0.78) 36 Months Maternal Care Only 9.65 (8.31) 2.33 10.92 (8.22) 8.52* 8.18 (7.20) 0.17 8.73 (8.15) 1.04 Non-Center-Based Care 7.51 (6.87) 7.22 (6.62) 7.64 (6.76) 7.37 (6.62) Center-Based Care 7.44 (7.25) 6.93 (6.23) 7.60 (6.82) 7.32 (6.81)

PAGE 264

242 A Pearson correlation coefficient was computed to assess the relationship between maternal report of child social competence at first grade and teacher report of child social competence at first grade. Th e resulting coefficient was r=0.22 indicating that only 4.84% of the variance in teache r report of child social competence can be explained by maternal report of child soci al competence. Although the correlation was found to be significant at p<.05, the association was low. Multivariate Analyses Modifications to the Plan After reviewing variable frequency distributions for each categorical classification of child social competence (less than average, or average or greater than average social competence) and display of pr oblem behaviors (borde rline or clinically significant/deviant, or normal/non-deviant di splay of problem behaviors), the decision was made to omit further investigation of these dependent variab les due to their low variances. For example, only 47 (6.92%) childr en from the study sample were classified by mothers to have less than average social competence, and only 47 (6.92%) children from the study sample were classified by t eachers to have less than average social competence. Further, only 57 (8.39%) children were classified as displaying borderline or clinically significant deviant problem be haviors. Based on literature suggesting the number of “events” needed per independent va riable to maximize a logistic regression model’s stability of predic tion (Peduzzi, Concato, Kemper Holford, & Feinstein, 1996), there were insufficient numbers of outcome even ts in the study sample to properly fit the proposed models including all sociodemogra phic characteristics. Researchers have

PAGE 265

243 shown that testing models with too few even ts per variable, or EPV, results in high variability of regression coefficients (Pedu zzi et al., 1996). Thus, multivariate analyses will not be conducted for each of the three beha vioral categorical classifications of child social competence and display of problem behaviors including mother rated social competence, teacher rated social competen ce, and mother rated display of problem behaviors. Overview This study outlined four major research hypotheses including: (1) maternal depression is associated with child social competence and display of problem behaviors at first grade; (2) the relationship betw een maternal depression and child social competence and display of problem behaviors at first grade is mediated by maternal sensitivity; (3) the effect of maternal depr ession on child social competence and display of problem behaviors at firs t grade is stronger among mothers who rate low in maternal sensitivity; and (4) the eff ect of maternal depression on child social competence and display of problem behaviors at first gr ade is weaker among children exposed to nonmaternal care. For each hypothesis, ma ternal depression is operationalized in a variety of ways to understand the impact of (1) ever exposure to maternal depression during the first 36 months postpartum (hypot hesis 1.1); (2) timing of initial onset (hypothesis 1.2); (3) severity of symptoms along the trajectory of initial onset (hypothesis 1.3); and (4) chronicity of symptoms (hypot hesis 1.4). Furthermore, each of these considerations of maternal de pression were explored for a number of dependent variables including six maternal reports of child soci al competence at first grade (total social competence, cooperation, assertion, responsibilit y, self control, and peer competence),

PAGE 266

244 five teacher reports of child social competence at first grad e (total social competence, cooperation, assertion, self cont rol, and peer competence), and three maternal report of child display of problem behaviors at first gr ade (total problem behaviors, internalizing problems, and externalizing problems). For presentation of findings, multivariate results will be organized by research hypothesis w ithin each operationalization of maternal depression. This will be done for each subset of dependent variable s (maternal report of social competence, teacher report of social competence, and maternal report of problem behaviors). Ever Exposure to Maternal Depression Hypothesis 1.1. Hypothesis 1.1 stated that maternal depres sion, in general, during the first three years postpartum, is negatively associated w ith child social competence and display of problem behaviors at first grade. To te st hypothesis 1.1, each dependent variable was regressed against ever exposure to ma ternal depression controlling for the sociodemographic control variables. A summary of significant fi ndings from OLS regression models is presented in Table 55 for child social competence at firs t grade and Table 56 for child display of problem behaviors at first grade. An ‘X’ in a designated cell indicates that a significant association between the independent variab le and dependent vari able was detected, controlling for all other variab les included in the model. Ever exposure to maternal depression during the first 36 months postp artum was negatively associated with maternal reports of total social competence, assertion, peer comp etence, total problem behaviors, internalizing problem behaviors, and externalizing problem behaviors. More

PAGE 267

245 specifically, children ever e xposed to maternal depressi on during the first 36 months postpartum had decreased levels of total social competence, assertion, and peer competence, and increased levels of total problem behaviors, internalizing problem behaviors, and externalizing pr oblem behaviors. Ever expo sure to maternal depression during the first 36 months postpartum was not significantly associated with maternal reports of cooperation, responsibility, or self control; nor teacher reports of total social competence, cooperation, assertion, self contro l, or peer competence. See Appendix B, Tables B1 through B2, for partial model re sults highlighting parameter estimates (b coefficients), t values, and model fit statistics from regression models testing the impact of ever exposed to maternal depression on ch ild social competence or display of problem behaviors, where a significant main effect of depression was detected.

PAGE 268

246 Table 55 Summary of Significant Findi ngs from OLS Regression Testing the Impact of Ever Depressed on Child Social Competence Variables Total Social Competence Cooperation Assertion Responsibility Self Control Peer Competence Maternal Report Child Sex X ns ns ns ns ns Child Birth Order X ns X X X X Maternal Age ns ns ns ns ns ns Maternal Race X ns X ns ns X Maternal Ethnicity ns ns ns ns ns ns Maternal Education X X X X X X Partner/Husband at Home ns ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns ns Ever Depressed X ns X ns ns X Teacher Report Child Sex ns X X na X X Child Birth Order ns ns ns na ns ns Maternal Age ns ns ns na ns ns Maternal Race X X ns na X X Maternal Ethnicity ns ns ns na X ns Maternal Education X X X na X X Partner/Husband at Home ns ns ns na ns ns Receipt of Public Assistance ns ns ns na ns ns Breastfeeding at One Month ns ns X na ns X Ever Depressed ns ns ns na ns ns

PAGE 269

247 Table 56 Summary of Significant Findi ngs from OLS Regression Testing the Impact of Ever Depressed on Child Display of Problem Behaviors Variables Total Problem Behaviors Internalizing Problems Externalizing Problems Maternal Report Child Sex ns ns ns Child Birth Order ns X ns Maternal Age ns ns ns Maternal Race ns X ns Maternal Ethnicity ns ns ns Maternal Education ns ns X Partner/Husband at Home ns ns ns Receipt of Public Assistance ns ns ns Breastfeeding at One Month ns ns ns Ever Depressed X X X Hypothesis 2.1. Hypothesis 2, that the relationship between maternal depression and child social competence and display of problem behaviors at first grade is mediated by maternal sensitivity, was tested by adding maternal sensitivity to the model containing the sociodemographic control variables and ever ex posure to maternal depression. Mediating effects of maternal sensitivity were only te sted in models for which ever exposure to maternal depression had a significant main effect (maternal reports of total social competence, assertion, peer competence, total problem behaviors, in ternalizing problems, and externalizing problems). Maternal sens itivity was operationalized as the average level of maternal sensitivity during the first 36 months postpartum. Given that maternal sensitivity at 36 months was rated via a 7point scale as opposed to the 4-point scale

PAGE 270

248 employed at six, 15, and 24 months, levels of ma ternal sensitivity at each time point were first standardized to have a mean of zero and standard deviation of one, prior to averaging maternal sensitivity over time. Prior to entering maternal sensitivity to each model in which a main effect of ever exposure to maternal depression was de tected, the associations between (1) ever exposure to maternal depression and maternal sensitivity; and (2) maternal sensitivity and each dependent variable under investigation while controlling for ever exposure to maternal depression, were first evaluated as a preliminary step in establishing mediation. Results are summarized in Ta ble 57. Ever exposure to mate rnal depression during the first 36 months postpartum was significantly associated with maternal sensitivity. Furthermore, maternal sensitivity was significan tly associated with maternal reports of total social competence, a ssertion, peer competence, total problem behaviors, and externalizing problem behaviors, while simu ltaneously controlling for ever exposure to maternal depression. No significant asso ciation was detected between maternal sensitivity and internalizing problem behaviors, thus no further investigation of a mediating effect of maternal sensitivity on the relationshi p between ever exposure to maternal depression and internaliz ing behaviors was explored.

PAGE 271

249 Table 57 Tested Associations for Establishing a Medi ating Effect of Maternal Sensitivity (MS) Tested Associations b coefficient t value (b=0) Ever Exposure X Maternal Sensitivity (MS) -0.14 -5.22* MS X Maternal Report of Total Social Competence 6.98 4.49* MS X Maternal Report of Assertion 0.64 2.61* MS X Maternal Report of Peer Competence 0.76 2.73* MS X Maternal Report of Tota l Problem Behaviors -2.02 -2.01* MS X Maternal Report of Internalizing Problems 0.19 0.21 MS X Maternal Report of Externalizing Problems -2.22 -2.16* *Indicates statistical significance at p<.05. Parameter estimates (b coefficients) and associated p values for ever exposure to maternal depression in both the reduced model (sociodemographi c control variables + ever exposure to maternal depression) and full model (sociodemographic control variables + ever exposure to maternal depres sion + maternal sensitivity) were compared to assess the mediating effect of maternal sensitivity. Partial model results (excluding sociodemographic control variables) are provided in Tables 58-62.

PAGE 272

250 Table 58 Mediating Effect of Maternal Sensitivity in the Associat ion Between Ever Exposure to Maternal Depression and Maternal Report of Total Social Competence1 Maternal Report of Total Social Competence Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Ever Depressed No 0.00 -0.00 -Yes -2.94 -2.57* -2.75 -2.40* Maternal Sensitivity --1.68 1.79 Model Fit Statistics (N=697) R2 0.11 0.11 Overall Model F 5.27* 5.16* *Indicates statistical significance at p<.05. 1Controlling for all sociodemographic control variables including child sex, child birth order, maternal age, maternal race, maternal ethnicity, maternal education, presence of partner/husband at home at one month postpartum, family receipt of public assistance at one month postpartum, and maternal br eastfeeding at one month postpartum. Table 58 above summarizes the evidence for maternal sensitivity as a mediating effect in the asso ciation between ever exposure to maternal depression during the first 36 months postpartum and maternal re port of total social competence at first grade. The hypothesis was only partially s upported. With the addition of maternal sensitivity to the reduced m odel (see full model results), ever depressed remained a significant predictor of maternal report of total social competence, thus complete mediation was not evident. However, the magnitude of effect of ever depressed decreased after adding maternal sensitivity to the reduced model. The unstandardized regression coefficient or b coefficient was -2.94 in the reduced model decreasing to -2.75 in the full model after the addition of maternal sensitivity. Thus, maternal sensitivity

PAGE 273

251 partially mediated the relationship between ever exposure to mate rnal depression during the first 36 months postpartum and maternal re port of total social competence at first grade. Table 59 Mediating Effect of Maternal Sensitivity in the Associat ion Between Ever Exposure to Maternal Depression and Maternal Report of Assertion1 Maternal Report of Assertion Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Ever Depressed No 0.00 -0.00 -Yes -0.46 -2.55* -0.45 -2.51* Maternal Sensitivity --0.05 0.31 Model Fit Statistics (N=697) R2 0.06 0.06 Overall Model F 2.85* 2.67* *Indicates statistical significance at p<.05. 1Controlling for all sociodemographic control variables including child sex, child birth order, maternal age, maternal race, maternal ethnicity, maternal education, presence of partner/husband at home at one month postpartum, family receipt of public assistance at one month postpartum, and maternal br eastfeeding at one month postpartum. Table 59 above summarizes the evidence for maternal sensitivity as a mediating effect in the asso ciation between ever exposure to maternal depression during the first 36 months postpartum and maternal re port of assertion at first grade. The hypothesis was partially supported. With the addition of maternal sensitivity to the reduced model (see full model results), ever de pressed remained a sign ificant predictor of maternal report of child assertion at first grad e, thus complete mediation was not evident.

PAGE 274

252 However, the magnitude of effect of ever depressed decreased after adding maternal sensitivity to the reduced model. The unstandardized regression coefficient or b coefficient was –0.46 in the reduced model decr easing to –0.45 in the full model after the addition of maternal sensitivity. Thus, maternal sensitivity partially mediated the relationship between ever exposure to mate rnal depression during the first 36 months postpartum and maternal report of assertion at first grade. Table 60 Mediating Effect of Maternal Sensitivity in the Associat ion Between Ever Exposure to Maternal Depression and Matern al Report of Peer Competence1 Maternal Report of Peer Competence Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Ever Depressed No 0.00 -0.00 -Yes -0.52 -2.51* -0.51 -2.46 Maternal Sensitivity --0.08 0.45 Model Fit Statistics (N=697) R2 0.07 0.07 Overall Model F 3.09* 2.90* *Indicates statistical significance at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal race, maternal ethnicity, maternal education, presence of partner/husband at home at one mont h postpartum, family receipt of public assistance at one month postpartum, and maternal breastfeeding at one month postpartum.

PAGE 275

253 Table 60 above summarizes the evidence for maternal sensitivity as a mediating effect in the asso ciation between ever exposure to maternal depression during the first 36 months postpartum and maternal re port of peer competence at first grade. The hypothesis was partially supported. With the addition of maternal sensitivity to the reduced model (see full model results), ever de pressed remained a sign ificant predictor of maternal report of child assertion at first grad e, thus complete mediation was not evident. However, the magnitude of effect of ever depressed decreased after adding maternal sensitivity to the reduced model. The unstandardized regression coefficient or b coefficient was –0.52 in the reduced model decr easing to –0.51 in the full model after the addition of maternal sensitivity. Thus, maternal sensitivity partially mediated the relationship between ever exposure to mate rnal depression during the first 36 months postpartum and maternal report of peer competence at first grade.

PAGE 276

254 Table 61 Mediating Effect of Maternal Sensitivity in the Associat ion Between Ever Exposure to Maternal Depression and Maternal Report of Total Problem Behaviors1 Total Problem Behaviors Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Ever Depressed No 0.00 -0.00 -Yes 2.19 2.92* 2.14 2.84* Maternal Sensitivity ---0.51 -0.83 Model Fit Statistics (N=697) R2 0.05 0.05 Overall Model F 2.32* 2.22* *Indicates statistical significance at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal race, maternal ethnicity, maternal education, presence of partner/husband at home at one mont h postpartum, family receipt of public assistance at one month postpartum, and maternal breastfeeding at one month postpartum. Table 61 above summarizes the evidence for maternal sensitivity as a mediating effect in the asso ciation between ever exposure to maternal depression during the first 36 months postpartum and maternal re port of total problem behaviors at first grade. Overall, the hypothesis is only part ially supported. With the addition of maternal sensitivity to the reduced m odel (see full model results), ever depressed remained a significant predictor of maternal report of total problem behaviors, thus complete mediation was not evident. However, the magnitude of effect of ever depressed decreased slightly after a dding maternal sensitivity to the reduced model. The unstandardized regression coefficient or b coefficient was 2.19 in the reduced model

PAGE 277

255 decreasing to 2.14 in the full model after the addition of maternal sensitivity. Thus, maternal sensitivity partially mediated the relationship between ever exposure to maternal depression during the first 36 months postpar tum and maternal report of total problem behaviors at first grade. Table 62 Mediating Effect of Maternal Sensitivity in the Associat ion Between Ever Exposure to Maternal Depression and Maternal Report of Externalizing Problems1 Externalizing Problems Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Ever Depressed No 0.00 -0.00 -Yes 1.96 2.58* 1.92 2.50* Maternal Sensitivity ---0.44 -0.71 Model Fit Statistics (N=697) R2 0.05 0.06 Overall Model F 2.55* 2.42* *Indicates statistical significance at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal race, maternal ethnicity, maternal education, presence of partner/husband at home at one mont h postpartum, family receipt of public assistance at one month postpartum, and maternal breastfeeding at one month postpartum. Table 62 above summarizes the evidence for maternal sensitivity as a mediating effect in the asso ciation between ever exposure to maternal depression during the first 36 months postpartum and maternal re port of externalizing problem behaviors at first grade. Overall, the hypothesis is only partially suppor ted. With the addition of maternal sensitivity to the reduced model (see full model results), ever depressed

PAGE 278

256 remained a significant predictor of maternal report of externalizing problem behaviors, thus complete mediation was not evident. However, the magnitude of effect of ever depressed decreased slightly af ter adding maternal sensitivity to the reduced model. The unstandardized regression coefficient or b coefficient was 1.96 in the reduced model decreasing to 1.92 in the full model after the addition of maternal sensitivity. Thus, maternal sensitivity partially mediated the relationship between ever exposure to maternal depression during the first 36 months postpar tum and maternal repor t of externalizing problem behaviors. Hypothesis 3.1. Hypothesis 3 stated that the effect of maternal depression on child social competence and display of problem behavior s at first grade is stronger among mothers who rate low in maternal sensitivity. To test the moderating effect of maternal sensitivity on the associations between ever exposure to maternal depression during the first 36 months postpartum and child social competence and display of problem behaviors at first grade, the interacti on term of ever exposure to mate rnal depression times maternal sensitivity (ever depressed X maternal se nsitivity) was added to each final model (sociodemographic control variables + ever e xposure to maternal depression + maternal sensitivity). Maternal sensitivity was operati onalized as the average level of maternal sensitivity during the first 36 months postpartum. A significant interaction term indicates the presence of a moderating effect. OLS regression results are summarized in Table 63. Maternal sensitivity was a significant moderator on the a ssociation between ever expos ure to maternal depression and teacher reported total social competence at first grade, with maternal depression

PAGE 279

257 having the greatest impact on total social competence among mothers rated low in maternal sensitivity. No other moderating in fluences of maternal sensitivity were detected. Table 63 OLS Regression Results Testing the Moderati ng Effect of Maternal Sensitivity on the Association Between Ever Exposure to Maternal Depression and Child Social Competence and Display of Problem Behaviors Ever Depressed X Maternal Sensitivity Dependent Variables b coefficient t value (b=0) Maternal Report of Social Competence Total 0.36 0.23 Cooperation 0.20 0.58 Assertion 0.48 1.88 Responsibility -0.07 -0.25 Self Control -0.22 -0.60 Peer Competence 0.43 1.47 Teacher Report of Social Competence Total 3.38 2.28* Cooperation 0.66 1.55 Assertion 0.76 1.81 Self Control 0.75 1.88 Peer Competence 0.70 1.80 Maternal Report of Problem Behaviors Total Problem Behaviors -0.37 -0.36 Internalizing Problem Behaviors 0.13 0.14 Externalizing Problem Behaviors 0.48 0.45 *Indicates statistical significance at p<.05.

PAGE 280

258 Hypothesis 4.1. Hypothesis 4 stated that the effect of maternal depression on child social competence and display of problem behaviors at first grade is weaker among children exposed to nonmaternal care. To test the moderating effect of exposure to nonmaternal care on the associations between ever exposur e to maternal depression during the first 36 months postpartum and child social competence and display of problem behaviors at first grade, the interaction term of ever exposure to matern al depression times exposure to nonmaternal care was added to each final m odel (sociodemographic control variables + ever exposure to maternal depression + mate rnal sensitivity + exposure to nonmaternal care). Exposure to nonmaternal care was opera tionalized in two ways: (1) ever exposure to nonmaternal (despite type of arrangemen t); and (2) ever exposure to center-based nonmaternal care. Ever expos ure to nonmaternal care consis ted of two response options (no and yes), and ever exposure to center-b ased nonmaternal care consisted of three response options (maternal care only, noncenter-based nonmaternal care, and centerbased nonmaternal care). A significant inte raction term indicates the presence of a moderating effect. OLS regression results are summarized in Table 64. Exposure to nonmaternal care, operationalized as both ev er exposure to nonmaternal care and ever exposure to center-based nonmaternal care, was not a significant moderator on the association between ever exposure to maternal depression and child social compet ence and display of problem behaviors at first grade, using either criteria to determine statistical significance.

PAGE 281

259 Table 64 OLS Regression Results Testing the Moderating Effect of Ex posure to Nonmaternal Care on the Association Between Ever Exposure to Maternal Depression and Child Social Competence and Display of Problem Behaviors Ever Depressed X Ever Exposure to Nonmaternal Care Ever Depressed X Ever Exposure to Center-Based Nonmaternal Care Dependent Variables b coefficient t value (b=0) b coefficient t value (b=0) Maternal Report of Social Competence Total -1.39 -0.35 -0.98 -0.54 Cooperation -0.37 -0.44 0.10 0.26 Assertion -0.15 -0.24 -0.24 -0.82 Responsibility 0.29 0.39 0.10 0.29 Self Control -0.50 -0.55 -0.44 -1.07 Peer Competence -0.60 -0.83 -0.36 -1.10 Teacher Report of Social Competence Total 0.75 0.20 1.22 0.73 Cooperation -0.14 -0.13 0.42 0.87 Assertion 1.29 1.23 0.47 0.97 Self Control 0.29 0.29 0.14 0.32 Peer Competence 0.85 0.88 0.37 0.85 Maternal Report of Problem Behaviors Total Problem Behaviors 1.94 0.74 1.37 1.15 Internalizing Problem Behaviors 2.18 0.90 0.94 0.85 Externalizing Problem Behaviors1.51 0.57 1.86 1.54 *Indicates statistical significance at p<.05.

PAGE 282

260 Timing of Initial Onset Hypothesis 1.2. Hypothesis 1.2 stated that ti ming of initial on set of maternal depression, during the first three years postpartum, is associated with child social competence and display of problem behaviors at first grade, with earlier onset episodes having the greatest negative impact. To test hypothesis 1.2, a series of models were tested. For each dependent variable, the dependent variab le was first regressed agains t maternal depression at one month, dichotomously scored, controlling fo r the sociodemographic control variables (n=679) (Model 1). Next, among those not prev iously depressed (n=558), the dependent variables were regressed against maternal depression at six months, dichotomously scored, controlling for the sociodemographic control variables (Model 2). Then, among those not previously depressed (n=510), the dependent variables were regressed against maternal depression at 15 months, dic hotomously scored, controlling for the sociodemographic control variables (Model 3) The same process was conducted to test the impact of initial onset of maternal depression at 24 and 36 months, among samples of previously nondepressed women of 486 a nd 447 respectively (Models 4 and 5). The significance of each maternal depression variab le across models was determined by the pvalue associated with its unstandardized regression coefficient. A summary of significant fi ndings from OLS regression models is presented in Table 65. An ‘X’ in a designated cell indicate s that a significant a ssociation between the independent variable and dependent variable was detected at p<.05, controlling for all other variables included in the model. An ‘na’ in a designated cell indicates that a particular independent variable was not tested in a given mo del (for example, depression

PAGE 283

261 at one month is not included in model 2 gi ven that the population for model 2 is among those not previously depressed). Only depende nt variables with at least one significant main effect of maternal depression across m odels tested (models 1-5) are presented in Table 65 including: maternal report of tota l social competence, maternal report of assertion, maternal report of p eer competence, teacher report of total social competence, teacher report of self control, teacher report of peer competence, and maternal reports of both total problem behaviors and externalizing problem behaviors. No main effect of maternal depression was detected across mode ls 1-5 for maternal report of cooperation, maternal report of responsibility, maternal report of self contro l, teacher report of cooperation, teacher report of assertion, a nd maternal report of internalizing problem behaviors.

PAGE 284

262 Table 65 Summary of Significant Findings from OLS Regression Testi ng the Impact of Timing of Initial Onset of Maternal Depression on Child Social Competence and Display of Problem Behaviors Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All ND1 ND ND ND Maternal Report of Total Social Competence Child Sex X ns ns X X Child Birth Order X X X X X Maternal Age ns ns ns ns ns Maternal Race X X X X X Maternal Ethnicity ns ns ns ns ns Maternal Education X X X X X Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depressed at One Month ns na na na na Depressed at Six Months na X na na na Depressed at 15 Months na na ns na na Depressed at 24 Months na na na ns na Depressed at 36 Months na na na na ns Maternal Report of Assertion Child Sex ns ns ns ns ns Child Birth Order X X X ns ns Maternal Age ns ns ns ns ns Maternal Race X X X X X Maternal Ethnicity ns ns ns ns ns Maternal Education X X X X ns Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depressed at One Month ns na na na na Depressed at Six Months na X na na na Depressed at 15 Months na na ns na na Depressed at 24 Months na na na ns na Depressed at 36 Months na na na na ns Continued on the next page

PAGE 285

263 Table 65 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All ND ND ND ND Maternal Report of Peer Competence Child Sex ns ns X ns ns Child Birth Order X X X X X Maternal Age ns ns ns ns ns Maternal Race X X X X X Maternal Ethnicity ns ns ns ns ns Maternal Education X ns ns ns ns Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depressed at One Month ns na na na na Depressed at Six Months na X na na na Depressed at 15 Months na na ns na na Depressed at 24 Months na na na ns na Depressed at 36 Months na na na na ns Teacher Report of Total Social Competence Child Sex ns ns ns ns ns Child Birth Order ns ns ns ns ns Maternal Age ns ns ns ns ns Maternal Race X ns ns ns ns Maternal Ethnicity ns ns ns ns ns Maternal Education X X X X X Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns X ns ns ns Depressed at One Month ns na na na na Depressed at Six Months na X na na na Depressed at 15 Months na na ns na na Depressed at 24 Months na na na ns na Depressed at 36 Months na na na na ns Continued on the next page

PAGE 286

264 Table 65 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All ND ND ND ND Teacher Report of Self Control Child Sex X X X X X Child Birth Order ns X X X X Maternal Age ns ns ns ns ns Maternal Race X ns ns ns ns Maternal Ethnicity X ns ns ns ns Maternal Education X X X X X Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depressed at One Month ns na na na na Depressed at Six Months na X na na na Depressed at 15 Months na na ns na na Depressed at 24 Months na na na ns na Depressed at 36 Months na na na na ns Teacher Report of Peer Competence Child Sex X X X X X Child Birth Order ns X X X X Maternal Age ns ns ns ns ns Maternal Race X ns ns ns ns Maternal Ethnicity ns ns ns ns X Maternal Education X X X X X Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month X X ns ns ns Depressed at One Month ns na na na na Depressed at Six Months na X na na na Depressed at 15 Months na na ns na na Depressed at 24 Months na na na ns na Depressed at 36 Months na na na na ns Continued on the next page

PAGE 287

265 Table 65 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All ND ND ND ND Maternal Report of Total Problems Child Sex ns ns ns ns ns Child Birth Order ns X X X X Maternal Age ns ns ns ns ns Maternal Race ns ns ns ns ns Maternal Ethnicity ns ns ns ns ns Maternal Education ns X X X X Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depressed at One Month X na na na na Depressed at Six Months na ns na na na Depressed at 15 Months na na ns na na Depressed at 24 Months na na na ns na Depressed at 36 Months na na na na ns Maternal Report of Externalizing Problems Child Sex ns ns ns ns ns Child Birth Order ns ns X X ns Maternal Age ns ns ns ns ns Maternal Race ns ns ns ns ns Maternal Ethnicity ns ns ns ns ns Maternal Education X X X X X Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depressed at One Month X na na na na Depressed at Six Months na ns na na na Depressed at 15 Months na na ns na na Depressed at 24 Months na na na ns na Depressed at 36 Months na na na na ns 1ND denotes ‘Never Depressed’

PAGE 288

266 Results from Table 65 above support hypothe sis 1.2, that timing of initial onset of maternal depression, during the first three y ears postpartum, is associated with child social competence and display of problem beha viors, with earlier onset episodes having the greatest negative impact. For both ma ternal and teacher reported total social competence and peer competence, maternal repor t of assertion, and teacher report of self control, maternal depression that initiated between the first and sixth postpartum month (and was present at six months postpartum) wa s significantly associated with child social competence at first grade. More specifical ly, children who were exposed to maternal depression at six months postpartum (among those who had not been exposed at one month postpartum) were rated by mothers as displaying significantly less total social competence, assertion, and peer competen ce; and rated by teacher as displaying significantly less total social competence, self control, and peer competence. Maternal depression that initiated after the sixt h postpartum month (by 15, 24, or 36 months postpartum) did not have statistically signi ficant impacts on child so cial competence at first grade. Thus, results indi cate a critical period of de pression developing between the first and sixth postpartum months related to child development of social competence. Furthermore, results from Table 65 hi ghlight the importance of maternal depression that initiated during and was still present by the fi rst postpartum month related to child display of total problem behavior s and externalizing problem behaviors, as reported by the mother. Only children exposed to depression that initi ated during the first month postpartum scored significantly hi gher in total problem behaviors and externalizing problem behaviors. Depression that initiated at later time points (six, 15, 24, or 36 months) did not significantly imp act maternal report s of child problem

PAGE 289

267 behaviors. See Appendix C, Tables C1 through C8, for partial model results highlighting parameter estimates (b coefficients), t values, and model fit statistics from regression models testing the impact of timing of initial onset of maternal depression on child social competence or display of problem behaviors, wh ere at least one significant main effect of maternal depression was detected acro ss models tested (models 1-5).

PAGE 290

268 Hypothesis 2.2. Hypothesis 2, that the relationship between maternal depression (timing of initial onset) and child social competence and displa y of problem behaviors at first grade is mediated by maternal sensitivity, was tested by adding maternal sensitivity to the models containing the sociodemographic control variables and presence of maternal depression at a given time point. Given that maternal sensitivity was only a ssessed in the NICHD SECC at six, 15, 24, and 36 months (no one-m onth assessment), mediating effects of maternal sensitivity were only tested in models for which maternal depression had a significant main effect at six, 15, 24 or 36 months (maternal reports of total social competence, assertion, and peer competence ; and teacher reports of total social competence, self control, and peer competence) Note that each of the significant main effects of maternal depressi on on child social competence wa s detected at six months postpartum only, among children not previous ly exposed to maternal depression. Although significant main effects of maternal depression were detected for maternal reports of total problem behaviors and exte rnalizing problems, both main effects were detected at one month postpartum only. No fu rther exploration of a mediating effect of maternal sensitivity was tested in these mode ls given that maternal sensitivity at one month postpartum was not assessed. Thus, mediating influences of maternal sensitivity were explored for the following dependent variables: (1) impact of maternal depression at six months postpartum on maternal report of total social competence, among children not previously exposed to maternal depression; (2) impact of maternal depre ssion at six months postpartum on maternal report of assertion, among child not previously exposed to

PAGE 291

269 maternal depression; (3) imp act of maternal depression at six months postpartum on maternal report of peer competence, among those not previously exposed to maternal depression; (4) impact of maternal depressi on at six months postpartum on teacher report of total social competence, among those not pr eviously exposed to maternal depression; (5) impact of maternal depression at six m onths postpartum on teacher report of self control, among those not previously exposed to maternal depression; and (6) impact of maternal depression at six months postpartu m on teacher report of peer competence, among those not previously exposed to matern al depression. For each of these dependent variables, maternal sensitivity was operati onalized as the mother’s observed rating of maternal sensitivity at six months postpartum (response options ranged from one to four). Prior to entering maternal sensitivity to each model in which a main effect of maternal depression, dichotomously scored, was detected, the following associations were explored as a preliminary step in esta blishing mediation: (1) maternal depression at six months postpartum, among those children not previously exposed to depression, and maternal sensitivity at six months; and (2) maternal sensitivity at six months and each dependent variable under investigation, while controlling for maternal depression at six months postpartum. Results are summarized in Table 66.

PAGE 292

270 Maternal depression at six months, am ong children not previously exposed to depression, was significantly associated with maternal sensitivity at six months. In addition, maternal sensitivity at six months was associated with both maternal and teacher reports of total social competence. No other significant relationships between maternal sensitivity at six months and maternal report of assertion or peer competence, or teacher report of self control or peer compet ence were detected. Fu rther investigation of a mediating effect of maternal sensitivity will only be conducted for maternal depression at six months, among children not previously exposed to maternal depression, on both maternal and teacher reports of total social competence. Table 66 Tested Associations for Establishing a Media ting Effect of Maternal Sensitivity (MS) in the Association Between Timing of Initial Onse t of Maternal Depression and Child Social Competence and Display of Problem Behaviors Tested Associations b coefficient t value (b=0) Maternal Depression at Six Months X Maternal Sensitivity at Six Months -0.19 -2.31* Maternal Sensitivity at Six Months X Maternal Report of Total Social Competence 2.39 2.43* Maternal Sensitivity at Six Months X Maternal Report of Assertion 0.22 1.41 Maternal Sensitivity at Six Months X Maternal Report of Peer Competence 0.13 0.76 Maternal Sensitivity at Six Months X Teacher Report of Total Social Competence 1.98 2.20* Maternal Sensitivity at Six Months X Teacher Report of Self Control 0.16 0.68 Maternal Sensitivity at Six Months X Teacher Report of Peer Competence 0.13 0.55 *Indicates statistical significance at p<.05.

PAGE 293

271 Parameter estimates (b coefficients) and associated p values for maternal depression at six months, among children not pr eviously exposed to depression, in both the reduced model (sociodemographic cont rol variables + expos ure to maternal depression at six months) and the full m odel (sociodemographic control variables + exposure to maternal depression at six months + maternal sensitivity) were compared to assess the mediating effect of maternal se nsitivity. Partial mo del results (excluding sociodemographic control variables) are provided in Tables 67-68. Table 67 Mediating Effect of Maternal Sensitivity in the Associati on Between Timing of Initial Onset of Maternal Depression and Matern al Report of Total Social Competence Maternal Report of Total Social Competence Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Depressed at Six Months No 0.00 -0.00 -Yes -4.26 -2.01* -4.41 -2.08* Maternal Sensitivity ---1.55 -1.34 Model Fit Statistics (N=558) R2 0.09 0.12 Overall Model F 4.81* 4.63* *Indicates statistical significance at p<.05.

PAGE 294

272 Table 67 above summarizes the evidence for maternal sensitivity as a mediating effect in the asso ciation between maternal de pression at six months among children not previously expos ed to depression and matern al report of total social competence at first grade. The hypothesis is not supported; there is no evidence of a mediating effect of maternal sensitivity. Af ter the addition of maternal sensitivity to the reduced model (see full model results) not onl y did maternal depres sion at six months remain a significant predictor of maternal report of total social competence, but also the magnitude of effect of ma ternal depression increased rather than decreased. Table 68 Mediating Effect of Maternal Sensitivity in the Associati on Between Timing of Initial Onset of Maternal Depression and Teache r Report of Total Social Competence Teacher Report of Total Social Competence Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Depressed at Six Months No 0.00 -0.00 -Yes -4.07 -2.03* -4.10 -2.04* Maternal Sensitivity ---0.32 -0.29 Model Fit Statistics (N=558) R2 0.07 0.07 Overall Model F 2.53* 2.38* *Indicates statistical significance at p<.05.

PAGE 295

273 Table 68 above summarizes the evidence fo r maternal sensitivity as a mediating effect in the association between maternal depression at six months among children not previously exposed to depression and teacher report of total social competence at first grade. The hypothesis is not supported; ther e is no evidence of a mediating effect of maternal sensitivity. After the addition of mate rnal sensitivity to the reduced model (see full model results) not only did maternal depr ession at six months remain a significant predictor of teacher report of total social competence, but also the magnitude of effect of maternal depression increased rather than decreased.

PAGE 296

274 Hypothesis 3.2. Hypothesis 3 stated that the effect of maternal depression (timing of initial onset) on child social competence and display of problem behaviors at first grade is stronger among mothers who rate low in maternal sensitivity. To te st the moderating effect of maternal sensitivity on the associ ations between timing of initial onset of maternal depression during the first 36 months postpartum and child social competence and display of problem behaviors at first grade, the interaction term of maternal depression at a given time point (six, 15, 24 or 36 months) times maternal sensitivity at the corresponding time point (six, 15, 24 or 36 months) was added to each final model (sociodemographic control variables + mate rnal depression + maternal sensitivity). Maternal sensitivity was not assessed at one month postpartum thus no interaction term of maternal depression initiati ng at one month and maternal sensitivity was tested. A significant interaction term indi cates the presence of a mode rating effect. OLS regression results are summarized in Table 69. Note that two separate criteria for judging statistical significance are presented in Table 69, p<.05 as the typica l criteria utilized, and a Bonferroni-adjusted p-value of <.01 given that five models were tested for each dependent variable (0.05/5=0.01).

PAGE 297

275 Table 69 OLS Regression Results Testing th e Moderating Effect of Maternal Sensitivity on the Association Between Timing of Initial Onset of Maternal Depression on Child Social Comp etence and Display of Problem Behaviors Maternal Depression X Maternal Sensitivity Model 1 Model 21 Model 32 Model 43 Model 54 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Dependent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Maternal Report of Social Competence Total na -2.15 (-0.61) -7.05 (-1.41) 4.00 (1.13) -3.36 (-0.77) Cooperation na -0.23 (-0.31) 1.28 (-1.21) -0.28 (-0.37) 0.05 (0.05) Assertion na -0.43 (-0.77) 0.11 (0.15) 1.26 (2.29)* -0.41 (-0.59) Responsibility na -0.24 (-0.38) -0.98 (-1.07) 0.46 (0.70) -1.55 (-1.93) Self Control na 0.03 (0.04) -2.16 (-1.86) 0.91 (1.11) -0.31 (-0.31) Peer Competence na -0.54 (-0.84) -0.97 (-1.06) 1.19 (1.84) 0.48 (0.59) Continued on the next page 1 Maternal depression at six months tim es maternal sensitivity at six months. 2 Maternal depression at 15 months times maternal sensitivity at 15 months. 3 Maternal depression at 24 months times maternal sensitivity at 24 months. 4 Maternal depression at 36 months times maternal sensitivity at 36 months.

PAGE 298

276 Table 69 (Continued) Maternal Depression X Maternal Sensitivity Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Dependent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Teacher Report of Social Competence Total na 4.31 (1.30) -4.38 (-0.93) 3.36 (1.00) 3.04 (0.74) Cooperation na 0.55 (0.58) 0.36 (0.26) -0.62 (-0.64) -0.10 (-0.08) Assertion na 0.96 (1.02) -3.60 (-2.75)** 0.38 (0.41) 1.24 (1.10) Self Control na 1.71 (1.98)* -0.02 (-0.01) 1.61 (1.87) -0.09 (-0.09) Peer Competence na 1.29 (1.51) -1.53 (-1.29) 1.08 (1.27) 0.31 (0.30) Maternal Report of Problem Behaviors Total Problem Behaviors na -0.34 (-0.15) 2.45 (0.77) 1.01 (0.44) -2.20 (-0.78) Internalizing Problems na 0.39 ( 0.19) 1.95 (0.66) 0.78 (0.37) -1.74 (-0.67) Externalizing Problems na -0.32 (-0.14) 3.76 (1.17) 2.01 (0.89) -1.88 (-0.66) *Indicates statistical significance at p<.05. **Indicates statistical significance at p<.01.

PAGE 299

277 Some evidence supports a moderating effect of maternal sensitivity on the relationship between timing of initial onset of maternal depression and child social competence. For example, the relationship be tween maternal depres sion initiating at six months postpartum and teacher report of self control was moderated by maternal sensitivity at six months, with depression ha ving the greatest impact among less sensitive mothers (interaction term estimates: b= 1.71, t=1.98, p<.05). Yet still, maternal sensitivity was detected to moderate th e association between maternal depression initiating at 24 months and mate rnal report of assertion. Fi rst time exposure to maternal depression at 24 months had a more negative impact on maternal report of assertion at first grade among mothers who rated lower in maternal sensitivity at 24 months (interaction term estimates: b=1.26, t=2.29, p<.05) But the opposite was true for teacher report of assertion. Although maternal se nsitivity at 15 months was a significant moderator on the association between maternal depression initiating at 15 months and teacher report of assertion, exposure to firs t time depression at 15 months had the most negative impact among more sensitive mother s (interaction term estimates: b=-3.60, t=2.75, p<.01.

PAGE 300

278 Hypothesis 4.2. Hypothesis 4 stated that the effect of maternal depression (timing of initial onset) on child social competence and display of problem behaviors at first grade is weaker among children exposed to nonmaternal care. To test the moderating effect of exposure to nonmaternal care on the associa tions between timing of initial onset of maternal depression during the first 36 months postpartum and child social competence and display of problem behaviors at first grade, the interaction term of maternal depression at a given time poi nt (six, 15, 24 or 36 months) times exposure to nonmaternal care at the corresponding time point (six, 15, 24 or 36 months) was added to each final model (sociodemographic contro l variables + maternal depression + maternal sensitivity + exposure to nonmaternal care ). The moderating effect of ever exposure to nonmaternal care at one month was not expl ored given that maternal se nsitivity, a control variable, was not assessed in the NICHD SECC at one month postpartum. Exposure to nonmaternal care was operationalized in tw o ways: (1) ever exposure to nonmaternal (despite type of arrangement); and (2) ever exposure to center-based nonmaternal care. Ever exposure to nonmaternal care consisted of two response options (no and yes), and ever exposure to center-based nonmaternal ca re consisted of three response options (maternal care only, non-center-based nonmate rnal care, and center-based nonmaternal care). A significant interaction term indicate s the presence of a moderating effect. Note that two separate criteria fo r judging statistical significance were applied including p<.05 as the typical criteria utilized, and a Bonferroni adjusted pvalue of <.005 given that ten models were tested for each dependent variable (0.05/10=0.005). No significant interaction effects were detected using the B onferroni adjusted p va lue. OLS regression

PAGE 301

279 results for the moderating effect of ever e xposure to nonmaternal care at each time point (among those not previously exposed to matern al depressed) are summarized in Table 70.

PAGE 302

280 Table 70 OLS Regression Results Testing th e Moderating Effect of Ever Ex posure to Nonmaternal Care on the Association Between Timing of Initial Onset of Maternal Depression on Child So cial Competence and Display of Problem Behaviors Maternal Depression X Ever Exposure to Nonmaternal Care Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Dependent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Maternal Report of Social Competence Total na 2.07 (0.47) -5.59 (-0.77) 1.96 (0.28) 0.70 (0.08) Cooperation na 0.60 (0.64) -0.62 (-0.40) 0.07 (0.05) 1.95 (1.00) Assertion na 0.06 (0.08) 0.67 (-0.59) 0.77 (0.70) -1.01 (-0.69) Responsibility na 0.22 (0.26) 0.37 (0.27) 1.05 (0.81) -1.27 (-0.74) Self Control na 0.52 (0.50) -3.05 (-1.80) -0.72 (-0.44) 0.23 (0.11) Peer Competence na 0.01 (0.01) -1.40 (-1.06) -0.37 (-0.29) -0.62 (-0.36) Continued on the next page

PAGE 303

281 Table 70 (Continued) Maternal Depression X Ever Exposure to Nonmaternal Care Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Dependent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Teacher Report of Social Competence Total na 8.58 (2.05)* 5.53 (0.81) 4.93 (0.74) -3.06 (8.68) Cooperation na 1.81 (1.50) 0.11 (-0.06) 0.78 (0.40) -4.88 (-1.92) Assertion na 2.91 (2.47)* 3.25 (1.70) 2.55 (1.39) 1.66 (0.69) Self Control na 1.81 (1.65) 0.36 (0.21) 1.16 (0.68) 0.16 (0.07) Peer Competence na 2.65 (2.47)* 1.21 (0.70) 2.49 (1.47) 0.92 (0.42) Maternal Report of Problem Behaviors Total Problem Behaviors na -2.59 (-0.89) 4.69 (1.01) -2.90 (-0.64) -2.39 (-0.40) Internalizing Problems na 0.33 ( 0.12) 5.42 (1.27) -2.51 (-0.60) 2.43 (0.44) Externalizing Problems na -3.41 (1.16) 8.17 (1.75) -2.60 (-0.58) -3.11 (-0.51) *Indicates statistical significance at p<.05.

PAGE 304

282 Ever exposure to nonmaternal care at si x months moderated the relationship between maternal depression at six months postpartum (among children not previously exposed) and teacher reports of total soci al competence (interaction term estimates: b=8.58, t=2.05, p<.05), assertion (interaction term estimates: b=2.91, t=2.47, p<.05), and peer competence (interaction term estim ates: b=2.65, t=2.47, p<.05). Among children ever exposed to nonmaternal care at six mont hs, maternal depression initiating between the first and sixth month postpartum and still present at six months had a weaker impact on teacher reports of total social competence, assertion, and peer competence. Thus, in the context of initial depre ssion at six months, exposure to nonmaternal care at six months had a buffering influence on teacher repo rted social competen ce (total, assertion, and peer competence). OLS regression results for the moderating eff ect of ever exposure to center-based nonmaternal care at each time point (among t hose not previously exposed to maternal depressed) are summarized in Table 71. Once again, no significan t interaction terms were detected using the Bonferroni adjusted p-value of <.005; all si gnificant in teraction terms presented are at p<.05. Ever exposure to center-based nonmaterna l care at six months moderated the relationship between maternal depression at six months (among children not previously exposed) and teacher report of assertion (interaction term estimates: b=2.16, t=2.23, p<.05). The negative impact of depression at six months, among children not previously exposed to maternal depression, on teacher report of assertion was weaker among children ever exposed to cente r-based nonmaternal care at six months. A similar trend was found for the moderating effect of expos ure to center-based nonmaternal care at 36

PAGE 305

283 months on the relationship between maternal depression at 36 months (among children not previously exposed) and maternal report of cooperation (interaction term estimates: b=2.00, t=1.99, p<.05). Exposure to center-b ased nonmaternal care at 36 months buffered the negative impact of maternal depression on cooperation.

PAGE 306

284 Table 71 OLS Regression Results Testing the Moderati ng Effect of Ever Exposure to Center-Bas ed Nonmaternal Care on the Association Between Timing of Initial Onset of Mate rnal Depression on Child Social Compet ence and Display of Problem Behaviors Maternal Depression X Ever Exposu re to Center-Based Nonmaternal Care Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Dependent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Maternal Report of Social Competence Total na 1.17 (0.32) -5.44 (-0.97) -0.51 (-0.14) 3.58 (0.75) Cooperation na -0.14 (-0.19) -0.54 (-0.46) 0.10 (0.12) 2.00 (1.99)* Assertion na -0.19 (-0.32) 1.00 (-1.14) -0.05 (-0.09) -0.27 (-0.36) Responsibility na 0.28 (0.41) 0.37 (0.36) -0.03 (-0.05) 0.69 (0.78) Self Control na 0.85 (1.00) 2.32 (-1.78) -0.25 (-0.29) -0.24 (-0.22) Peer Competence na -0.25 (-0.38 ) -1.94 (-1.91) -0.10 (-0.15) 0.32 (0.36) Continued on the next page

PAGE 307

285 Table 71 (Continued) Maternal Depression X Ever Exposu re to Center-Based Nonmaternal Care Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Dependent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Teacher Report of Social Competence Total na 5.08 (1.48) 2.16 (0.41) 4.53 (1.28) 2.80 (0.63) Cooperation na 0.79 (0.79) 0.40 (0.26) 1.11 (1.08) -0.78 (-0.60) Assertion na 2.16 (2.23)* 2.17 (1.48) 1.25 (1.28) 0.94 (0.76) Self Control na 0.88 (0.98) -0.77 (-0.57) 0.75 (0.82) 1.02 (0.89) Peer Competence na 1.53 (1.73) -0.05 (-0.05) 1.50 (1.67) 0.90 (0.80) Maternal Report of Problem Behaviors Total Problem Behaviors na -0.97 (-0.41) 3.12 (0.87) -1.86 (-0.78) -0.97 (-0.32) Internalizing Problems na 1.90 ( 0.87) 4.56 (1.39) -0.26 (-0.12) 0.48 (0.17) Externalizing Problems na -1.98 (-0.82) 3.68 (1.02) -1.24 (-0.52) 0.30 (0.10) *Indicates statistical significance at p<.05.

PAGE 308

286 Severity of Symptoms Hypothesis 1.3. Hypothesis 1.3 stated that se verity of maternal depres sion along the trajectory of initial onset is associated with child social competence and display of problem behaviors at first grade, with severity of symptoms ha ving the greatest negative impact with earlier onset episodes. To understand the impact of depression severity, maternal depression scores obtained from the CES-D (ranging from zero to 60) were recoded so that (1) individuals scoring below the cut point designating depression (<16) had values of zero; and (2) individuals experienci ng depression (>15) had values equal to 15 less than their original CES-D scores. For example, a woma n scoring 16 on the CES-D was recoded as having a value of one; a woman scoring 17 on the CES-D was recoded as having a value of two, et cetera. The pur pose of recoding the maternal depression variables in this manner was to assess the impact of symp tom severity above and beyond the threshold level recommended by the CES-D. To test hypothesis 1.3, a series of mode ls were tested. For each dependent variable, the dependent variab le was first regressed agains t maternal depression at one month, continuously scored, controlling fo r the sociodemographi c control variables (n=679) (Model 1). Next, among those not prev iously depressed (n=558), the dependent variables were regressed against maternal de pression at six months, continuously scored, controlling for the sociodemographic control variables (Model 2). Then, among those not previously depressed (n=510), the depe ndent variables were regressed against maternal depression at 15 months, c ontinuously scored, controlling for the sociodemographic control variables (Model 3) The same process was conducted to test

PAGE 309

287 the impact of severity of maternal depre ssion at 24 and 36 months, among samples of previously nondepressed women of 486 a nd 447 respectively (Models 4 and 5). The significance of each maternal depression variab le across models was determined by the pvalue associated with its unstandardized regression coefficient. A summary of significant fi ndings from OLS regression models is presented in Table 72. An ‘X’ in a designated cell indicate s that a significant a ssociation between the independent variable and dependent variable was detected at p<.05, controlling for all other variables included in the model. An ‘na’ in a designated cell indicates that a particular independent variable was not tested in a given mo del (for example, depression severity at one month is not included in M odel 2 given that the population for Model 2 is among those not previously depressed). Only dependent variables with at least one significant main effect of depression severi ty across models tested (models 1-5) are presented in Table 72 including: maternal re port of total social competence, maternal report of self control, maternal report of peer competence, te acher report of peer competence, and maternal reports of both total problem behaviors and externalizing problem behaviors. No main effect of mate rnal depression was dete cted across models 15 for maternal report of coope ration, maternal re port of assertion, ma ternal report of responsibility, teacher report of total social competence, teacher report of cooperation, teacher report of assertion, teacher report of self control, and maternal report of internalizing problem behaviors.

PAGE 310

288 Table 72 Summary of Significant Findings from OLS Regression Testing the Impact of Severity of Maternal Depression Along the Trajectory of Initial Onset on Child Social Competence and Display of Problem Behaviors Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All ND1 ND ND ND Maternal Report of Total Social Competence Child Sex X ns ns X X Child Birth Order X X X X X Maternal Age ns ns ns ns ns Maternal Race X X X X X Maternal Ethnicity ns ns ns ns ns Maternal Education X X X X X Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depression Severity at On e Month X na na na na Depression Severity at Six Months na ns na na na Depression Severity at 15 Months na na ns na na Depression Severity at 24 Months na na na ns na Depression Severity at 36 Months na na na na ns Maternal Report of Self Control Child Sex ns ns ns ns ns Child Birth Order X X X X X Maternal Age ns ns ns ns ns Maternal Race ns ns X X X Maternal Ethnicity ns ns ns ns ns Maternal Education X X ns ns ns Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns X Breastfeeding at One Month ns ns ns ns ns Depression Severity at On e Month X na na na na Depression Severity at Six Months na ns na na na Depression Severity at 15 Months na na ns na na Depression Severity at 24 Months na na na ns na Depression Severity at 36 Months na na na na ns Continued on the next page

PAGE 311

289 Table 72 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All ND ND ND ND Maternal Report of Peer Competence Child Sex ns ns X ns ns Child Birth Order X X X X X Maternal Age ns ns ns ns ns Maternal Race X X X X X Maternal Ethnicity ns ns ns ns ns Maternal Education X ns ns ns ns Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depression Severity at On e Month X na na na na Depression Severity at Six Months na ns na na na Depression Severity at 15 Months na na ns na na Depression Severity at 24 Months na na na ns na Depression Severity at 36 Months na na na na ns Teacher Report of Peer Competence Child Sex X X X X X Child Birth Order ns X X X X Maternal Age ns ns ns ns ns Maternal Race X ns ns ns ns Maternal Ethnicity ns X ns ns X Maternal Education X X X X X Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month X X ns ns ns Depression Severity at One Month ns na na na na Depression Severity at Six Months na X na na na Depression Severity at 15 Months na na ns na na Depression Severity at 24 Months na na na ns na Depression Severity at 36 Months na na na na ns Continued on the next page

PAGE 312

290 Table 72 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All ND ND ND ND Maternal Report of Total Problem Behaviors Child Sex ns ns ns ns ns Child Birth Order ns ns X X X Maternal Age ns ns ns ns ns Maternal Race ns ns ns ns ns Maternal Ethnicity ns ns ns ns ns Maternal Education ns X X X X Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depression Severity at On e Month X na na na na Depression Severity at Six Months na ns na na na Depression Severity at 15 Months na na ns na na Depression Severity at 24 Months na na na ns na Depression Severity at 36 Months na na na na ns Maternal Report of Externalizing Behaviors Child Sex ns ns ns ns ns Child Birth Order ns ns X X X Maternal Age ns ns ns ns ns Maternal Race ns ns ns ns ns Maternal Ethnicity ns ns ns ns ns Maternal Education X X X X X Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depression Severity at On e Month X na na na na Depression Severity at Six Months na ns na na na Depression Severity at 15 Months na na ns na na Depression Severity at 24 Months na na na ns na Depression at 36 Months na na na na ns 1ND denotes ‘Never Depressed’

PAGE 313

291 Results from Table 72 above support hypothesis 1.3 for at least some of the dependent variables investigated: that severity of maternal depressi on along the trajectory of initial onset is associated with child social competence and display of problem behaviors at first grade, with severity of symptoms having the greatest negative impact with earlier onset episodes For maternal reports of total social competence, self control, peer competence, total problem behaviors, and externalizing problem behaviors, depression severity at one month postpartum significantly impacted levels of social competence and problem behaviors. As depr ession severity increased, maternal rated social competence decreased and mother rate d display of problem behaviors increased. Depression severity did not significantly influence maternal rated child social competence and display of problem behaviors among depressive episod es that initiated after the first postpartum month. Depressi on severity at six, 15, 24, or 36 months, among children not previously expos ed to depression, had no signi ficant impact on maternal reports of total social competence, self control, peer competence, total problem behaviors, or externalizing problem behavior s. For teacher rated social competence, depression severity at six months among child ren not previously exposed to depression significantly impacted teacher report of p eer competence. As depression severity increased, teacher rated peer competence decreas ed. No other significant associations of depression severity along the trajectory of in itial onset were detected. See Appendix D, Tables D1 through D6, for partial model re sults highlighting parameter estimates (b coefficients), t values, and model fit statistics from regression models testing the impact of depression severity on child social comp etence or display of problem behaviors, where

PAGE 314

292 at least one significant main effect of maternal depressi on was detected across models tested (models 1-5). Hypothesis 2.3. Hypothesis 2, that the relationship betw een maternal depression (severity of symptoms above the CES-D cut point desi gnating depression) and child social competence and display of problem behaviors at first grade is mediated by maternal sensitivity, was tested by adding maternal sensitivity to the models containing the sociodemographic control variables and depres sion severity at a given time point. Given that maternal sensitivity was only assesse d in the NICHD SECC at six, 15, 24, and 36 months (no one-month assessment), mediating e ffects of maternal sensitivity were only tested in models for which depression severi ty had a significant main effect at six, 15, 24 or 36 months (teacher report of peer competence). Note that the significant main effect of depression severity on teacher report of peer competence was detected at six months postpartum, among children not previously e xposed to maternal depression. Although significant main effects of depre ssion severity were detected for maternal reports of total social competence, self control, peer competence, total problem behaviors and externalizing problems, these main effects we re detected at one month postpartum only. No further exploration of a mediating effect of maternal sensitivity was tested in these models given that maternal sensitivity at one month postpartum was not assessed. Thus, the mediating influence of maternal sens itivity was only explored for impact of depression severity at six months postpartu m on teacher report of peer competence, among children not previously exposed to matern al depression. For this model, maternal

PAGE 315

293 sensitivity was operationalized as the mother’s observed rating of ma ternal sensitivity at six months postpartum (response options ranged from one to four). Prior to entering maternal sensitivity to the model containing depression severity, continuously scored, and the block of sociode mographic control vari ables, the following associations were explored as a preliminary step in establishing mediation: (1) depression severity at six months postpartum, among those children not previously exposed to depression, and maternal sensitivity at six m onths; and (2) maternal sensitivity at six months and teacher report of peer competen ce, while controlling for depression severity at six months postpartum. Depression severity at six months was statistically associated with maternal sensitivity at six months (b=-0.01, t=-2.32, p<.05). However, maternal sensitivity at six months was not significantl y associated with teacher report of peer competence, while controlling for depression severity at six months postpartum (b=0.14, t=0.59, ns). Thus no further examination of a me diating effect of maternal sensitivity on the relationship between depression severi ty and teacher report of assertion was examined. The hypothesis was not supported.

PAGE 316

294 Hypothesis 3.3. Hypothesis 3 stated that the effect of maternal depression (depression severity) on child social competence and display of pr oblem behaviors at fi rst grade is stronger among mothers who rate low in maternal sens itivity. To test the moderating effect of maternal sensitivity on the associations between depression severity along the trajectory of initial onset during the first 36 months postpartum and child social competence and display of problem behaviors at first grade, th e interaction term of depression severity at a given time point (six, 15, 24 or 36 mont hs) times maternal sensitivity at the corresponding time point (six, 15, 24 or 36 months) was added to each final model (sociodemographic control variables + depression severity + maternal sensitivity). Maternal sensitivity was not assessed at one month postpartum thus no interaction term of depression severity at one month and mate rnal sensitivity was te sted. A significant interaction term indicates the presence of a moderating effect. OLS regression results are summarized in Table 73 Note that two se parate criteria for judging statistical significance are presented in Table 73 p<.05 as the typical criteria utilized, and a Bonferroni-adjusted p-value of <.01 given that five models were tested for each dependent variable (0.05/5=0.01). Evidence of a moderating effect of maternal sensitivity was detected for the interaction term of depression severity at 24 months, among children not previously exposed to depression, and maternal sensit ivity at 24 months (b=0.13, t=2.46, p<.05). Depression severity had a str onger impact on maternal report of assertion among children with mothers rated lower in maternal sensitivity.

PAGE 317

295 Similar trends were detected for teache r reports of total social competence (interaction term estimates: b=0.77, t=2.44, p<.05) assertion (interaction term estimates: b=0.22, t=2.47, p<.05), self control (inter action term estimates: b=0.21, t=2.56, p<.05), and peer competence (interaction term estimates: b=0.21, t=2.63, p<.01); however the significant interaction terms were with depres sion severity at six months, among children not previously exposed to depr ession, and maternal sensitiv ity at six months. Depression severity at six months had a more negative impact on teacher reports of total social competence, assertion, self control, and peer competence among children of mothers rated lower in maternal sensitivity.

PAGE 318

296 Table 73 OLS Regression Results Testing the Moderati ng Effect of Maternal Sensitivity on the Association Between Depression Severity on Child Social Competence and Display of Problem Behaviors Depression Severity X Maternal Sensitivity Model 1 Model 21 Model 32 Model 43 Model 54 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Dependent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Maternal Report of Social Competence Total na -0.04 (-0.11) -1.12 (-1.64) 0.19 (0.58) -0.20 (-0.43) Cooperation na 0.01 (0.08) 0.17 (-1.15) -0.05 (-0.65) 0.01 (0.06) Assertion na 0.01 (0.17) 0.05 (-0.48) 0.13 (2.46)* 0.00 (0.06) Responsibility na 0.01 (0.19) -0.13 (-1.06) 0.00 (0.07) -0.06 (-0.66) Self Control na -0.01 (-0.23) -0.30 (-1.89) 0.02 (0.26) -0.08 (-0.73) Peer Competence na -0.02 (-0.31) -0.17 (-1.35) 0.12 (1.90) 0.07 (0.78) Continued on the next page 1 Maternal depression at six months tim es maternal sensitivity at six months. 2 Maternal depression at 15 months times maternal sensitivity at 15 months. 3 Maternal depression at 24 months times maternal sensitivity at 24 months. 4 Maternal depression at 36 months times maternal sensitivity at 36 months.

PAGE 319

297 Table 73 (Continued) Depression Severity X Maternal Sensitivity Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Dependent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Teacher Report of Social Competence Total na 0.77 (2.44)* -0.42 (-0.66) 0.11 (0.33) -0.23 (-0.51) Cooperation na 0.16 (1.78) -0.06 (-0.34) -0.05 (-0.53) -0.13 (-1.04) Assertion na 0.22 (2.47)* 0.27 (-1.49) 0.01 (0.17) 0.07 (0.60) Self Control na 0.21 (2.56)* -0.03 (-0.17) 0.05 (0.64) -0.14 (-1.23) Peer Competence na 0.21 (2.63)** -0.09 (-0.57) 0.02 (0.22) -0.07 (-0.59) Maternal Report of Problem Behaviors Total Problem Behaviors na -0.33 (-1.51) 0.51 (1.17) 0.23 (1.05) -0.47 (-1.54) Internalizing Problems na -0.20 (-1.00) 0.30 (0.74) 0.16 (0.82) -0.50 (-1.79) Externalizing Problems na -0.30 (-1.34) 0.52 (1.20) 0.41 (1.93) -0.38 (-1.24) *Indicates statistical significance at p<.05. ** Indicates statistical significance at p<.01.

PAGE 320

298 Hypothesis 4.3. Hypothesis 4 stated that the effect of maternal depression (depression severity) on child social competence and display of pr oblem behaviors at first grade is weaker among children exposed to nonmaternal care. To test the moderating effect of exposure to nonmaternal care on the associations betw een depression severity along the trajectory of initial onset during the first 36 months postpartum and child social competence and display of problem behaviors at first grade, th e interaction term of depression severity at a given time point (six, 15, 24 or 36 months ) times exposure to nonmaternal care at the corresponding time point (six, 15, 24 or 36 months) was added to each final model (sociodemographic control variables + depr ession severity + maternal sensitivity + exposure to nonmaternal care). The moderati ng effect of ever e xposure to nonmaternal care at one month was not expl ored given that maternal se nsitivity, a control variable, was not assessed in the NICHD SECC at one month postpartum. Exposure to nonmaternal care was operationalized in tw o ways: (1) ever exposure to nonmaternal (despite type of arrangement); and (2) ever exposure to center-based nonmaternal care. Ever exposure to nonmaternal care consisted of two response options (no and yes), and ever exposure to center-based nonmaternal ca re consisted of three response options (maternal care only, non-center-based nonmate rnal care, and center-based nonmaternal care). A significant interaction term indicate s the presence of a moderating effect. Two criteria for assessing statistical signi ficance were applied including p<.05 and a Bonferroni adjusted p-value of <.005 to acc ount for the ten models tested for each dependent variable (0.05/10=0.005) OLS regression results fo r the moderating effect of ever exposure to nonmaternal care at each time point (among those not previously

PAGE 321

299 exposed to maternal depression) are summari zed in Table 74 Note that no significant interaction terms were detected at p<.005.

PAGE 322

300 Table 74 OLS Regression Results Testing th e Moderating Effect of Ever Ex posure to Nonmaternal Care on the Association Between Depression Severity Along the Trajectory of Initial Onset on Ch ild Social Competence and Display of Problem Behaviors Depression Severity X Ever Exposure to Nonmaternal Care Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Dependent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Maternal Report of Social Competence Total na 0.08 (0.19) -0.89 (-0.89) -0.43 (-0.71) -0.09 (-0.02) Cooperation na 0.07 (0.82) 0.19 (-0.90) -0.01 (-0.09) 0.73 (0.88) Assertion na 0.02 (0.25) 0.09 (-0.57) -0.06 (-0.67) 0.05 (0.09) Responsibility na -0.04 (-0.46) 0.03 (0.15) -0.01 (-0.05) -0.80 (-1.09) Self Control na 0.01 (0.15) 0.33 (-1.44) -0.17 (-1.20) -0.23 (-0.25) Peer Competence na 0.02 (0.27) -0.21 (-1.14) -0.13 (-1.16) -0.17 (-0.23) Continued on the next page

PAGE 323

301 Table 74 (Continued) Depression Severity X Ever Exposure to Nonmaternal Care Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Dependent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Teacher Report of Social Competence Total na 0.49 (1.22) 0.98 (1.05) 0.61 (1.08) -2.35 (-0.63) Cooperation na 0.09 (0.82) 0.18 (-0.66) 0.06 (0.37) -2.09 (-1.91) Assertion na 0.22 (1.95) 0.56 (2.15)* 0.31 (2.00)* 0.20 (0.19) Self Control na 0.11 (1.07) 0.13 (0.54) 0.04 (0.26) -0.02 (-0.02) Peer Competence na 0.17 (1.68) 0.26 (1.08) 0.12 (0.85) -0.11 (-0.12) Maternal Report of Problem Behaviors Total Problem Behaviors na -0.07 (-0.27) 0.50 (0.79) 0.20 (-0.52) -1.32 (-0.51) Internalizing Problems na 0.12 ( 0.47) 0.65 (1.11) -0.25 (-0.70) -0.63 (-0.27) Externalizing Problems na -0.05 (0.19) 0.81 (1.27) -0.28 (-0.72) -1.16 (-0.45) *Indicates statistical significance at p<.05.

PAGE 324

302 Significant interaction terms were only dete cted for teacher report of assertion at 15 and 24 months. For this dependent vari able, the hypothesis was supported. Results indicated that depression severity at 15 months, among children not previously exposed to depression, had a weaker impact on teacher report of assertion if the child had ever been exposed to nonmaternal care at 15 m onths (interaction term estimates: b=0.56, t=2.15, p<.05). Furthermore, ever exposure to nonmaternal care at 24 months buffered the negative impact of depression severity at 24 months, among children not previously exposed to depression, on teacher report of assertion (interaction term estimates: b=0.31, t=2.00, p<.05). OLS regression results for the moderating eff ect of ever exposure to center-based nonmaternal care at each time point (among t hose not previously exposed to maternal depressed) are summarized in Table 75 Once again, no significant interaction terms were detected using the Bonferroni adjusted p-va lue of <.005. All signifi cant interaction terms presented are at p<.05. Significant interaction terms were only dete cted for teacher report of assertion at 15 and 24 months. For this dependent vari able, the hypothesis was supported. Results indicated that depression severity at 15 months, among children not previously exposed to depression, had a weaker impact on teacher re port of assertion at fi rst grade if the child had ever been exposed to center-based nonmate rnal care at 15 mont hs (interaction term estimates: b=0.54, t=2.14, p<.05). Furthermor e, ever exposure to center-based nonmaternal care at 24 months buffered the nega tive impact of depression severity at 24 months, among children not previously expos ed to depression, on teacher report of assertion at first grade (interacti on term estimates: b=0.23, t=1.97, p<.05).

PAGE 325

303 Table 75 OLS Regression Results Testing the Moderati ng Effect of Ever Exposure to Center-Bas ed Nonmaternal Care on the Association Between Depression Severity Along the Trajectory of Initial On set on Child Social Competence and Display of Problem Behaviors Depression Severity X Ever Exposu re to Center-Based Nonmaternal Care Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Dependent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Maternal Report of Social Competence Total na -0.06 (-0.18) -0.94 (-0.97) -0.42 (-0.92) 0.26 (0.37) Cooperation na -0.00 (-0.02) 0.16 (-0.80) -0.04 (-0.44) 0.28 (1.86) Assertion na -0.02 (-0.36) 0.15 (-0.97) -0.03 (-0.45) -0.06 (-0.55) Responsibility na -0.03 (-0.45 ) 0.24 (0.14) -0.03 (-0.39) 0.01 (0.06) Self Control na 0.03 (0.40) 0.32 (-1.43) -0.14 (-1.30) -0.05 (-0.30) Peer Competence na -0.03 (-0.44 ) -0.27 (-1.53) -0.06 (-0.71) 0.07 (0.55) Continued on the next page

PAGE 326

304 Table 75 (Continued) Depression Severity X Ever Exposu re to Center-Based Nonmaternal Care Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Dependent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Teacher Report of Social Competence Total na0.36 (1.10)0.94 (1.05)0.69 (1.62)-0.36 (-0.55) Cooperation na0.06 (0.67)-0.11 (-0.41)0.12 (0.98)-0.09 (-0.46) Assertion na0.17 (1.82)0.54 (2.14)*0.23 (1.97)*-0.11 (-0.61) Self Control na0.09 (1.00)0.11 (0.48)0.07 (0.67)-0.05 (-0.27) Peer Competence na0.12 (1.46)0.22 (0.95)0.12 (1.07)-0.11 (-0.69) Maternal Report of Problem Behaviors Total Problem Behaviors na0.00 (-0.01)0.41 (0.67)-0.09 (-0.29)0.13 (0.28) Internalizing Problems na0.17 (0.82)0.64 (1.13)-0.00 (-0.02)0.09 (0.21) Externalizing Problems na-0.01 (-0.04)0.62 (1.00)-0.09 (-0.30)0.16 (0.35) *Indicates statistical significance at p<.05.

PAGE 327

305 Chronicity of Symptoms Hypothesis 1.4. Hypothesis 1.4 stated that ch ronicity of maternal depr ession is associated with child social competence and display of problem behaviors at first grade, with chronicity of symptoms having the greates t negative impact at later tim e points. To understand the impact of depression chronicity, a sequence of dummy variables we re created at each time point (one month, six months, 15 mont hs, 24 months, and 36 months) to compare women with either (1) new or first time depression; (2) remitted depression (had a previous depressive episode but no current depression); (3) recurr ent depression (had a previous depressive episode followed by re mission with current depression); or (4) chronic depression (continuously depressed). For example, to understand the impact of depression chronicity at one month postpartum, a single dummy variable was used (new depression). For depression ch ronicity at six months, th e following dummy variables were used: new depression, remitted depressi on, and chronic depression. For depression chronicity at 15, 25, and 36 months, the fo llowing dummy variables were used: new depression, remitted depression, recurrent depression, and chronic depression. Women with no depression were the reference categ ory for each sequence of dummy variables across time points. To test hypothesis 1.4, for each dependent variable, the dependent variable was first regressed against the depression ch ronicity dummy vari ables at one month controlling for the sociodemographic control variables (n=679) (Model 1). Next, the dependent variables were regressed against th e depression chronicity dummy variables at six months controlling for the sociodemogra phic control variables (n=679) (Model 2).

PAGE 328

306 The same process was conducted to test the im pact of depression chronicity at 15, 24 and 36 months, among the entire sample of 679 children (Models 3, 4 and 5). The significance of each depression chronicity dummy variable within and across models was determined by the p-value associated with its unstandardized regressi on coefficient. To understand the relative importance of each significant dummy variable within a given model, standardized regression coefficients were compared; to understand the relative importance of each significant dummy variable across the models, unstandardized regression coefficients were compared. No te that the population for each model in a series consisted of the entire sample of 679 children, not subsets of children with no previous exposure to maternal depre ssion at each assessment time point. A summary of significant fi ndings from OLS regression models is presented in Table 76. An ‘X’ in a designated cell indicate s that a significant a ssociation between the independent variable and dependent variable was detected at p<.05, controlling for all other variables included in the model. An ‘na’ in a designated cell indicates that a particular independent variable was not tested in a given mo del (for example, depression chronicity at one month was not included in Model 2 given that the sequence of dummy variables included in Model 2 to represent depression chronicity at six months incorporated depression chronicity at one m onth postpartum). Only dependent variables with at least one significant main effect of depression chro nicity across models tested (models 1-5) are presented in Table 76 incl uding: maternal report of total social competence, maternal report of cooperation, mate rnal report of asser tion, maternal report of responsibility, maternal repo rt of self control, maternal report of peer competence, teacher report of total social competence, teac her report of self control, teacher report of

PAGE 329

307 peer competence, maternal report of tota l problem behaviors, internalizing problem behaviors, and externalizing pr oblem behaviors. No main e ffect of depression chronicity was detected across models 1-5 for teacher re port of cooperation or teacher report of assertion.

PAGE 330

308 Table 76 Summary of Significant Findings from OLS Regression Testing the Impact of Depression Chronicity on Child Social Competence and Display of Problem Behaviors Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Population All All All All All Maternal Report of Total Social Competence Child Sex X X X X X Child Birth Order X X X X X Maternal Age ns ns ns ns ns Maternal Race X X X X X Maternal Ethnicity ns ns ns ns ns Maternal Education X X X X X Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depression Chronicity at On e Month ns na na na na Depression Chronicity at Si x Months na X na na na Depression Chronicity at 15 Months na na X na na Depression Chronicity at 24 Months na na na X na Depression Chronicity at 36 Months na na na na X Maternal Report of Cooperation Child Sex ns ns ns ns ns Child Birth Order ns ns ns ns ns Maternal Age ns ns ns ns ns Maternal Race ns ns ns ns ns Maternal Ethnicity ns ns ns ns ns Maternal Education X X X X X Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depression Chronicity at On e Month ns na na na na Depression Chronicity at Si x Months na ns na na na Depression Chronicity at 15 Months na na ns na na Depression Chronicity at 24 Months na na na X na Depression Chronicity at 36 Months na na na na X Continued on the next page

PAGE 331

309 Table 76 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Population All All All All All Maternal Report of Assertion Child Sex ns ns ns ns ns Child Birth Order X ns X X X Maternal Age ns ns ns ns ns Maternal Race X X X X X Maternal Ethnicity ns ns ns ns ns Maternal Education X X X ns X Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depression Chronicity at On e Month ns na na na na Depression Chronicity at Si x Months na X na na na Depression Chronicity at 15 Months na na X na na Depression Chronicity at 24 Months na na na X na Depression Chronicity at 36 Months na na na na X Maternal Report of Responsibility Child Sex ns ns ns ns ns Child Birth Order X X X X X Maternal Age ns ns ns ns ns Maternal Race ns ns ns ns ns Maternal Ethnicity ns ns ns ns ns Maternal Education X X ns X X Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depression Chronicity at On e Month ns na na na na Depression Chronicity at Si x Months na X na na na Depression Chronicity at 15 Months na na X na na Depression Chronicity at 24 Months na na na X na Depression Chronicity at 36 Months na na na na X Continued on the next page

PAGE 332

310 Table 76 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Population All All All All All Maternal Report of Self Control Child Sex ns ns ns ns ns Child Birth Order X X X X X Maternal Age ns ns ns ns ns Maternal Race ns ns ns ns ns Maternal Ethnicity ns ns ns ns ns Maternal Education X X X X X Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depression Chronicity at On e Month ns na na na na Depression Chronicity at Si x Months na ns na na na Depression Chronicity at 15 Months na na X na na Depression Chronicity at 24 Months na na na X na Depression Chronicity at 36 Months na na na na X Maternal Report of Peer Competence Child Sex ns ns ns ns ns Child Birth Order X X X X X Maternal Age ns ns ns ns ns Maternal Race X X X X X Maternal Ethnicity ns ns ns ns ns Maternal Education X X X ns X Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depression Chronicity at On e Month ns na na na na Depression Chronicity at Si x Months na X na na na Depression Chronicity at 15 Months na na X na na Depression Chronicity at 24 Months na na na X na Depression Chronicity at 36 Months na na na na X Continued on the next page

PAGE 333

311 Table 76 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Population All All All All All Teacher Report of Total Social Competence Child Sex ns ns ns ns ns Child Birth Order ns ns ns ns ns Maternal Age ns ns ns ns ns Maternal Race X X X X X Maternal Ethnicity ns ns ns ns ns Maternal Education X X X X X Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns X X X Depression Chronicity at On e Month ns na na na na Depression Chronicity at Si x Months na ns na na na Depression Chronicity at 15 Months na na X na na Depression Chronicity at 24 Months na na na X na Depression Chronicity at 36 Months na na na na ns Teacher Report of Self Control Child Sex X X X X X Child Birth Order ns X X ns X Maternal Age ns ns ns ns ns Maternal Race X X X X X Maternal Ethnicity X X X ns X Maternal Education X X X X X Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depression Chronicity at On e Month ns na na na na Depression Chronicity at Si x Months na X na na na Depression Chronicity at 15 Months na na X na na Depression Chronicity at 24 Months na na na ns na Depression Chronicity at 36 Months na na na na ns Continued on the next page

PAGE 334

312 Table 76 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Population All All All All All Teacher Report of Peer Competence Child Sex X X X X X Child Birth Order ns ns ns ns ns Maternal Age ns ns ns ns ns Maternal Race X X X X X Maternal Ethnicity ns ns ns ns ns Maternal Education X X X X X Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month X X X X X Depression Chronicity at On e Month ns na na na na Depression Chronicity at Si x Months na X na na na Depression Chronicity at 15 Months na na X na na Depression Chronicity at 24 Months na na na ns na Depression Chronicity at 36 Months na na na na ns Maternal Report of Total Problem Behaviors Child Sex ns ns ns ns ns Child Birth Order ns X X ns ns Maternal Age ns ns ns ns ns Maternal Race ns ns ns ns ns Maternal Ethnicity ns ns ns ns ns Maternal Education ns ns ns ns ns Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depression Chronicity at One Month X na na na na Depression Chronicity at Si x Months na X na na na Depression Chronicity at 15 Months na na X na na Depression Chronicity at 24 Months na na na X na Depression Chronicity at 36 Months na na na na X Continued on the next page

PAGE 335

313 Table 76 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Population All All All All All Maternal Report of Intern alizing Problem Behaviors Child Sex ns ns ns ns ns Child Birth Order X X X X X Maternal Age ns ns ns ns ns Maternal Race X X X X X Maternal Ethnicity ns ns ns ns ns Maternal Education ns ns ns ns ns Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depression Chronicity at On e Month ns na na na na Depression Chronicity at Si x Months na ns na na na Depression Chronicity at 15 Months na na X na na Depression Chronicity at 24 Months na na na X na Depression Chronicity at 36 Months na na na na X Maternal Report of Externalizing Problem Behaviors Child Sex ns ns ns ns ns Child Birth Order ns ns ns ns ns Maternal Age ns ns ns ns ns Maternal Race ns ns ns ns ns Maternal Ethnicity ns ns ns ns ns Maternal Education X X X ns ns Partner/Husband at Home ns ns ns ns ns Receipt of Public Assistance ns ns ns ns ns Breastfeeding at One Month ns ns ns ns ns Depression Chronicity at One Month X na na na na Depression Chronicity at Si x Months na ns na na na Depression Chronicity at 15 Months na na X na na Depression Chronicity at 24 Months na na na X na Depression Chronicity at 36 Months na na na na X

PAGE 336

314 Partial model results highlighting signifi cant depression chronicity variables for dependent variables with at least one significa nt main effect of depression chronicity across models tested (models 1-5) are pr esented in Appendix E, Tables E1-E12. Examination of results reveal that for at least some of the dependent variables investigated, hypothesis 1.4, that depression chronicity is asso ciated with child social competence and display of problem behaviors at first grade with chr onicity of symptoms having the greatest negative impact at later tim e points, is supported. However, the time point at which depression chr onicity had the greatest magnitu de of negative influence on child social competence and display of pr oblem behaviors was 24 months postpartum rather than 36 months postpartum. This tr end was detected for the following dependent variables: maternal report of total social co mpetence, maternal report of self control, maternal report of peer competence, matern al report of total problem behaviors and maternal report of externa lizing problem behaviors. For maternal reports of total social competence, peer competence, and total problem behaviors, depression chronicity at six, 15, 24, and 36 months were statistically significant. Comparison of the unstandardized regression coeffi cients or b coefficients across six, 15, 24, and 36 months revealed that for each of these dependent variables, the parameter estimates for depression chronicity increased in magnitude up to 24 months postpartum, at which they began to decline. For example, the b coefficients for depression chronicity at si x, 15, 24, and 36 months postpartu m for maternal report of total social competence were –5.54, -9.03, -12.88, and –12.79 respectively. The same trend was observed for maternal report of p eer competence and maternal report of total

PAGE 337

315 problem behaviors in which the parame ter estimates were –0.79, -1.31, -2.29, and –2.27 for peer competence, and 3.67, 6.80, 9.40, a nd 9.20 for total problem behaviors. Furthermore, depression chronicity at 15, 24, and 36 months was significant for maternal report of self control and maternal report of externalizing problem behaviors, with depression chronicity at 24 months having the greatest magnitude of negative influence. For maternal report of self control, unstandardized regression coefficients at 15, 24, and 36 months for depression chronicity were –2.19, -3.48, and – 3.44. For maternal report of externalizing pr oblem behaviors, unstandardized regression coefficients at 15, 24, and 36 months fo r depression chronicity were 6.71, 10.26, and 10.18. Chronicity of depression was significan t at both 24 and 36 months postpartum for maternal report of cooperation, however the magnitude of negative impact was the same at both time points (b coeffi cient=-2.44 for depression chronicity at 24 months; b coefficient=-2.44 for depression ch ronicity at 36 months). Recurrent depression, when significant ac ross more than one model for a single dependent variable, was detected to have a stronger negative imp act on the dependent variable at later time points for both maternal report of assertion and maternal report of responsibility. Recurrent depr ession, or depression that recurs after remittance, was significant at both 24 and 36 months for these outcome variables, w ith recurrence at 36 months having the greatest magnitude of influe nce. For maternal re port of assertion, the b coefficients for depression recurrence at 24 and 36 months postpartum were –0.84 and –0.86. For maternal report of responsibility, the b coefficients for depression recurrence at 24 and 36 months postpartum were –0.96 and –1.37.

PAGE 338

316 For maternal reports of total problem behaviors and inte rnalizing problem behaviors, recurrent depression was also detected to be sign ificant across more than one model, specifically at 15, 24, and 36 months postpartum. This time however, it was recurrent depression at 24 months that had the greatest negative impact on the outcome variable. Unstandardized regression coeffici ents for depression recurrence at 15, 24, and 36 months for maternal reports of total pr oblem behaviors and in ternalizing problem behaviors were as follows: (1) 4.41, 5.94, 4.86; and (2) 5.62, 6.24, 4.48. For maternal report of externalizing problem behaviors, although recurrent depr ession was significant at 15, 24, and 36 months postpartum, depressi on recurrence at 15 months had the greatest negative impact on exte rnalizing problem behaviors (b coefficients of 3.62, 3.51, and 3.28 at 15, 24, and 36 months respectively. In addition to comparing significant dumm y variables across models for a single dependent variable (for example compari ng chronic depression at 24 months versus chronic depression at 36 months for maternal re port of total social competence), results from Appendix E, Tables E1-E12, also illu minate the relative importance of each significant dummy variable with in a single model. For example, both new and chronic depression were significant at six months for maternal reports of total social competence and peer competence. A comparison of sta ndardized regression co efficients, or Beta coefficients (B), for these si gnificant dummy variables reve aled that chronic depression had a slightly greater magnitude of infl uence (-0.08 versus –0.09) on total social competence, as well as for peer competence (-0.08 versus –0.11). Chronic depression and an additional dummy variable for depressi on chronicity were al so significant in a single model for maternal reports of peer competence and externalizing problem

PAGE 339

317 behaviors 15, 24, and 36 months. For maternal report of peer competence, both remitted and chronic depression were significant at 15 months, with chronic depression having slightly greater influence on peer compet ence (B=-0.08 versus –0.09). For both maternal reports of peer competence and externalizing problem behaviors, ch ronic depression at 24 and 36 months had larger standardized regressi on coefficients as compared to recurrent depression (B=-0.09 for recurrent depression an d B=-0.10 for chronic depression for peer competence at 24 months; B=-0.09 for recu rrent depression and B=-0.10 for chronic depression for peer competence at 36 m onths; B=0.08 for recurrent depression and B=0.12 for chronic depression for externaliz ing problem behaviors at 24 months; B=0.09 for recurrent depression and 0.12 for chroni c depression for externalizing problem behaviors at 36 months). This same tre nd was detected for recurrent and chronic depression at 15 months for externalizi ng problem behaviors (B=0.08 for recurrent depression and B=0.12 for chronic depression), as well as maternal report of total problem behaviors (B=0.10 for recurrent depre ssion and B=0.12 for chronic depression). Also significant in a single model were recurrent and chronic depression at 36 months for maternal report of total soci al competence. However it was recurrent depression at 36 months (depress ion that recurred after at least one single episode of remittance) that had the greatest magnitude of influence (B=-0.12) rather than chronic depression (B=-0.10). This same trend was dete cted for maternal re port of total problem behaviors at both 24 and 36 months. At 24 months, both recurrent and chronic depression were significant with recurrent depression having the greatest magnitude of influence (B=0.14 for recurrent depression a nd B=0.12 for chronic depression). At 36 months, recurrent and chroni c depression were again si gnificant with recurrent

PAGE 340

318 depression having the greatest magnitude of in fluence (B=0.13 for recurrent depression and B=0.11 for chronic depression). For mate rnal report of inte rnalizing problem behaviors, both recurrent depression and chronic depression we re significant at 15 months with recurrent depression having th e greatest effect on outcome (B=0.13 for recurrent depression and B=0.08 for chronic depr ession). So for at least some of the dependent variables, recurre nt depression, or depressi on that presents itself inconsistently, has the greatest magnitude of negative impact on outcome. Furthermore, for maternal report of assertion and teacher reports of total soci al competence, self control, and peer competence, recurrent depression rather than chronic depression emerged as a significant predictor of the de pendent variable. R ecurrent depression was statistically associated with maternal asse rtion at both 24 and 36 m onths postpartum; and recurrent depression at 15 months was statistically associated with teacher reports of total social competence, self control, and peer competence.

PAGE 341

319 Hypothesis 2.4. Hypothesis 2, that the relationship between maternal depression (chronicity of symptoms) and child social competence and disp lay of problem behaviors at first grade is mediated by maternal sensitivity, was tested by adding maternal sensitivity to the models containing the sociodemographic control va riables and depression chronicity dummy variables at a given time point Given that maternal sensit ivity was only assessed in the NICHD SECC at six, 15, 24, and 36 months (no one-month assessment), mediating effects of maternal sensitivity were only test ed in models for which depression chronicity had a significant main effect at six, 15, 24 or 36 months including: (1 ) maternal report of total social competence (depression chr onicity at six, 15, 24, and 36 months); (2) maternal report of cooperation (depression chro nicity at 24 and 36 months); (3) maternal report of assertion (depression chronicity at 6, 15, 24, and 36 months); (4) maternal report of responsibility (depression ch ronicity at six, 15, 24, and 36 months); (5) maternal report of self control (depression chronicity at 15, 24, and 36 months); (6) maternal report of peer competence (depression chronicity at six, 15, 24, and 36 months); (7) teacher report of total social competence (depression chronicity at 15 months); (8) te acher report of self control (depression chronicity at six and 15 months); (9 ) teacher report of peer competence (depression chronicity at six and 15 months); (10) mate rnal report of total problem behaviors (depression chronicity at six, 15, 24, and 36 months ); (11) maternal report of internalizing problem behavior s (depression chronicity at 15, 24, and 36 months); and (12) maternal report of externalizing problem behaviors (depression chronicity at 15, 24, and 36 months). Alt hough significant main effects of depression chronicity at one month were detected on mate rnal reports of total problem behaviors and

PAGE 342

320 externalizing problem behaviors, no further ex ploration of a mediati ng effect of maternal sensitivity was tested in these models gi ven that maternal sensitivity at one month postpartum was not assessed. Maternal sensit ivity was operationalized as the mother’s observed rating of maternal sensitivity at six months for significant main effects of depression chronicity detected at six months postpartum, 15 months for significant main effects of depression chroni city detected at 15 months postpartum, 24 months for significant main effects of depression chroni city detected at 24 months, and 36 months for significant main effect of depression chro nicity detected at 36 months postpartum. Prior to entering maternal sensitiv ity to each model containing the sociodemographic and depression chronicity dummy variables in which a main effect of depression chronicity was detected, two pre liminary steps were completed. First, the associations between depression chronicity at six, 15, 24, and 36 months postpartum and maternal sensitivity at six, 15, 24, and 36 months, correspondingly, was assessed. Results are summarized in Ta ble 77. New depression at six months was associated with maternal sensitivity at six months. New, re mitted, recurrent, and chronic depressions at 15 months were statistically associated with maternal sensitivity at 15 months. Only new and recurrent depressions at 24 months were significantly associated with maternal sensitivity at 24 months, and only remitted, recurrent, and chronic depression at 36 months were statistically associated with maternal sensitivity at 36 months. Based on these results, no further investigation of a me diating effect of maternal sensitivity will be tested for the following models : (1) depression chronicity at 24 months on maternal report of total social competence given that chronic depression at 24 months was not statistically associated with maternal sensitiv ity at 24 months; (2) depression chronicity at

PAGE 343

321 24 months on maternal report of cooperation given that chro nic depression at 24 months was not statistically associated with matern al sensitivity at 24 months; (3) depression chronicity at six months on maternal report of assertion given that chronic depression at six months was not statistically associated with maternal sensitivity at six months; (4) depression chronicity at six months on maternal report of responsibility given that chronic depression at six months was not statistically associated with maternal sensitivity at six months; (5) depression chronicity at 24 months on maternal report of self control given that chronic depression at 24 months was not statistically associated with maternal sensitivity at 24 months; and (6) depression ch ronicity at six months on maternal report of total problem behaviors given that ch ronic depression at six months was not statistically associated with mate rnal sensitivity at six months.

PAGE 344

322 Table 77 Main Effect of Depression Chronicity on Mate rnal Sensitivity as a Preliminary Step for Establishing Maternal Sensitivity (MS) as a Mediating Effect in th e Association Between Depression Chronicity and Child Social Competence and Display of Problem Behaviors b coefficient t value (b=0) Depression Chronicity at Six Months X MS at Six Months Never 0.00 -New -0.19 -2.26* Remitted -0.12 -1.79 Chronic -0.10 -1.09 Depression Chronicity at 15 Months X MS at 15 Months Never 0.00 -New -0.21 -2.00* Remitted -0.15 -3.05* Recurrent -0.25 -2.68* Chronic -0.26 -2.30* Depression Chronicity at 24 Months X MS at 24 Months Never 0.00 -New -0.30 -3.36* Remitted -0.10 -1.94 Recurrent -0.26 -2.72* Chronic 0.09 0.51 Depression Chronicity at 36 Months X MS at 24 Months Never 0.00 -New -0.18 -1.09 Remitted -0.21 -3.13* Recurrent -0.30 -2.54* Chronic -0.87 -3.39* *Indicates statistical significance at p<.05.

PAGE 345

323 The second preliminary step completed to provide necess ary evidence of a mediating effect of maternal sensitivity involved testi ng the associations between maternal sensitivity at each time point in wh ich a significant main effect of depression chronicity was detected (either six, 15, 24, or 36 months) and each dependent variable under investigation, while controlling for the depression chronicity dummy variables at the corresponding time point. This step wa s not completed for models in which no evidence of an association between depressi on chronicity and maternal sensitivity was detected in preliminary step number one. Results are summarized in Table 78. Significant associations were detected for: (1) maternal sensitivity at six, 15, and 36 months and maternal report of total social competence; (2) maternal sensitivity at 36 months and maternal report of cooperation; (3) maternal sensitiv ity at 36 months and maternal report of assertion; (4) matern al sensitivity at 15, 24, and 36 months and maternal report of responsibility; (5) maternal sensitivity at 15 and 36 months and maternal report of self control; (6) maternal sensitivity at 24 and 36 months and maternal report of peer competence; (7) maternal sens itivity at 15 months and teacher report of total social competence; (8) ma ternal sensitivity at 36 months and maternal report of total problem behaviors; and (9) maternal sensitiv ity at 24 and 36 months and maternal report of externalizing problem behaviors. Thus, fu rther investigation of a mediating effect of maternal sensitivity will be tested for the above models only.

PAGE 346

324 Table 78 Main Effect of Maternal Sens itivity on Select Dependent Va riables of Social Competence and Display of Problem Behaviors as a Prelim inary Step for Est ablishing Mediation b coefficient t value (b=0) Maternal Report of Total Social Competence Maternal Sensitivity at Six Months 2.35 2.39* Maternal Sensitivity at 15 Months 2.38 2.14* Maternal Sensitivity at 36 Months 3.35 4.70* Maternal Report of Cooperation Maternal Sensitivity at 36 Months 0.57 3.81* Maternal Report of Assertion Maternal Sensitivity at 15 Months 0.27 1.56 Maternal Sensitivity at 24 Months 0.23 1.39 Maternal Sensitivity at 36 Months 0.36 3.18* Maternal Report of Responsibility Maternal Sensitivity at 15 Months 0.45 2.15* Maternal Sensitivity at 24 Months 0.50 2.63* Maternal Sensitivity at 36 Months 0.42 3.15* Maternal Report of Self Control Maternal Sensitivity at 15 Months 0.72 2.85* Maternal Sensitivity at 36 Months 0.75 4.63* Maternal Report of Peer Competence Maternal Sensitivity at Six Months 0.12 0.70 Maternal Sensitivity at 15 Months 0.26 1.29 Maternal Sensitivity at 24 Months 0.48 2.58* Maternal Sensitivity at 36 Months 0.39 3.06* Teacher Report of Total Social Competence Maternal Sensitivity at 15 Months 2.22 2.17* Teacher Report of Self Control Maternal Sensitivity at Six Months 0.14 0.57 Maternal Sensitivity at 15 Months 0.47 1.71 Continued on the next page

PAGE 347

325 Table 78 (Continued) b coefficient t value (b=0) Teacher Report of Peer Competence Maternal Sensitivity at Six Months 0.11 0.44 Maternal Sensitivity at 15 Months 0.31 1.15 Maternal Report of Total Problem Behaviors Maternal Sensitivity at 15 Months -0.44 -0.62 Maternal Sensitivity at 24 Months -1.22 -1.86 Maternal Sensitivity at 36 Months -1.08 -2.33* Maternal Report of Intern alizing Problem Behaviors Maternal Sensitivity at 15 Months 0.25 0.38 Maternal Sensitivity at 24 Months 0.23 0.38 Maternal Sensitivity at 36 Months -0.05 -0.13 Maternal Report of Externalizing Problem Behaviors Maternal Sensitivity at 15 Months -0.71 -0.98 Maternal Sensitivity at 24 Months -1.49 -2.22* Maternal Sensitivity at 36 Months -1.37 -2.93* *Indicates statistical significance at p<.05. Following preliminary steps one and two, a total of 16 models for nine dependent variables remained eligible for con tinued exploration of a mediating effect of maternal sensitivity. To test these models parameter estimates (b coefficients) and associated p values for significant depressi on chronicity dummy variables in both the reduced models (sociodemographic control variables + depression chronicity dummy variables) and the full mode ls (sociodemographic cont rol variables + depression chronicity dummy variables + maternal sensit ivity) were compared. Due to the large number of models tested, onl y partial OLS regression results are presented (parameter estimates for depression chronicity dummy va riables, parameter estimates for maternal sensitivity, associated t values, and model fit statistics).

PAGE 348

326 Results testing the mediating effects of ma ternal sensitivity in the associations between depression chronicity and the following dependent variables are presented in Tables 79-93: (1) maternal report of total social competence (Tables 79-81); (2) maternal report of cooperation (Table 82); (3) maternal report of asser tion (Table 83); (4) maternal report of responsibility (Tables 84-85); (5) maternal report of self control (Tables 86-87); (6) maternal report of peer competence (Table s 88-89); (7) teacher report of total social competence (Table 90); (8) maternal report of total problem behavior s (Table 91); and (9) maternal report of externa lizing problem behaviors (Tables 92-93). A summary of findings is described after each table.

PAGE 349

327 Table 79 Partial OLS Regression Results Testing the Medi ating Effect of Mate rnal Sensitivity at Six Months in the Association Between Depression Chronicity at Six Months and Maternal Report of Total Social Competence at First Grade Maternal Report of Total Social Competence Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Depression Chronicity at Six Months Never 0.00 -0.00 -New -4.30 -2.02* -4.30 -2.02* Remitted -1.27 -0.75 -1.26 -0.75 Chronic -5.54 -2.31* -5.55 -2.31* Maternal Sensitivity at Six Months --0.08 0.08 Model Fit Statistics N 679 679 R2 0.11 0.11 Overall Model F 4.77* 4.50* *Indicates statistical significance at p<.05. Table 79 above summarizes the evidence for maternal sensitivity at six months as a mediator in the relationship between depres sion chronicity at six months and maternal report of total social competence at first gr ade. The hypothesis was not supported. After the addition of maternal sens itivity to the reduced model (see full model results) new depression remained a significant predictor of ma ternal report of total social competence, and no change in the magnitude of influence was detected. Potential change in chronic depression was not interpreted in the above table given chronic depression was not a significant predictor of maternal sensitivity at six months (preliminary step two).

PAGE 350

328 Table 80 Partial OLS Regression Results Testing the Medi ating Effect of Mater nal Sensitivity at 15 Months in the Association Be tween Depression Chronicity at 15 Months and Maternal Report of Total Social Competence at First Grade Maternal Report of Total Social Competence Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Depression Chronicity at 15 Months Never 0.00 -0.00 -New -2.34 -0.80 -2.39 -0.81 Remitted -2.47 -1.71 -2.49 -1.72 Recurrent -2.38 -0.87 -2.41 -0.88 Chronic -9.03 -2.80* -9.07 -2.81* Maternal Sensitivity at 15 Months ---0.37 -0.30 Model Fit Statistics N 679 679 R2 0.11 0.11 Overall Model F 4.60* 4.36* *Indicates statistical significance at p<.05. Table 80 above summarizes the evidence for maternal sensitivity at 15 months as a mediator in the relationship between depres sion chronicity at 15 months and maternal report of total social competence at first gr ade. The hypothesis was not supported. After the addition of maternal sensit ivity to the reduced model (s ee full model results) chronic depression remained a significant predictor of ma ternal report of total social competence, and its parameter estimate increa sed rather than decreased.

PAGE 351

329 Table 81 Partial OLS Regression Results Testing the Medi ating Effect of Mater nal Sensitivity at 36 Months in the Association Be tween Depression Chronicity at 36 Months and Maternal Report of Total Social Competence at First Grade Maternal Report of Total Social Competence Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Depression Chronicity at 36 Months Never 0.00 -0.00 -New 0.23 0.08 0.53 0.18 Remitted -1.79 -1.39 -1.65 -1.29 Recurrent -7.16 -3.27* -6.77 -3.10* Chronic -12.79 -2.72* -11.64 -2.47* Maternal Sensitivity at 36 Months --2.09 2.74* Model Fit Statistics N 679 679 R2 0.12 0.13 Overall Model F 5.05* 5.23* *Indicates statistical significance at p<.05. Table 81 above summarizes the evidence for maternal sensitivity at 36 months as a mediator in the relationship between depres sion chronicity at 36 months and maternal report of total social competence at first gr ade. This time, the hypothesis was partially supported. After the addition of maternal sens itivity to the reduced model (see full model results), although both recurrent and chronic de pression remained significant predictors of maternal report of total social competence, their associated parameter estimates decreased in magnitude.

PAGE 352

330 Table 82 Partial OLS Regression Results Testing the Medi ating Effect of Mater nal Sensitivity at 36 Months in the Association Be tween Depression Chronicity at 36 Months and Maternal Report of Cooperation at First Grade Maternal Report of Cooperation Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Depression Chronicity at 36 Months Never 0.00 -0.00 -New -0.23 -0.36 -0.17 -0.26 Remitted -0.37 -1.34 -0.34 -1.24 Recurrent -0.47 -1.00 -0.39 -0.83 Chronic -2.44 -2.40* -2.19 -2.16* Maternal Sensitivity at 36 Months --0.45 2.74* Model Fit Statistics N 679 679 R2 0.04 0.06 Overall Model F 1.71* 2.03* *Indicates statistical significance at p<.05. Table 82 above summarizes the evidence for maternal sensitivity at 36 months as a mediator in the relationship between depres sion chronicity at 36 months and maternal report of cooperation at first grade. The hypothesis was part ially supported. After the addition of maternal sensitivity to the reduced model (see full model results), although chronic depression remained a significant predic tor of maternal report of cooperation, its associated parameter estimate decreased in magnitude.

PAGE 353

331 Table 83 Partial OLS Regression Results Testing the Medi ating Effect of Mater nal Sensitivity at 36 Months in the Association Be tween Depression Chronicity at 36 Months and Maternal Report of Assertion at First Grade Maternal Report of Assertion Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Depression Chronicity at 36 Months Never 0.00 -0.00 -New -0.34 -0.70 -0.31 -0.64 Remitted -0.36 -1.75 -0.35 -1.69 Recurrent -0.86 -2.43* -0.82 -2.33* Chronic -0.84 -1.11 -0.74 -0.97 Maternal Sensitivity at 36 Months --0.19 1.55 Model Fit Statistics N 679 679 R2 0.06 0.07 Overall Model F 2.48* 2.48* *Indicates statistical significance at p<.05. Table 83 above summarizes the evidence for maternal sensitivity at 36 months as a mediator in the relationship between depres sion chronicity at 36 months and maternal report of assertion at first grade. The hypothesis was partially supported. After the addition of maternal sensitivity to the reduced model (see full model results), although recurrent depression remained a significant pred ictor of maternal re port of assertion, its associated parameter estimate decreased in magnitude.

PAGE 354

332 Table 84 Partial OLS Regression Results Testing the Medi ating Effect of Mater nal Sensitivity at 15 Months in the Association Be tween Depression Chronicity at 15 Months and Maternal Report of Responsibility at First Grade Maternal Report of Responsibility Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Depression Chronicity at 15 Months Never 0.00 -0.00 -New -0.57 -1.06 -0.58 -1.06 Remitted -0.20 -0.75 -0.20 -0.75 Recurrent -0.08 -0.15 -0.08 -0.16 Chronic -1.54 -2.59* -1.55 -2.59* Maternal Sensitivity at 15 Months ---0.03 -0.11 Model Fit Statistics N 679 679 R2 0.09 0.09 Overall Model F 3.73* 3.53* *Indicates statistical significance at p<.05. Table 84 above summarizes the evidence for maternal sensitivity at 15 months as a mediator in the relationship between depres sion chronicity at 15 months and maternal report of responsibility at first grade. The hypothesis was not supported. After the addition of maternal sensitivity to the reduced model (see full model results) chronic depression remained a significant predictor of maternal report of responsibility, and its parameter estimate increased slight ly rather than decreased.

PAGE 355

333 Table 85 Partial OLS Regression Results Testing the Medi ating Effect of Mater nal Sensitivity at 24 Months in the Association Be tween Depression Chronicity at 24 Months and Maternal Report of Responsibility at First Grade Maternal Report of Responsibility Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Depression Chronicity at 24 Months Never 0.00 -0.00 -New -0.40 -0.92 -0.35 -0.81 Remitted -0.25 -1.00 -0.24 -0.98 Recurrent -0.96 -1.99* -0.94 -1.96 Chronic -1.09 -1.24 -1.14 -1.30 Maternal Sensitivity at 24 Months --0.20 1.03 Model Fit Statistics N 679 679 R2 0.09 0.09 Overall Model F 3.63* 3.49* *Indicates statistical significance at p<.05. Table 85 above summarizes the evidence for maternal sensitivity at 24 months as a mediator in the relationship between depres sion chronicity at 24 months and maternal report of responsibility at first grade. Th e hypothesis was supported. After the addition of maternal sensitivity to th e reduced model (see full model results), recurrent depression was no longer a significant predic tor of maternal report of re sponsibility at first grade.

PAGE 356

334 Table 86 Partial OLS Regression Results Testing the Medi ating Effect of Mater nal Sensitivity at 15 Months in the Association Be tween Depression Chronicity at 15 Months and Maternal Report of Self Control at First Grade Maternal Report of Self Control Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Depression Chronicity at 15 Months Never 0.00 -0.00 -New -0.67 -1.00 -0.66 -0.98 Remitted -0.29 -0.87 -0.28 -0.86 Recurrent -0.65 -1.05 -0.65 -1.03 Chronic -2.19 -2.97* -2.18 -2.95* Maternal Sensitivity at 15 Months --0.11 0.39 Model Fit Statistics N 679 679 R2 0.10 0.10 Overall Model F 4.10* 3.89* *Indicates statistical significance at p<.05. Table 86 above summarizes the evidence for maternal sensitivity at 15 months as a mediator in the relationship between depres sion chronicity at 15 months and maternal report of self control at first grade. Th e hypothesis was partially supported. After the addition of maternal sensitivity to the reduced model (see full model results), although chronic depression remained a significant predicto r of maternal report of self control, its associated parameter estimate slightly decreased in magnitude.

PAGE 357

335 Table 87 Partial OLS Regression Results Testing the Medi ating Effect of Mater nal Sensitivity at 36 Months in the Association Be tween Depression Chronicity at 36 Months and Maternal Report of Self Control at First Grade Maternal Report of Self Control Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Depression Chronicity at 36 Months Never 0.00 -0.00 -New 0.80 1.15 0.86 1.25 Remitted -0.22 -0.77 -0.20 -0.67 Recurrent -1.58 -3.15* -1.50 -3.00* Chronic -3.44 -3.20* -3.21 -2.98* Maternal Sensitivity at 36 Months --0.43 2.47* Model Fit Statistics N 679 679 R2 0.12 0.12 Overall Model F 4.77* 4.88* *Indicates statistical significance at p<.05. Table 87 above summarizes the evidence for maternal sensitivity at 36 months as a mediator in the relationship between depres sion chronicity at 36 months and maternal report of self control at first grade. Th e hypothesis was partially supported. After the addition of maternal sensitivity to the reduced model (see full model results), although both recurrent and chronic depression remained significant predictors of maternal report of self control, their associated para meter estimates decreased in magnitude.

PAGE 358

336 Table 88 Partial OLS Regression Results Testing the Medi ating Effect of Mater nal Sensitivity at 24 Months in the Association Be tween Depression Chronicity at 24 Months and Maternal Report of Peer Competence at First Grade Maternal Report of Peer Competence Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Depression Chronicity at 24 Months Never 0.00 -0.00 -New -0.31 -0.74 -0.25 -0.59 Remitted -0.42 -1.72 -0.41 -1.69 Recurrent -1.05 -2.22* -1.03 -2.18* Chronic -2.29 -2.66* -2.35 -2.74* Maternal Sensitivity at 24 Months --0.26 1.36 Model Fit Statistics N 679 679 R2 0.07 0.08 Overall Model F 2.94* 2.89* *Indicates statistical significance at p<.05. Table 88 above summarizes the evidence for maternal sensitivity at 24 months as a mediator in the relationship between depres sion chronicity at 24 months and maternal report of peer competence at first grade. The hypothesis was partially supported. After the addition of maternal sens itivity to the reduced model (see full model results), although recurrent depression remained a signif icant predictor, its associated parameter estimate decreased slightly. Potential change in chronic depression was not interpreted given that chronic depression at 24 months was not a signif icant predictor of maternal sensitivity at 24 months (preliminary step two).

PAGE 359

337 Table 89 Partial OLS Regression Results Testing the Medi ating Effect of Mater nal Sensitivity at 36 Months in the Association Be tween Depression Chronicity at 36 Months and Maternal Report of Peer Competence at First Grade Maternal Report of Peer Competence Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Depression Chronicity at 36 Months Never 0.00 -0.00 -New -0.16 -0.30 -0.13 -0.24 Remitted -0.39 -1.66 -0.37 -1.60 Recurrent -0.90 -2.26* -0.87 -2.17* Chronic -2.27 -2.64* -2.15 -2.50* Maternal Sensitivity at 36 Months --0.21 1.46 Model Fit Statistics N 679 679 R2 0.07 0.07 Overall Model F 2.94* 2.89* *Indicates statistical significance at p<.05. Table 89 above summarizes the evidence for maternal sensitivity at 36 months as a mediator in the relationship between depres sion chronicity at 36 months and maternal report of peer competence at first grade. The hypothesis was partially supported. After the addition of maternal sens itivity to the reduced model (see full model results), although both recurrent and chronic depres sion remained significant predictors of maternal report of peer competence, their associated parameter estimates decreased in magnitude.

PAGE 360

338 Table 90 Partial OLS Regression Results Testing the Medi ating Effect of Mater nal Sensitivity at 15 Months in the Association Between Depression Chronicity at 15 Months and Teacher Report of Total Social Competence at First Grade Teacher Report of Total Social Competence Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Depression Chronicity at 15 Months Never 0.00 -0.00 -New -0.18 -0.06 -0.18 -0.06 Remitted -1.18 -0.88 -1.18 -0.88 Recurrent -5.04 -1.99* -5.04 -1.99* Chronic 0.90 0.30 0.90 0.30 Maternal Sensitivity at 15 Months ---0.00 -0.00 Model Fit Statistics N 679 679 R2 0.07 0.07 Overall Model F 2.94* 2.78* *Indicates statistical significance at p<.05. Table 90 above summarizes the evidence for maternal sensitivity at 15 months as a mediator in the relationship between depr ession chronicity at 15 months and teacher report of total social competence at first gr ade. The hypothesis was not supported. After the addition of maternal sensit ivity to the reduced model (see full model results) recurrent depression remained a significant predictor of teacher report of total social competence, and no change in the magnitude of influence was detected.

PAGE 361

339 Table 91 Partial OLS Regression Results Testing the Medi ating Effect of Mater nal Sensitivity at 36 Months in the Association Be tween Depression Chronicity at 36 Months and Maternal Report of Total Problem Behaviors at First Grade Maternal Report of Total Problem Behaviors Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Depression Chronicity at 36 Months Never 0.00 -0.00 -New -0.69 -0.35 -0.79 -0.40 Remitted 1.55 1.85 1.51 1.79 Recurrent 4.86 3.38* 4.73 3.29* Chronic 9.20 2.98* 8.81 2.84* Maternal Sensitivity at 36 Months ---0.72 -1.42 Model Fit Statistics N 679 679 R2 0.07 0.07 Overall Model F 2.64* 2.61* *Indicates statistical significance at p<.05. Table 91 above summarizes the evidence for maternal sensitivity at 36 months as a mediator in the relationship between depres sion chronicity at 36 months and maternal report of total problem behaviors at first gr ade. The hypothesis was partially supported. After the addition of maternal sensitivity to the reduced model (see full model results), although both recurrent and chronic depres sion remained significant predictors of maternal report of total problem behavior s, their associated parameter estimates decreased in magnitude.

PAGE 362

340 Table 92 Partial OLS Regression Results Testing the Medi ating Effect of Mater nal Sensitivity at 24 Months in the Association Be tween Depression Chronicity at 24 Months and Maternal Report of Externalizing Problem Behaviors at First Grade Maternal Report of Externalizing Problems Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Depression Chronicity at 24 Months Never 0.00 -0.00 -New 1.87 1.20 1.61 1.03 Remitted 1.57 1.75 1.53 1.72 Recurrent 3.51 2.03* 3.43 1.99* Chronic 10.26 3.26* 10.56 3.36* Maternal Sensitivity at 24 Months ---1.15 -1.61 Model Fit Statistics N 679 679 R2 0.07 0.07 Overall Model F 2.66* 2.66* *Indicates statistical significance at p<.05. Table 92 above summarizes the evidence for maternal sensitivity at 24 months as a mediator in the relationship between depres sion chronicity at 24 months and maternal report of externalizing problems at first grad e. The hypothesis was partially supported. After the addition of maternal sensitivity to the reduced model (see full model results), although recurrent depression remained a signif icant predictor, its associated parameter estimate decreased in magnitude. Potentia l change in chronic depression was not interpreted given that chronic depression at 24 months was not a significant predictor of maternal sensitivity at 24 months (preliminary step two).

PAGE 363

341 Table 93 Partial OLS Regression Results Testing the Medi ating Effect of Mater nal Sensitivity at 36 Months in the Association Be tween Depression Chronicity at 36 Months and Maternal Report of Externalizing Problem Behaviors at First Grade Maternal Report of Externalizing Problems Reduced Model Full Model Variables b coefficient t value (b=0) b coefficient t value (b=0) Depression Chronicity at 36 Months Never 0.00 -0.00 -New -0.11 -0.06 -0.25 -0.12 Remitted 1.51 1.77 1.45 1.70 Recurrent 3.28 2.24* 3.11 2.12* Chronic 10.18 3.24* 9.68 3.07* Maternal Sensitivity at 36 Months ---0.91 -1.78 Model Fit Statistics N 679 679 R2 0.07 0.07 Overall Model F 2.67* 2.70* *Indicates statistical significance at p<.05. Table 93 above summarizes the evidence for maternal sensitivity at 36 months as a mediator in the relationship between depres sion chronicity at 36 months and maternal report of externalizing problems at first grad e. The hypothesis was partially supported. After the addition of maternal sensitivity to the reduced model (see full model results), although both recurrent and chronic depres sion remained significant predictors of maternal report of externaliz ing problem behaviors, their a ssociated parameter estimates decreased in magnitude.

PAGE 364

342 Hypothesis 3.4. Hypothesis 3 stated that the effect of maternal depression (depression chronicity) on child social competence and disp lay of problem behaviors at first grade is stronger among mothers who rate low in maternal sensitivity. To te st the moderating effect of maternal sensitivity on the associ ations between depre ssion chronicity during the first 36 months postpartum and child so cial competence and display of problem behaviors at first grade, th e interaction terms of each depression chronicity dummy variable at a given time point (six, 15, 24 or 36 months) times maternal sensitivity at the corresponding time point (six, 15, 24 or 36 months) were added as a block to each final model (sociodemographic cont rol variables + depression ch ronicity dummy variables + maternal sensitivity). Maternal sensitivity was not assessed at one month postpartum thus no interaction term of depression chroni city at one month and maternal sensitivity was tested. A significant intera ction term indicates the presence of a moderating effect. A summary of significant findings from OLS regression models testing the moderating effect of maternal sensitiv ity on the association between depression chronicity on child social competence and di splay of problem behaviors is presented in Table 94. Model 2 highlights the significance of depression chronic ity dummy variables at six months postpartum times maternal sens itivity at six months. Model 3 highlights the significance of depression chronicity dum my variables at 15 months postpartum times maternal sensitivity at 15 months; and Mode ls 4 and 5 highlight the significance of depression chronicity dummy variables at 24 and 36 months postpartum times maternal sensitivity at 24 and 36 months respectively.

PAGE 365

343 An ‘X’ in a designated cell i ndicates that at least one in teraction term in a given model was significant at p<.05, controlling for al l the other variables in the model. An “XX” in a designated cell indicates that at least one interaction term in a given model was significant at p<.01 (Bonferroni adjustment to account for the five models tested for each dependent variable, 0.05/5=0.01), controlling for all the other variables in the model. For example, three interaction terms (one for each depression chronicity dummy variable including new, remitted, and chronic) were entered to assess the moderating effect of maternal sensitivity at six months, and four interaction terms (one for each depression chronicity dummy variable including new, remitted, recurrent, and chronic) were entered to assess the moderating effect of maternal sensitivity at 15, 24, and 36 months. Only a single interaction term needed to be signifi cant for an “X” or “XX” to appear in the appropriate cell in Table 94. Significant moderating effects of maternal sensitivity were detected for: (1) maternal report of cooperation (depression chronicity at six months X maternal sensitivity at six months); (2) teacher re port of total social competence (depression chronicity at 15 months X maternal sensit ivity at 15 months; (3) teacher report of assertion (depression chronicity at 15 months X maternal sensitivity at 15 months); (4) teacher report of self control (depression ch ronicity at 15 months X maternal sensitivity at 15 months); and (5) teacher report of peer competence (depression chronicity at 15 months X maternal sensi tivity at 15 months).

PAGE 366

344 Table 94 Summary of Findings from OLS Regression Te sting the Moderating Effect of Maternal Sensitivity on the Association Between Depression Chronicity on Child Social Co mpetence and Display of Problem Behaviors Depression Chronicity X Maternal Sensitivity Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Maternal Report of Social Competence Total na ns ns ns ns Cooperation na XX ns ns ns Assertion na ns ns ns ns Self Control na ns ns ns ns Responsibility na ns ns ns ns Peer Competence na ns ns ns ns Teacher Report of Social Competence Total na ns X ns ns Cooperation na ns ns ns ns Assertion na ns XX ns ns Self Control na ns XX ns ns Peer Competence na ns X ns ns Maternal Report of Problem Behaviors Total Problem Behaviors na ns ns ns ns Internalizing Problems na ns ns ns ns Externalizing Problems na ns ns ns ns

PAGE 367

345 Table 95 below provides mo re detail related to th e significant moderating effects of maternal sensitivit y detected for maternal repo rt of cooperation and teacher reports of total social compet ence, assertion, self control, and peer competence. Nonsignificant results from testing the moderati ng effect of maternal sensitivity on the relationship between depression chronicity and child social competence and display of problem behaviors are not presented. Un standardized regression coefficients and associated t values are provide d for dummy variable interac tion terms included in a single model where at least one significant interacti on term was detected. Results are organized by time point (six, and 15 months) given that the depression chroni city dummy variables vary across time point at which chronici ty was assessed (new, remitted, and chronic depression at six months; new, remitted, recurre nt, and chronic depression at 15 months). Note that two separate criteria for judging statistical significance are presented in Table 95, p<.05 as the typical criteria utilized, and a Bonferroni-adjusted p-value of <.01 given that five models were tested for each dependent variable (0.05/5=0.01).

PAGE 368

346 Table 95 Significant Interaction Terms of Depressi on Chronicity and Maternal Sensitivity (MS) b coefficient t value (b=0) Six Months Maternal Report of Cooperation New Depression X MS at Six Months -0.10 -0.13 Remitted Depression X MS at Six Months 1.82 3.09** Chronic Depression X MS at Six Months 1.30 0.89 15 Months Teacher Report of Total Social Competence New Depression X MS at 15 Months -4.35 -0.94 Remitted Depression X MS at 15 Months 6.28 2.42* Recurrent Depression X MS at 15 Months 5.43 1.22 Chronic Depression X MS at 15 Months 1.78 0.28 Teacher Report of Assertion New Depression X MS at 15 Months -3.52 -2.69** Remitted Depression X MS at 15 Months 1.74 2.37* Recurrent Depression X MS at 15 Months 1.91 1.52 Chronic Depression X MS at 15 Months 1.54 0.86 Teacher Report of Self Control New Depression X MS at 15 Months -0.10 -0.08 Remitted Depression X MS at 15 Months 1.55 2.24** Recurrent Depression X MS at 15 Months 1.04 0.88 Chronic Depression X MS at 15 Months -0.24 -0.14 Teacher Report of Peer Competence New Depression X MS at 15 Months -1.58 -1.31 Remitted Depression X MS at 15 Months 1.71 2.52* Recurrent Depression X MS at 15 Months 1.24 1.07 Chronic Depression X MS at 15 Months 0.92 0.56 *Indicates statistical significance at p<.05. **Indicates statistical significance at p<.01.

PAGE 369

347 Overall, for most dependent variables in which a moderating effect of maternal sensitivity was detected, hypothesis 3 was s upported. In fact, of the six significant interaction terms detected, five supported th e hypothesis that the effect of depression chronicity is stronger among mothers who rate low in mate rnal sensitivity. For example, maternal sensitivity at six months moderated the relationship between remitted depression at six months postpartum and maternal report of cooperation. Remitted depression at six months (or depression pres ent at one month postp artum and not present at six months postpartum) had the strongest negative impact on cooperation at first grade among children with mothers rated lower in maternal sensitivit y (interaction term estimates: b=1.82, t=3.09, p<.01). In addition, at fifteen months postpartum, the relationships between depre ssion chronicity and teacher reports of total social competence, assertion, self-control, and peer competence, were moderated by maternal sensitivity at 15 months. Mo re specifically, remitted depression at 15 months had a greater negative impact on (1) te acher report of total social competence (interaction term estimates: b=6.28, t=2.42, p<.05); (2) teacher re port of assertion (interaction term estimates: b=1.74, t=2.37, p<.05); (3) teacher repo rt of self control (interaction term estimates: b=1.55, t=2.24, p<.01); and (4) teacher report of peer competence (interaction term estimates: b=1.71, t=2.52, p<.05); among ch ildren with mothers rated lower in maternal sensitivity. Although the moderating eff ect of maternal sensitivity at 15 months on the relationship between depression chronici ty (new depression) on teacher report of assertion was statisticall y significant (interaction te rm estimates: b=-3.52, t=-2.69, p<.01), the effect of new de pression at 15 months was stronger among children with

PAGE 370

348 mothers rated higher in maternal sensitivit y. This latter finding does not support hypothesis 3. Hypothesis 4.4. Hypothesis 4 stated that the effect of maternal depression (depression chronicity) on child social competence and disp lay of problem behaviors at first grade is weaker among children exposed to nonmaternal care. To test the moderating effect of ever exposure to nonmaternal care on the asso ciations between depr ession chronicity during the first 36 months postpartum and child social competence and display of problem behaviors at first grade, the inte raction terms of each depression chronicity dummy variable at a given time point (six, 15, 24 or 36 months) times ever exposure to nonmaternal care at the corresponding time poi nt (six, 15, 24 or 36 months) were added as a block to each final model (sociode mographic control variables + depression chronicity dummy variables + maternal sens itivity + exposure to nonmaternal care). The moderating effect of ever e xposure to nonmaternal care at one month was not explored given that maternal sensitivity, a control variable, was not asse ssed in the NICHD SECC at one month postpartum. Exposure to nonmaternal care was operationalized in two ways: (1) ever exposure to nonmaternal (despi te type of arrangement); and (2) ever exposure to center-based nonmate rnal care. Ever exposure to nonmaternal care consisted of two response options (no and yes), and ever exposure to center-based nonmaternal care consisted of three response options (mat ernal care only, non-cente r-based nonmaternal care, and center-based nonmaternal care). A significant interaction term indicates the presence of a moderating effect.

PAGE 371

349 A summary of significant findings from OLS regression models testing the moderating effect of ever exposure to non maternal care on the association between depression chronicity on child social compet ence and display of problem behaviors is presented in Table 96. Model 2 highlights the significance of depression chronicity dummy variables at six months postpartum tim es ever exposure to nonmaternal care at six months. Model 3 highlights the signi ficance of depression chronicity dummy variables at 15 months postpartum times ever exposure to nonmaternal care at 15 months; and Models 4 and 5 highlight the significance of depression chronicity dummy variables at 24 and 36 months postpartum times ever exposure to nonmaternal care at 24 and 36 months respectively. An ‘X’ in a designated cell indicates th at at least one interaction term in a given model was significant at p< .05, controlling for all th e other variables in the model. An “XX” in a designated cell indicates that at least one interaction term in a given model was significant at p<.005 (Bonfe rroni adjustment to account for the ten models tested for each dependent variable 0.05/10=0.01), controlling for all the other variables in the model. Only a single interac tion term needed to be significant for an “X” or “XX” to appear in the appropriate cell in Table 96. Significant moderating effects of ever expos ure to nonmaternal were detected for: (1) maternal report of assert ion (depression chronicity six, 24, and 36 months X ever exposure to nonmaternal care at six, 24, and 36 months, respec tively); (2) teacher report of total social competence (depression chronici ty at six and 24 months X ever exposure to nonmaternal care at six and 24 months, respec tively); (6) teacher re port of cooperation (depression chronicity at 24 and 36 months X ever exposure to nonmaternal care at 24 and 36 months, respectively); (7) teacher report of assertion (d epression chronicity at six

PAGE 372

350 and 24 months X ever exposure to nonmaterna l care at six and 24 months, respectively; (8) teacher report of self cont rol (depression chroni city at 24 months X ever exposure to nonmaternal care at 24 months; (9) teacher report of peer competence (depression chronicity at six, 24, and 36 months X ever to nonmaternal care at six, 24, and 36 months, respectively); and (10) maternal report of externalizing problem behaviors (depression chronicity at 15 months X ever exposure to nonmaternal care at 15 months).

PAGE 373

351 Table 96 Summary of Findings from OLS Regression Test ing the Moderating Effect of Ever Exposure to Nonmaternal Care on the Association Between Depression Chronicity on Child Social Competence and Display of Problem Behaviors Depression Chronicity X Ever Exposure to Nonmaternal Care Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Maternal Report of Social Competence Total na ns ns ns ns Cooperation na ns ns ns ns Assertion na X ns XX X Self Control na ns ns ns ns Responsibility na ns ns ns ns Peer Competence na ns ns ns ns Teacher Report of Social Competence Total na X ns XX ns Cooperation na ns ns XX X Assertion na X ns X ns Self Control na ns ns X ns Peer Competence na X ns XX X Maternal Report of Problem Behaviors Total Problem Behaviors na ns ns ns ns Internalizing Problems na ns ns ns ns Externalizing Problems na ns X ns ns

PAGE 374

352 Table 97 below provides mo re detail related to th e significant moderating effects of ever exposure to nonmaternal care de tected for maternal re ports of total social competence, assertion, responsibility and peer competence, teacher repor ts of total social competence, cooperation, assertion, self c ontrol and peer competence, and maternal report of externalizing problem behaviors. Non-significant results from testing the moderating effect of ever exposure to nonmaternal on the relationship between depression chronicity and child social comp etence and display of problem behaviors are not presented. Unstandardized regression coefficients a nd associated t values are provided for dummy variable interaction term s included in a single model where at least one significant interaction term was detecte d. Results are organized by time point (six, 15, 24 and 36 months) given that the depressi on chronicity dummy variables vary across time point at which chronicity was assessed (new, remitted, and chronic depression at six month; new, remitted, recurrent, and chronic depression at 15, 24, and 36 months). Note that two separate criteria fo r judging statistical significan ce are presented in Table 97, p<.05 as the typical criteria utilized, and a Bonferroni-adj usted p-value of <.005 given that ten models were tested for each dependent variable (0.05/10=0.005).

PAGE 375

353 Table 97 Significant Interaction Terms of Depr ession Chronicity and Ever Exposure to Nonmaternal Care on the Asso ciation Between Depression Ch ronicity and Child Social Competence and Display of Problem Behaviors b coefficient t value (b=0) Six Months Maternal Report of Assertion New Depression X Ever E xposure at Six Months 0.02 0.03 Remitted Depression X Ever E xposure at Six Months -0.76 -1.36 Chronic Depression X Ever Exposure at Six Months 1.82 2.40* Teacher Report of Total Social Competence New Depression X Ever E xposure at Six Months 8.56 2.07* Remitted Depression X Ever E xposure at Six Months -0.71 -0.22 Chronic Depression X Ever Exposure at Six Months 2.96 0.67 Teacher Report of Assertion New Depression X Ever E xposure at Six Months 2.89 2.44* Remitted Depression X Ever Exposure at Six Months 0.18 0.20 Chronic Depression X Ever Exposure at Six Months 0.97 0.76 Teacher Report of Peer Competence New Depression X Ever E xposure at Six Months 2.56 2.38* Remitted Depression X Ever E xposure at Six Months -0.49 -0.57 Chronic Depression X Ever Exposure at Six Months 1.08 0.93 15 Months Maternal Report of Externalizing Problem Behaviors New Depression X Ever E xposure at 15 Months 9.68 2.01* Remitted Depression X Ever Exposure at 15 Months 3.32 1.36 Recurrent Depression X Ever Exposure at 15 Months 2.11 0.54 Chronic Depression X Ever Exposure at 15 Months 5.15 1.19 Continued on the next page

PAGE 376

354 Table 97 (Continued) b coefficient t value (b=0) 24 Months Maternal Report of Assertion New Depression X Ever E xposure at 24 Months 0.81 0.73 Remitted Depression X Ever E xposure at 24 Months -0.54 -0.98 Recurrent Depression X Ever Exposure at 24 Months 0.18 0.14 Chronic Depression X Ever Exposure at 24 Months 5.27 2.91** Teacher Report of Total Social Competence New Depression X Ever E xposure at 24 Months 5.14 0.79 Remitted Depression X Ever E xposure at 24 Months -1.53 -0.48 Recurrent Depression X Ever Exposure at 24 Months 8.58 1.15 Chronic Depression X Ever Exposure at 24 Months 31.73 3.02** Teacher Report of Cooperation New Depression X Ever E xposure at 24 Months 0.82 0.44 Remitted Depression X Ever Exposure at 24 Months 0.35 0.38 Recurrent Depression X Ever Exposure at 24 Months 1.38 0.65 Chronic Depression X Ever Exposure at 24 Months 9.18 3.04** Teacher Report of Assertion New Depression X Ever E xposure at 24 Months 2.64 1.43 Remitted Depression X Ever E xposure at 24 Months -0.29 -0.32 Recurrent Depression X Ever Exposure at 24 Months 2.45 1.16 Chronic Depression X Ever Exposure at 24 Months 7.12 2.37* Teacher Report of Self Control New Depression X Ever E xposure at 24 Months 1.19 0.68 Remitted Depression X Ever E xposure at 24 Months -0.69 -0.80 Recurrent Depression X Ever Exposure at 24 Months 1.43 0.72 Chronic Depression X Ever Exposure at 24 Months 7.30 2.58* Teacher Report of Peer Competence New Depression X Ever E xposure at 24 Months 2.53 1.49 Remitted Depression X Ever E xposure at 24 Months -0.53 -0.63 Recurrent Depression X Ever Exposure at 24 Months 1.76 0.91 Chronic Depression X Ever Exposure at 24 Months 8.48 3.08** Continued on the next page

PAGE 377

355 Table 97 (Continued) b coefficient t value (b=0) 36 Months Maternal Report of Assertion New Depression X Ever Expos ure at 36 Months -1.08 -0.74 Remitted Depression X Ever Exposure at 36 Months 0.41 0.58 Recurrent Depression X Ever Exposure at 36 Months -2.55 -2.36* Chronic Depression X Ever Exposure at 36 Months 4.85 2.01* Teacher Report of Cooperation New Depression X Ever Expos ure at 36 Months -4.99 -2.05* Remitted Depression X Ever Exposure at 36 Months 0.93 0.79 Recurrent Depression X Ever Exposure at 36 Months -2.05 -1.13 Chronic Depression X Ever Exposure at 36 Months 6.27 1.55 Teacher Report of Peer Competence New Depression X Ever E xposure at 36 Months 1.07 0.48 Remitted Depression X Ever Exposure at 36 Months 0.82 0.76 Recurrent Depression X Ever Exposure at 36 Months -0.28 -0.17 Chronic Depression X Ever Exposure at 36 Months 8.01 2.17* *Indicates statistical significance at p<.05. **Indicates statistical significance at p<.005. Significant moderating effects of ever e xposure to nonmaternal care detected at six and 24 months supported hypothesis 3, w ith depression chronicity (either new depression or chronic depression) having a weaker impact on child social competence and display of problem behaviors among child ren ever exposed to nonmaternal care at the corresponding time points. For example, either new or chronic depression at six months had a weaker influence on maternal report of assertio n (interaction term estimates: b=1.82, t=2.40, p<.05), teacher report of total social competence (interaction term estimates: b=8.56, t=2.07, p<.05), teacher report of assertion (interaction term estimates: b=2.89, t=2.44, p<.05), and teacher re port of peer competence (interaction

PAGE 378

356 term estimates: b=2.56, t=2.38, p<.05), among childre n ever exposed to nonmaternal care at six months. A buffering effect of ever exposure to nonmaternal care was also detected for the associations between chronic de pression at 24 months and the following dependent variables: (1) maternal report of assertion (interaction term estimates: b=5.27, t=2.91, p<.005); (2) teache r report of total social compet ence (interaction term estimates: b=31.73, t=3.02, p<.005); (3) teacher report of c ooperation (interaction term estimates: b=9.18, t=3.04, p<.005); (4) teacher report of assertion (interaction term estimates: b=7.12, t=2.37, p<.05); (5) teacher report of se lf control (interaction term estimates: b=7.30, t=2.58, p<.05); and (6) t eacher report of peer co mpetence (interaction term estimates: b=8.48, t=3.08, p<.005). Furthermore, ever exposure to nonmater nal care at 36 months moderated the association between chronic depression at 36 months and maternal report of assertion (interaction term estimates: b=4.85, t=2.01, p<.05) and teacher report of peer competence (interaction term estimates: b=8.01, t=2.17, p<.05). The impact of chronic depression at 36 months on these two aspects of child so cial competence was weaker among children ever exposed to nonmaternal care at 36 months. The opposite trend however was detected for the moderating effect of ev er exposure to nonmaternal care on the relationship between: (1) recurrent depressi on at 36 months and maternal report of assertion (interaction term estimates: b=2.55, t=-2.36, p<.05); and (2) new depression at 36 months and teacher report of cooperati on (interaction term estimates: b=-4.99, t=-2.05, p<.05). For both of these moderating effects, ch ronicity of depression (either recurrent or new) had stronger negative impacts on outcome among children ever exposed to nonmaternal care at 36 months. Children exposed to maternal care only faired better in

PAGE 379

357 the context of recurrent or ne w depression at 36 months in re lation to maternal report of assertion and teacher report of cooperation, respectively. On the other hand however, ever exposure to nonmaternal care at 15 months exacerbated the negative impact of new depre ssion at 15 months on maternal report of externalizing problem behavi ors (interaction term estimat es: b=9.69, t=2.01, p<.05). The impact of new depression at 15 months wa s stronger among children ever exposed to nonmaternal care. A summary of significant findings from OLS regression models testing the moderating effect of ever exposure to cente r-based nonmaternal care on the association between depression chronicity on child so cial competence and display of problem behaviors is presented in Ta ble 98. Model 2 highlights th e significance of depression chronicity dummy variables at six months postpartum times ever exposure to centerbased nonmaternal care at six months. Mode l 3 highlights the significance of depression chronicity dummy variables at 15 months pos tpartum times ever exposure to center-based nonmaternal care at 15 months; and Models 4 and 5 highlight the significance of depression chronicity dummy variables at 24 and 36 months postpartum times ever exposure to center-based nonmate rnal care at 24 and 36 months respectively. An ‘X’ in a designated cell indicates that at least one interaction term in a given model was significant at p<.05, controlling for all the othe r variables in the model. No interaction terms were detected to be significant usi ng a Bonferroni adjusted p value of <.005 (to account for the ten models tested for each dependent variable, 0.05/10=0.01). Only a single interaction term needed to be significan t for an “X” to appear in the appropriate cell in Table 98.

PAGE 380

358 Significant moderating effects of ever e xposure to center-based nonmaternal were detected for: (1) maternal report of coopera tion (depression chroni city at 36 months X ever exposure to center-based nonmaternal ca re at 36 months; (2) maternal report of assertion (depression chronicity at six and 15 months X ever exposure to center-based nonmaternal care at six and 15 months, re spectively); (3) matern al report of peer competence (depression chronicity at 15 m onths X ever exposure to center-based nonmaternal care at 15 months); (4) teacher re port of total social competence (depression chronicity at 24 months X ever exposure to center-based nonmaternal care at 24 months); (5) teacher report of cooperation (depression chronicity at 15 and 24 months X ever exposure to center-based nonmate rnal care at 15 and 24 months respectively); (6) teacher report of assertion (depression chronicity at six months X ever exposure to center-based nonmaternal care at six months; (7) teacher report of self control (depression chronicity at 24 months X ever exposure to center-based nonmaternal care at 24 months; (8) teacher report of peer competence (depression chronic ity at 24 and 36 months X ever exposure to center-based nonmaternal care at 24 and 36 months, respec tively); and (9) maternal report of externalizing problem behaviors (d epression chronicity at 24 months X ever exposure to center-based nonmate rnal care at 24 months).

PAGE 381

359 Table 98 Summary of Findings from OLS Regression Test ing the Moderating Effect of Ever Exposu re to Center-Based Nonmaternal Care on the Association Between Depression Ch ronicity on Child Social Competence and Display of Problem Behaviors Depression Chronicity X Ever Exposure to Center-Based Nonmaternal Care Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Maternal Report of Social Competence Total na ns ns ns ns Cooperation na ns ns ns X Assertion na X ns X ns Self Control na ns ns ns ns Responsibility na ns ns ns ns Peer Competence na ns X ns ns Teacher Report of Social Competence Total na ns ns X ns Cooperation na ns X X ns Assertion na X ns ns ns Self Control na ns ns X ns Peer Competence na ns ns X X Maternal Report of Problem Behaviors Total Problem Behaviors na ns ns ns ns Internalizing Problems na ns ns ns ns Externalizing Problems na ns ns X ns

PAGE 382

360 Table 99 below provides more detail related to the signi ficant moderating effects of ever exposure to center-based nonmaternal care detected for maternal reports of total social competence, cooperati on, assertion, responsibility a nd peer competence, teacher reports of total social competence, coope ration, assertion, self control and peer competence, and maternal report of external izing problem behaviors. Non-significant results from testing the moderating effect of ever exposure to center-based nonmaternal on the relationship between depression chr onicity and child social competence and display of problem behaviors are not presente d. Unstandardized regression coefficients and associated t values are provided for dumm y variable interaction terms included in a single model where at least one significant in teraction term was detected. Results are organized by time point (one, six, 15, 24 and 36 months) given that the depression chronicity dummy variables vary across time point at which chronicity was assessed. Recall that no interaction terms were detect ed to be significant using a Bonferroni adjusted p value of <.005.

PAGE 383

361 Table 99 Significant Interaction Terms of Depressi on Chronicity and Ever Exposure to CenterBased (CB) Nonmaternal Care on the Association Between Depression Chronicity and Child Social Competence and Display of Problem Behaviors b coefficient t value (b=0) Six Months Maternal Report of Assertion New Depression X Exposure to CB at Six Months -0.23 -0.39 Remitted Depression X Exposure to CB at Six Months -0.84 -1.96* Chronic Depression X Exposur e to CB at Six Months 1.29 2.26* Teacher Report of Assertion New Depression X Exposure to CB at Six Months 2.11 2.18* Remitted Depression X Exposure to CB at Six Months -0.07 -0.10 Chronic Depression X Exposur e to CB at Six Months 0.28 0.29 15 Months Maternal Report of Peer Competence New Depression X Exposure to CB at 15 Months -2.03 -2.01* Remitted Depression X Exposure to CB at 15 Months -0.40 -0.86 Recurrent Depression X Expos ure to CB at 15 Months 0.02 0.02 Chronic Depression X Exposure at to CB 15 Months -0.20 -0.26 Teacher Report of Cooperation New Depression X Exposure to CB at 15 Months 0.14 0.09 Remitted Depression X Exposur e to CB at 15 Months 0.88 1.29 Recurrent Depression X Exposur e to CB at 15 Months -0.40 -0.34 Chronic Depression X Exposur e at to CB 15 Months 2.22 1.96* Continued on the next page

PAGE 384

362 Table 99 (Continued) b coefficient t value (b=0) 24 Months Maternal Report of Assertion New Depression X Exposure to CB at 24 Months -0.09 -0.14 Remitted Depression X Exposure to CB at 24 Months -0.75 -2.20* Recurrent Depression X Expos ure to CB at 24 Months 0.82 1.11 Chronic Depression X Exposur e to CB at 24 Months 1.72 1.79 Teacher Report of Total Social Competence New Depression X Exposure to CB at 24 Months 3.55 1.03 Remitted Depression X Exposure to CB at 24 Months -0.56 -0.28 Recurrent Depression X Expos ure to CB at 24 Months 2.96 0.68 Chronic Depression X Exposur e to CB at 24 Months 13.38 2.38* Teacher Report of Cooperation New Depression X Exposure to CB at 24 Months 0.87 0.88 Remitted Depression X Exposur e to CB at 24 Months 0.49 0.86 Recurrent Depression X Expos ure to CB at 24 Months 0.82 0.66 Chronic Depression X Exposur e to CB at 24 Months 4.33 2.69* Teacher Report of Self Control New Depression X Exposure to CB at 24 Months 0.42 0.46 Remitted Depression X Exposure to CB at 24 Months -0.24 -0.44 Recurrent Depression X Expos ure to CB at 24 Months 0.28 0.24 Chronic Depression X Exposur e to CB at 24 Months 3.71 2.47* Teacher Report of Peer Competence New Depression X Exposure to CB at 24 Months 1.21 1.35 Remitted Depression X Exposure to CB at 24 Months -0.35 -0.66 Recurrent Depression X Expos ure to CB at 24 Months 0.45 0.40 Chronic Depression X Exposur e to CB at 24 Months 4.06 2.77* Maternal Report of Externalizing Problem Behaviors New Depression X Exposure to CB at 24 Months -0.40 -0.16 Remitted Depression X Exposur e to CB at 24 Months 1.30 0.91 Recurrent Depression X Expos ure to CB at 24 Months 6.23 2.01* Chronic Depression X Exposure to CB at 24 Months -2.20 -0.55 Continued on the next page

PAGE 385

363 Table 99 (Continued) b coefficient t value (b=0) 36 Months Maternal Report of Cooperation New Depression X Ever E xposure at 36 Months 2.00 1.98* Remitted Depression X Ever E xposure at 36 Months -0.18 -0.42 Recurrent Depression X Ever Exposure at 36 Months 0.21 0.29 Chronic Depression X Ever Exposure at 36 Months 1.50 1.01 Teacher Report of Peer Competence New Depression X Ever E xposure at 36 Months 0.63 0.55 Remitted Depression X Ever Exposure at 36 Months 0.15 0.30 Recurrent Depression X Ever Exposure at 36 Months 0.38 0.46 Chronic Depression X Ever Exposure at 36 Months 3.45 2.04* *Indicates statistical significance at p<.05. Significant moderating effects of ever e xposure to center-based nonmaternal care on the associations between new and chronic depression on social competence detected at six, 15, 24 and 36 months all supported hypothesi s 3, with depression chronicity (either new depression or chronic depression) ha ving a weaker impact on child social competence among children ever exposed to center-based nonmater nal care at the corresponding time point. For example, am ong children ever expos ed to center-based nonmaternal care at six months, chronic depres sion at six months had a weaker influence on maternal report of assertion (interacti on term estimates: b=1.29, t=2.26, p<.05), as did new depression on teacher repo rt of assertion (interacti on term estimates: b=2.11, t=2.18, p<.05). The same trend was detected for chr onic depression at 15 months and its impact on teacher report of cooperation, with a weak er impact among children ever exposed to center-based nonmaternal care at 15 months (interaction term estimates: b=2.22, t=1.96, p<.05).

PAGE 386

364 A buffering effect of ever exposure to center-based nonmaternal care was also detected for the associations between chroni c depression at 24 months and the following dependent variables: (1) maternal report of total social competence (interaction term estimates: b=13.38, t=2.38, p<.05); (2) teacher report of cooperation (interaction term estimates: b=4.33, t=2.69, p<.-05); (3) teacher re port of self contro l (interaction term estimates: b=3.71, t=2.47, p<.05); and (4) teacher report of peer competence (interaction term estimates: b=4.06, t=2.77, p<.05). Furthermor e, the impact of both new and chronic depression at 36 months on maternal report of cooperation (interaction term estimates: b2.00, t=1.98, p<.05) and teacher report of peer competence (interaction term estimates: b=3.45, t=2.04, p<.05) respectively, were mode rated by ever exposure to center-based nonmaternal care at 36 months, with the impact being weaker among children ever exposed to center-based nonmaternal care. Negative influences of ever exposure to center-based nonmaternal care, in opposition to hypothesis four, were detected for the following significant moderating effects: (1) the impact of remitted depre ssion at six months on maternal report of assertion was stronger among children ever e xposed to center-based nonmaternal care at six months (interaction term estimates: b= -0.84, t=-1.96, p<.05); (2 ) the impact of new depression at 15 months on maternal repor t of peer competence was stronger among children ever exposed to cente r-based nonmaternal care at 15 months (interaction term estimates: b=2.22, t=1.96, p<.05); (3) the impact of remitted depression at 24 months on maternal report of assertion was stronger among children ever exposed to center-based nonmaternal care at 24 months (interaction term estimates : b=-0.75, t=-2.20, p<.05); and (4) the impact of recurrent depression at 24 months on maternal report of externalizing

PAGE 387

365 problem behaviors was stronger among ch ildren ever exposed to center-based nonmaternal care (b=6.23, t=2.01, p<.05). Support for Hypotheses The following table outlines each of the study hypotheses and indicates whether this dissertation f ound no support, some support, moderate support, or strong support for each hypothesis.

PAGE 388

366 Table 100 Summary of Study Findings Study Hypothesis Level of Support 1.1. Maternal depression, in genera l, during the first three years postpartum is negatively associated with child social competence and display of problem be haviors at first grade. Strong support for mother rated social competence. No support for teacher rated social competence. Strong support for mother rated problem behaviors. 2.1 The relationship between ever exposure to maternal depression, during the first three years postpartum, and child social competence and display of problem behaviors at first grade is mediated by maternal sensitivity. Some support for mother rated social competence. Not tested for teacher rated social competence. Some support for mother rated problem behaviors. 3.1 The effect of ever exposure to maternal depression, during the first three years postpartum, on child social competence and display of problem behaviors at first grad e is stronger among mothers who rate low in maternal sensitivity. No support for mother rated social competence. Some support for teacher rated social competence. No support for mother rated problem behaviors. 4.1 The effect of ever exposure to maternal depression, during the first three years postpartum, on child social competence and display of problem behaviors at first grade is weaker among children exposed to nonmaternal care. No support for mother rated social competence. No support for teacher rated social competence. No support for mother rated problem behaviors. 1.2. Timing of initial onset of matern al depression, during the first three years postpartum, is associated with child social competence and display of problem behaviors at first grade, with earlier onset episodes having the greatest negative impact. Strong support for mother rated social competence. Strong support for teacher rated social competence. Strong support for mother rated problem behaviors. 2.2. The relationship between timing of initial onset of maternal depression and child social comp etence and display of problem behaviors at first grade is medi ated by maternal sensitivity. No support for mother rated social competence. No support for teacher rated social competence. Not tested for mother rated problem behaviors. Continued on the next page

PAGE 389

367 Table 100 (Continued) Study Hypothesis Level of Support 3.2. The effect of timing of initial onset of maternal depression on child social competence and display of problem behaviors at first grade is stronger among mothers who rate low in maternal sensitivity. No support for mother rated social competence. Some support for teacher rated social competence. No support for mother rated problem behaviors. 4.2. The effect of timing of initial ons et of maternal depression on child social competence and display of problem behaviors at first grade is weaker among children exposed to nonmaternal care. No support for mother rated social competence. No support for teacher rated social competence. No support for mother rated problem behaviors. 1.3. Severity of maternal depression al ong the trajectory of initial onset is associated with child social competence and display of problem behaviors at first grade, with severity of symptoms having the greatest negative impact w ith earlier onset episodes. Strong support for mother rated social competence. Moderate support for teacher rated social competence. Strong support for mother rated problem behaviors. 2.3. The relationship between severity of maternal depression, along the trajectory of initial onset, and ch ild social competence and display of problem behaviors at first gr ade is mediated by maternal sensitivity. Not tested for mother rated social competence. No support for teacher rated social competence. Not tested for mother rated problem behaviors. 3.3. The effect of severity of mate rnal depression, along the trajectory of initial onset, on child social competence and display of problem behaviors at first grade is str onger among mothers who rate low in maternal sensitivity. No support for mother rated social competence. Some support for teacher rated social competence. No support for mother rated problem behaviors. 4.3. The effect of severity of mate rnal depression, along the trajectory of initial onset, on child social competence and display of problem behaviors at first grade is w eaker among children exposed to nonmaternal care. No support for mother rated social competence. No support for teacher rated social competence. No support for mother rated problem behaviors. Continued on the next page

PAGE 390

368 Table 100(Continued) Study Hypothesis Level of Support 1.4. Chronicity of maternal depression is associated with child social competence and display of problem behaviors at first grade, with chronicity of symptoms having the greatest negative impact at later time points. Strong support for mother rated social competence. No support for teacher rated social competence. Strong support for mother rated problem behaviors. 2.4. The relationship between chronic ity of maternal depression and child social competence and displa y of problem behaviors at first grade is mediated by maternal sensitivity. Some support for mother rated social competence. Not tested for teacher rated social competence. Some support for mother rated problem behaviors. 3.4. The effect of chronicity mate rnal depression on child social competence and display of problem behaviors at first grade is stronger among mothers who rate low in maternal sensitivity. Some support for mother rated social competence. Some support for teacher rated social competence. No support for mother rated problem behaviors. 4.4. The effect of chronicity mate rnal depression on child social competence and display of problem behaviors at first grade is weaker among children exposed to nonmaternal care. Some support for mother rated social competence. Strong support for teacher rated social competence. No support for mother rated problem behaviors.

PAGE 391

369 Chapter Five: Discussion Synthesis of Research Findings Impact of Maternal Depression Rates of maternal depression during th e first three postpartum years for this study were similar to rates of postpartum de pression presented elsewhere, except for a slightly higher rate of mate rnal depression detected at one month postpartum. It is generally accepted that 10-15% of U.S. mo thers experience an ep isode of non-psychotic depression during the postpartum period (O 'Hara, 1995a; O'Hara & Swain, 1996), but 17.82% of mothers in this study had elevated CES-D scores at one month postpartum. Others have reported prevalence rates as hi gh as 31% for first-time mothers (Leathers, Kelley, & Richman, 1997), thus the rate of ma ternal depression at one month postpartum reported in this study falls with in a probable range. All ot her assessments of maternal depression (six, 15, 24, and 36 months) were within the estimated 10-15% ranging from 10.75% to 12.81%. Furthermore, previous literature (Campbell & Cohn, 1997; Stuart, Couser, Schilder, O'Hara, & Gorman, 1998) has sugge sted that for most women, maternal depression is highest during th e first few months postpartum after which it significantly declines. This study found that although maternal depression scores initially declined after one month postpartum through 15 months, maternal depression scores increased at 24 months, possibly coinciding with the child’ s more oppositional stag e of development. It has been documented that child ne gativity peaks around 24 months (Belsky,

PAGE 392

370 Woodworth, & Crnic, 1996), possibly making it more difficult for mothers to care for and cope with child behavior as compared to earlier, more compliant developmental stages (NICHD Early Child Care Research Network, 1999b). An alternative explanation for the slightly higher prevalence of maternal depression at 24 months postpartum observed in this study may also be related to a subse quent pregnancy or delivery. However, the percent of mothers depressed at 24 months (11.93%) remained lower than the percent of mothers depressed at six months (12.81%). Horowitz and Goodman (Horowitz & Goodman, in press) also found an initial decline in maternal depression scores after 4 to 8 weeks postpartum, followed by an increase at 24 months. Ever exposure to maternal depression during the first 36 months postpartum was found to be significantly associ ated with maternal reports of total social competence, assertion, peer competence, total problem be haviors, internalizi ng problem behaviors, and externalizing problem behaviors. Ever exposure to maternal depression was not significantly associated with maternal repo rts of cooperation, re sponsibility, or self control; nor teacher reports of total social competence, cooperation, assertion, self control, or peer competence. The fact that ever exposure to mate rnal depression had no impact on any of the teacher reported aspect s of child social competence implicates a possible reporting bias of mothers. Although this study purposefully excluded participants with mothers depressed at first grade to reduce maternal reporting bias in the context of concurrent depression, results from this study indicate a possible reporting bias of mothers who had ever experienced ma ternal depression during the first three postpartum years. Although several studies have documented a maternal reporting bias of child problem behaviors in the context of concurrent depression (Fergusson, Lynskey

PAGE 393

371 & Horwood, 1993; Lyons-Ruth, Easterbrooks & Davidson Cibelli 1997; Briggs-Gowan, Carter &Schwab-Stone, 1996 #665] only one study was found that investigated the extent of maternal reporting bias of probl em behaviors in the context of previous psychiatric illness [Chilcoat & Breslau, 1997 #666]. In this empirical study, Chilcoat and Breslau found a maternal reporting bias in the context of previous maternal psychopathology only for externalizing problem behaviors, not internalizing problem behaviors. More specificall y, previous history of maternal psychopathology was found to be associated with both teacher and mother reports of internalizing problem behaviors, but maternal history of major depression was only associated with mother rated externalizing problem behaviors, not teache r rated externalizing problem behaviors. These findings suggest that children of depressed moth ers do in fact have more internalizing problem behaviors, but that mothers overstate their child’s display of externalizing problem behaviors in the context of previous history of maternal psychiatric illness. It has been theorized that depressed mothers perceive and report more negative child behaviors due to cognitive distortions, a decreased tolerance for misbehavior, or increased sensitivity to depressive symptoms (Baumann, Pelham, Lang, Jacob, & Blumenthal, 2004). Perhaps these same mech anisms operate to explain the increased maternal ratings of negative externalizing be haviors in the context of prior exposure to maternal depression. No studies were found investigating a possibl e reporting bias of depressed mothers in rating child social competence. It is also possible that ev er exposure to maternal depr ession during the first three postpartum years had no significant impact on te acher reports of child social competence as a result of children behaving differently in home and school environments. Several

PAGE 394

372 researchers have supported the notion that disparities among mother and teacher rated behaviors are related, in part, due to the inherent experiences that the different informants have with the children (Offord, Boyl e & Racine, 1989; Achenbach, McConaughy & Howell, 1987). For example, parents largely observe children in home and family environments, where they possess historical information on the child’s behaviors, whereas teachers observe children in school environments and have more training on ageappropriate behavior (Briggs-Gow an, Carter & Schwab-Stone 1996). Timing of initial onset of maternal depression, during the first three years postpartum, was also found to be significantl y associated with child social competence and display of problem behaviors, with earl ier onset episodes having the greatest negative impact. For child social competence at firs t grade (maternal and teacher reported total social competence and peer competence, matern al report of assertion, and teacher report of self-control), the critical pe riod of risk for initial onset was six months. Children who were exposed to maternal depression at si x months postpartum (among those who had not been exposed at one month postpartum) were rated by mothers as displaying significantly less total social competence, assertion, and peer competence; and rated by teacher as displaying significantly less total social compet ence, self control, and peer competence. Maternal depression that in itiated after the sixth postpartum month (by 15, 24, or 36 months postpartum) did not have statisti cally significant impacts on child social competence at first grade. For child displa y of total problem beha viors and externalizing problem behaviors at first grade as reported by the mother, the critical period of risk for initial onset was one month. Only children e xposed to maternal depr ession that initiated at one month postpartum scored significan tly higher in total problem behaviors and

PAGE 395

373 externalizing problem behaviors. Depression that initiated at later time points (six, 15, 24, or 36 months) did not significantly imp act maternal report s of child problem behaviors. Given that mate rnal depression was only assessed at one, six, 15, 24, and 36 months postpartum, a significant finding of maternal depression initiating at one month more appropriately represents depression that initiated be tween delivery and one month postpartum and was still present at one month postpartum, given we are uncertain of the exact initiation time point betw een the data collection interv als. Likewise, a significant finding of maternal depression initiating at six months more appropriately represents depression that initiated between one and six months postpartum and was still present at six months postpartum. The significant imp act of earlier onset depressive episodes on child social competence and display of problem behaviors detected in this study is similar to another study (Brennan et al., 2000) that documented symptom onset at six months as a sensitive period of risk rela ted to child problem behaviors at age five (Brennan et al., 2000). Although Brennan (2000) found depression onset at six months as the critical period of risk related to child problem behavi ors, as compared to depression onset at one month detected in the present study, the ti me points assessed in Brennan’s study only included during pregnancy, immediately pos tpartum, six months, and age five. Severity of maternal depr ession along the trajectory of initial onset was found to be associated with child social competence and display of problem behaviors at first grade, with severity of symptoms (above thre shold levels indicating probable depression) having the greatest negative impact with earlie r onset episodes, for maternal reports of total social competence, self control, peer competence, total problem behaviors, and externalizing problem behaviors; and teacher report of total social competence. For

PAGE 396

374 maternal reports of social competence and pr oblem behaviors (total social competence, self control, peer competence, total probl em behaviors, and externalizing problem behaviors), depression severi ty at one month postpartum only, significantly impacted levels of social competence and problem beha viors. As depression severity increased, maternal rated social competence decreased and mother rated display of problem behaviors increased. Depressi on severity did not significantl y influence maternal rated child social competence or display of probl em behaviors among depressive episodes that initiated after the first postpa rtum month (six, 15, 24, or 36 mo nths). For teacher rated total social competence, depression severity at six months among children not previously exposed to depression, significantly impacted teacher report of peer competence. As depression severity increased, teacher ra ted peer competence decreased. Results indicated that severity of maternal depressive symptoms above the threshold level indicating probable depression was only significantly associated with maternal reports of child social competence and problem beha viors for depressive episodes initiating between delivery and one month postpartum a nd still present at one month postpartum, and is only associated with teacher report of total social competence for depressive episodes initial between one and six months postpartum and still present at six months postpartum. Severity of depressive sy mptoms beyond levels indicative of probable depression was not found to be significantly as sociated with child social competence of display of problem behaviors for depressive episodes that initia ted after six months postpartum (15, 24, or 36 months postpartum). Depression chronicity was also found to be significantly associated with child social competence and display of problem beha viors at first grade with chronicity of

PAGE 397

375 symptoms having the greatest negative impact at later time points. However, the time point at which depression chr onicity had the greatest magnitu de of negative influence on child social competence and display of pr oblem behaviors was 24 months postpartum rather than 36 months postpartum. This tre nd was detected for maternal reports of total social competence, self control, peer compet ence, total problem beha viors, internalizing problem behaviors, and externalizing problem behaviors. These results indicate that chronic depression does in fact pose a serious risk to child social competence and display of problem behaviors, but the negative impact of depression chronic ity does not increase after 24 months postpartum. Thus, chr onic depression contin uing beyond 24 months postpartum does not appear to contribute a dded risk to child social competence and display of problem behaviors. Note however that chronic depression was not significantly associated with any aspect of teacher repor ted social competence, calling into question again the potential for a maternal reporting bi as of mothers with pr evious experience with maternal depression. Other studies have docum ented a significant main effect of chronic maternal depression on child social comp etence and display of problem behaviors (Brennan et al., 2000; Civic & Holt, 2000; Dawson et al., 2003; NICHD Early Child Care Research Network, 1999b). Recurrent depression, or depression that recu rs after an episode of remittance, was also found to be significantly associated w ith child social compet ence and display of problem behaviors at first grade, with r ecurrent symptoms having the greatest negative impact at later time points (24 or 36 months postpartum). For the impact of recurrent depression on child social competence at first grade, in particular maternal reports of assertion and responsibility, recurrent depres sion at 36 months had a stronger negative

PAGE 398

376 impact than recurrent depression at 24 mont hs. For the impact of recurrent depression on child display of problem behaviors at first grad e, in particular maternal reports of total problem behavior and internalizing problem be haviors, recurrent depression at 24 months had a stronger negative impact than recu rrent depression at 15 or 36 months. This study also investigated the relative importance of each depression chronicity dummy variable within a singl e model revealing that when more than one chronicity dummy variable (new depression, remitted de pression, recurrent depression, or chronic depression) was significant, chronic depression had the greatest magnitude of negative impact, as judged by comparing each variable’s standardized regression coefficient, for maternal reports of peer competence and exte rnalizing problem behaviors. This trend on peer competence was detected for chroni c depression at six, 15, 25, and 36 months postpartum. For externalizing problem be haviors, this trend was found for chronic depression at 15, 24, and 36 months postpartum. Although chronic de pression at six and 15 months were found to have the greatest detr imental impact on maternal report of total social competence and maternal report of tota l problem behaviors resp ectively), at later time assessments (24 and 36 months) it was re current depression not chronic depression that posed the greatest risk to maternal re port of total social competence and maternal report of total problem behaviors. These re sults indicate that bot h recurrent and chronic depression pose serious thr eats to child social competence and display of problem behaviors. Mediating Effect of Maternal Sensitivity Evidence of a partial mediating effect of maternal sensitivity in the relationship between ever exposure to maternal depressi on during the first 36 m onths postpartum and

PAGE 399

377 child social competence and display of problem behaviors at first grade was detected. Average maternal sensitivity was found to par tially mediate the rela tionship between ever exposure to maternal depression and (1) matern al report of total so cial competence; (2) maternal report of assertion; (3) maternal re port of peer competence; (4) maternal report of total problem behaviors; and (5) maternal report of externalizi ng problem behaviors. However, the evidence for partial mediation wa s modest, particularly for maternal reports of assertion and peer competence. For ex ample, using an equation proposed by Harnish, Dodge, and Valente (Harnish, Dodge, & Valente, 1995), after adding the proposed mediating effect of maternal se nsitivity to the reduced model for maternal reports of total social competence, assertion, peer comp etence, total problem behaviors, and externalizing problem behaviors, the effect sizes of ever exposure to maternal depression during the first 36 months postpartum d ecreased 6.26%, 1.1%, 1.63%, 2.55%, and 2.48% respectively. No support of a mediating eff ect of maternal sensitivity was detected for maternal sensitivity at six months in the relationship between maternal depression initiating at six months postpartum and either (1) maternal report of total social competence; or (2) teacher report of total social competence. Fu rther, no evidence of a mediating effect of maternal sensitivity was detected for maternal sensitivity at six months in the relationship between depression severity at six months postpartum and teacher report of peer competence. Evidence of a mediating effect of maternal sensitivity in the association between chronicity of maternal depression and child social competence a nd display of problem behaviors was mixed. Maternal sensitivity at 36 months was found to partially mediate

PAGE 400

378 the association between chronic depression at 36 months (mothe rs were depressed at one, six, 15, 24, and 36 months) and (1) maternal report of total social competence; (2) maternal report of cooperation; (3) maternal report of self control; (4) maternal report of peer competence; (5) maternal report of tota l problem behaviors; and (6) maternal report of externalizing problem behaviors. Reductions in effect sizes of depression chronicity after the addition of maternal sensitivity to the regre ssion models were 8.99%, 10.25%, 6.69%, 5.29%, 4.24%, and 4.91%, respectively. Furthermore, maternal sensitivity at 36 months was found to partially mediate the relationship between recurrent depression at 36 months (mothe rs experienced an episode of depression at 36 months after an episode of remittance) and (1) maternal report of total social competence; (2) maternal report of assert ion; (3) maternal report of self control; (4) maternal report of peer competence; (5) mate rnal report of total pr oblem behaviors; and (6) maternal report of externa lizing problem behaviors. Redu ctions in effect sizes of recurrent depression at 36 months after the addition of maternal sensitivity were 5.45%, 4.65%, 5.06%, 3.33%, 2.67%, and 5.18%, respectively. In addition, maternal sensitivity at 24 months partially mediated the asso ciation between recurrent depression at 24 months and (1) maternal report of peer competence; and (2) maternal report of externalizing problem behaviors. Reductions in effect sizes of recu rrent depression at 24 months after the addition of maternal sens itivity were 1.90% and 2.28%. Last, maternal sensitivity at 24 months fully mediated the impact of recurrent depression on maternal report of responsibility. Models testing the mediating effect of ma ternal sensitivity at 15 months in the relationship between depression chronicity at 15 months and child social competence at

PAGE 401

379 first grade found no evidence of a mediating e ffect except for maternal report of self control. For this outcome variable, maternal sensitivity at 15 months partially mediated the impact of chronic depression at 15 mont hs (mothers were depr essed at one, 6, and 15 months). The effect size of chronic de pression at 15 months was reduced by 0.46% after the addition of maternal sensitiv ity, thus the partial mediation was small. No evidence of a mediating effect of maternal sensitivity at six months was found in the relationship between new depression at six months and mate rnal report of total social competence. Overall, evidence of a mediating effect of maternal sensitivity in the relationship between maternal depression and child so cial competence and display of problem behaviors was mixed. Maternal sensitivity appears to partially explain the negative impact of maternal depre ssion on child social competen ce and display of problem behaviors only when averaged across the first 36 months postpartum, or when assessed at later time points in the postpartum peri od, particularly 24 or 36 months. Maternal sensitivity at six months was never detected to mediate the impact of maternal depression (initial onset or seve rity) on total social competence; and maternal sensitivity at 15 months was only found to partially mediate th e association between depression chronicity and child social competence in one of the four models tested. On the other hand, maternal sensitivity at 24 and 36 months were consistently found to partially mediate (and in one instance fully mediate) the impact of recurrent or chronic depression on child social competence. It is important to em phasis however that all partial mediations detected were modest in magnitude accounting for only a small percent of the total effect of maternal depression on child social co mpetence and display of problem behaviors, ranging from 0.46% to 10.25%.

PAGE 402

380 No other studies have inves tigated the mediating effect of maternal sensitivity in the association between maternal depression and child developmental outcomes to the same degree of specificity as this study. Th erefore it is difficult to compare study results to other literature. For example, one study (Harnish et al., 1995) did investigate the mediating effect of harsh and insensitive pa renting in the association between maternal depression and child problem behaviors at first grade. This study however assessed maternal depression and mother-child interact ion, of which maternal sensitivity was a large component, much later in the post partum period – during the summer between kindergarten and first grade. Results did show that by including the mediating effect of interaction quality between mother and chil d, the effect of maternal depression was reduced by 19%. Another study also investigat ed the mediating effect of maternal sensitivity on the association between chronic depr ession classification and matern al reports of child social competence and display of problem behaviors at 36 months, revealing no evidence of an indirect relationship (NICHD Early Child Care Research Network, 1999b). This study also conducted secondary data analysis of the NICHD SECC and operationalized maternal sensitivity as the average level of ma ternal sensitivity across the first 36 months postpartum. The maternal sensitivity com posite however was created in a slightly different manner. Whereas this study created the average maternal sensitivity composite by first standardizing maternal sensitivity at each time point to have a mean of zero and standard deviation of one before summing and averaging the items (due to the fact that maternal sensitivity was assessed on a fourpoint scale at six, 15, and 24 months and a seven-point scale at 36 months ), the researchers in the ot her study recalibrated the seven-

PAGE 403

381 point scale to a four-point scale. This other study al so operationalized depression differently than the current study, both conditions of which may have impacted differential study findings. Nevertheless, all partial mediations of ma ternal sensitivity detected in this study were modest in magnitude, accounting for at most 10.25% of the effect of maternal depression. In addition, it has increasingly been shown that among depressed mothers, there is wide variation in sensitivity levels and paren ting skill (Campbell & Cohn, 1997; Frankel & Harmon, 1996). These finding implic ate that something other than maternal sensitivity is acting as the primary mechanism through which maternal depression poses a risk to child social competence and displa y of problem behaviors, warranting further research.

PAGE 404

382 Moderating Effect of Maternal Sensitivity Maternal sensitivity was found to be a significant moderator on the association between ever exposure to maternal depressi on during the first 36 m onths postpartum and teacher report of total social competence at first grade, with maternal depression having the greatest negative impact on total social competence among mothers rated lower in maternal sensitivity. No other moderating in fluences of maternal sensitivity on the association between ever exposure to matern al depression and child social competence and display of problem behaviors we re detected in the present study. Furthermore, maternal sensitivity was also found to be a significant moderator on the association between timing of initial onset of maternal depression and child social competence. Using a Bonferroni-adjuste d p value of p<.01 to judge statistical significance (0.05/5 given five in teractions were tested for each dependent variable), maternal sensitivity at 15 months was f ound to moderate the association between maternal depression initiating at 15 months postpartum and teacher report of assertion (interaction term estimates: b=-3.60, t=-2.75, p< .01). However, exposure to first time depression at 15 months had the most detr imental impact among children with more sensitive mothers. It appears that children with more sensitive mothers face greater risk to the development of assertive skills in the context of exposure to later onset depressive episodes, as compared to children with less sensitive mothers who perhaps have already compensated for their mothers’ lower leve l of responsiveness to their needs. Evidence of a moderating effect of maternal sensitivity at six months was detected on the relationship between depression severity at six months, among children not

PAGE 405

383 previously exposed to maternal depression, a nd teacher report of peer competence using a Bonferroni-adjusted p value of <.01 to judge statistical significance (interaction term estimates: b=0.21, t=2.63, p<.01). Depression se verity at six months had a stronger negative impact on teacher repo rt of peer competence among children with mothers rated lower in maternal sensitivity. Several moderating effects of maternal sensitivity on the association between depression chronicity and child social comp etence were detected using a Bonferroni adjusted p value of <.01. Maternal sensitiv ity at six months was found to moderate the association between remitted depression at six months (or depression present at one month postpartum and not present at six m onths postpartum) and maternal report of cooperation (interaction term estimates: b= 1.82, t=3.09, p<.01). For this interaction term, remitted depression had a more negative impact on children with mothers rated lower in maternal sensitivity. On the othe r hand, maternal sensitivity at 15 months was found to moderate the associ ation between new depression at 15 months and teacher report of assertion (interaction term es timates: b=-3.52, t=-2.69, p<.01), with new depression at 15 months having a greater nega tive effect on teacher reported assertion among children with mothers rated higher in maternal sensitivity. Thus, maternal sensitivity was found to buffer the detrimen tal impact of remitted depression at six months, and pose additional risk on the impact of new depression at 15 months on teacher reported assertion. Overall, moderate support of a buffering e ffect of high maternal sensitivity in the context of maternal depression was dete cted. Evidence however suggests that the protective influence of having a mother rated higher in ma ternal sensitivity only holds

PAGE 406

384 true for children exposed to earlier onset ep isodes. For each modera ting effect detected for maternal depression initia ting after six months postpar tum (i.e. 15 months), higher levels of maternal sensitivity were no longer a protective factor but instead a risk factor. It is also interesting to no te that no moderating effects of maternal sensitivity were detected for chronic or recurrent depression occurring at 24 or 36 months, implicating no buffering influence of high maternal sensitiv ity, or additional risk of low maternal sensitivity, for longer sustained depressive ep isodes or those that come and go over the first two to three years postpartum. Last, it is interesting to note that when using a less stringent p value to determine statistical significance (p<.05), maternal sens itivity at 24 months was found to moderate the associations between (1) depression init iating at 24 months postpartum and maternal report of assertion (interaction term estim ates: b=1.26, t=2.29, p<.05); and (2) depression severity at 24 months, among children not previously exposed to depression, and maternal report of assertion (interaction term estimates: b=-0.13, t=2.46, p<.01). These moderating effects however show a protective influence of hi gh maternal sensitivity on the relationship between depressi on initiation at 24 months (o r depression severity at 24 months among children not prev iously exposed to depression ) and maternal report of assertion, in contrast to th e negative influence of high maternal sensitivity reported earlier. Thus, based on maternal perception of assertion at first grade, high maternal sensitivity buffers the negative impact of depr ession initiating at 24 m onths (whether it is presence of depression or severity of symp toms); and based on teacher perception of assertion at first grade, high maternal se nsitivity amplifies the negative impact of depression initiating at 24 mont hs (whether it is presence of depression or severity of

PAGE 407

385 symptoms). Given that child assertion en compasses behaviors such as making friends easily, initiating conversations with others, joining gr oup activities without being prompted, and inviting others to join activities, perhaps children display different behaviors in school and home environments thus explaining the c ontradictory findings related to the moderating effect of maternal sensitivity. Then again, it is important to recall that these buffering effects of matern al sensitivity on the association between later onset depression (24 months) on maternal re port of cooperation were only significant at p<.05, not the Bonferroni adjusted value of p<.01, questioning whether interpretation is warranted. Moderating Effect of Expos ure to Nonmaternal Care Exposure to nonmaternal care, operati onalized as both ever exposure to nonmaternal care and ever exposure to center-based nonmaternal care, was not a significant moderator on the a ssociation between ever expos ure to maternal depression and child social competence and display of problem behaviors at first grade. In addition, using a Bonferroni-adjusted p value of p<.005 to judge statistical significance (0.05/10 given that ten sets of interaction terms we re tested for each dependent variable), no moderating effects of ever e xposure to nonmaternal care or ever exposure to center-based nonmaternal care were detected on the associat ions between (1) timing of initial onset of maternal depression and child social competen ce and display of problem behaviors; or (2) severity of depression along the trajectory of initial onset on child social competence and display of problem behaviors. For chronicity of maternal depressio n, significant moderating effects of ever exposure to nonmaternal care were detected, us ing a Bonferroni adjusted p value of less

PAGE 408

386 than .005, on the associations between chroni c depression at 24 months (mothers were depressed at one, six, 15, and 24 months) and maternal report of assertion (interaction term estimates: b=5.27, t=2.91, p<.005), teache r report of total social competence (interaction term estimates: b=31.73, t=3.02, p<.055), teacher report of cooperation (interaction term estimates: b=9.18, t=3.04, p<.005), and teacher report of peer competence (interaction term estimates : b=8.48, t=3.08, p<.005). For each of these moderating effects of ever e xposure to nonmaternal care at 24 months, ever exposure to nonmaternal care buffered the detrimental imp act of chronic depression on child social competence. No moderating effects of cen ter-based nonmaternal care on the association between depression chronicity and child so cial competence and display of problem behaviors were detected using p<.005 to judge statistical significance. Although other moderating effects of ever exposure to nonmaternal care and ever exposure to center-based nonmaternal care we re detected at p<.05, they will not be discussed in further detail he re given the number of moderati ng effects detected. Overall however, results are mixed showing both a posit ive and negative effect of ever exposure to nonmaternal care (in general or center-bas ed) for depression chronicity at later time points, with more buffering effects detect ed than not. When ever exposure to nonmaternal care (in general or center-based) was found to pot entiate to negative impact of depression chronicity on child social competence, it wa s increased risk for children exposed to remitted or recurrent depression, not new or chronic. This later finding raises the possibility that when attempting to understa nd the moderating effect of ever exposure to nonmaternal care, positive effects are usua lly detected when maternal states are

PAGE 409

387 relatively stable as they are with new or chronic depression, rather than inconsistent as they are with remitted or recurrent depression. This study hypothesized a buffering effect of exposure to nonmaternal care in the context of maternal depr ession given that other resear ch has suggested a protective effect of child care among children from disa dvantaged circumstances including poverty. For this study, exposure to maternal depres sion was perceived to be the condition of disadvantage facing children. For example, one U.S. study (O'Brien Caughy, DiPietro, & Strobino, 1994) found that day-care participa tion during the first th ree years of life, particularly initiation before the child’s firs t birthday, was associated with higher reading recognition scores at ages five and si x among children from impoverished home environments. This study also concluded th at center-based care more strongly buffered the impact of poverty on mathematic scores Another study (Desai, Chase-Lansdale, Michael, & 1989, 1989) also found a significan t interaction between nonmaternal care and poverty, however this study detected no effect of nonmaternal care on cognitive development of 4-year-old children from low-income families, but found a negative impact of nonmaternal care on cognitive abil ities among boys from high-income families. Other studies have not found a protective eff ect of exposure to nonmaternal care among disadvantaged youth related to child deve lopmental outcomes (Burchinal, PeisnerFeinberg, Bryant, & Cliffor d, 2000; NICHD Early Child Care Research Network, 2002; Stipek, Feiler, Daniels, & M ilburn, 1995). Thus the literatu re investigating a positive influence of exposure to non maternal care is mixed. This study however did detect a buffe ring impact of ever exposure to nonmaternal care, however only for chronic de pression. For children exposed to mothers

PAGE 410

388 with chronic depression through the first two y ears of life, having at least some exposure to nonmaternal care during the same time pe riod helps offset the negative effect of depression on child social competence. Th e fact that this trend was found for both maternal and teacher reported aspects of soci al competence, and was deemed true using a stricter criteria to assess statistical significance (p<.005), suggests stronger support for this conclusion. Attachment Theory Although not the main purpose of this study, inferences can be made regarding the usefulness of attachment theory in explai ning child social competence and display of problem behaviors at first grade in the cont ext of maternal depre ssion during the first 36 months postpartum. Maternal sensitivity, a ke y construct in attachme nt theory, is defined as a mother’s ability to notice infant si gnals, interpret signals accurately, and respond reasonably promptly and appropriately to such signals (Ainsworth & Bell, 1974). According to attachment theory, variations in the quality of the attachment bond between mother and child across mother-infant dyads result in part due to differential levels of maternal sensitivity (Bowlby, 1969). Further, infants and children construct internal working models, or internalized representa tions of the caretaker’s accessibility and responsiveness, based on their history of prior interaction with the ca regiver. It is the internal working model that guides all future in teractions with the ca regiver and serves as the template of interaction between the child and other individuals in the child’s social environment (Bowlby, 1973; Thompson, 1999). It is this mechanism through which the present study hypothesized a mediating effect of maternal sensitivity in the association

PAGE 411

389 between maternal depression during the firs t 36 months postpartum and child social competence and display of problem behaviors at first grade. Results from this study provide mixed evid ence of a mediating effect of maternal sensitivity. Maternal sensitivity was found to partially explain the negative impact of maternal depression on child social competen ce and display of problem behaviors, but only when averaged across the first 36 months pos tpartum, or when assessed at later time points in the postpartum period, particularly 24 or 36 months. Maternal sensitivity at six months was never detected to mediate the imp act of maternal depres sion on total social competence; and maternal sensitivity at 15 months was only found to partially mediate the association between maternal depression and child social competence in one of the four models tested. On the other hand, ma ternal sensitivity at 24 and 36 months were consistently found to mediate (either partially of fully) the impact of recurrent or chronic depression on child social competence. Th e significant mediating effect of maternal sensitivity assessed at later time points may be a function of the four phases outlined by Bowlby through which the attachment relations hip between child and primary caregiver develops. Bowlby stated that true emotiona l attachment between infant and caregiver does not emerge until phase thr ee of attachment development, “Maintenance of Proximity to a Discriminated Figure by Means of Locomo tion as Well as Signals,” which is stated to occur anytime between six months and three y ears. This is the time point at which the emotional bond between mother and child has fo rmed and is becoming internalized as a working model to guide future social interacti ons. Thus, perhaps maternal sensitivity at 24 and 36 months were found to mediate the ne gative impact of maternal depression on child social competence and display of problem behaviors, as compared to non-

PAGE 412

390 significant findings for maternal sensitivity at earlier time points, gi ven that children in earlier phases of attachment development have yet to internalize their expectations of their mothers’ responsiveness. Although study results provided some support of the use of attachment theory in explaining child social competence and displa y of problem behavior s in the context of maternal depression during the first 36 months pos tpartum, it is critical to highlight that all partial mediations of maternal sensitivity detected were modest in magnitude accounting for only a small percent of the tota l effect of maternal depression on child social competence and display of problem behaviors, ranging from 0.46% to 10.25%. Thus, despite the statistical si gnificance of maternal sensitiv ity detected in some models tested in this study, findings implicate that maternal sens itivity is not the primary mechanism through which maternal depression poses a risk to child social competence and display of problem behavi ors, suggesting that other developmental theories may prove more valuable in understanding the causal mechanisms of risk.

PAGE 413

391 Study Limitations Results from this study should be interpreted with caution due to the following limitations. First, the study was longitudinal in nature co llecting data in two phases during which 144 (10.6%) families were lost to attrition. These families were more likely to have mothers that were young, non-White, less educated, living with no husband or partner at home, and receiving public assistan ce. The sample was further limited by only including families that had complete data on all variables of interest in the study and those that did not have depressed mothers at first grade. Families omitted due to exclusion criteria (n=544; 50.22% of the original 1364 enroll ed in Phase I) were more likely to have mothers that were young, nonWhite, less educated, living with no husband or partner at home, receiving public assi stance, not breastfeeding at one month postpartum, experienced depression at some point during the first 36 months postpartum, and had lower levels of maternal sensitivity at one month postpartum. Therefore, respondents included in this study tended to ha ve less social and dem ographic risk factors than those that excluded from the study. Caution should be used in generalizing the findings of this study beyond the study sample used. Further, the CES-D utilized to measure ma ternal depression in this study is not a diagnostic tool for maternal depression, but instead assesses depr essive symptomatology in community samples. Women scoring at or above suggested cut points indicative of depression may not in fact be experiencing actual depression, but instead elevated depressive symptoms. Further, the use of depression assessment tools developed for community samples on postpartum women has b een criticized (J. Cox et al., 1987). Critics state that a lthough these instruments have va lidity among general populations, use

PAGE 414

392 among postpartum women may elevate the prev alence of depression given that they emphasize somatic symptoms of psychiatric di sorders that characterize expected and normal physiological changes associated w ith the postpartum pe riod including fatigue, poor appetite, and altered sleep patterns. Nevertheless, the CES-D instrument has been employed in numerous studies to assess th e impact of postpartum depression on child social competence and display of problem be haviors (Carter et al ., 2001; Civic & Holt, 2000; Dawson et al., 2003; Hubbs-Tait et al ., 1996; Lyons-Ruth et al., 1997; NICHD Early Child Care Research Network, 1999b). Further, rates of maternal depression during the first three postpartu m years observed in this study were similar to rates of postpartum depression presented elsewhere. The assessment tool used to rate observed maternal sensitivity was created for implementation in the NICHD SECC, with little documentation of psychometric properties including reliability or validity. In addition, th e response options to rate maternal sensitivity at six, 15, and 24 months uti lized a four-point scal e. Given that little variance in levels of maternal sensitivity we re detected in the present study, perhaps a more finely graduated rating instrument would have resulted in broader ranges of observed maternal sensitivity. Maternal sensit ivity was also observed in relatively brief samples of semi-structured play situations. It is possible that more extensive and naturalistic measurements of maternal sensitivity may have produced different conclusions regarding the mediating and mode rating effects of maternal sensitivity. Due to lack of variance in categorical cl assifications of child social competence and display of problem behavi ors, the effect of maternal depression on these dependent variables could not be studied. For example, for both maternal and teacher reports of

PAGE 415

393 child total social competence, only 47 (6.92 %) children were rated to have less than average social competence. In addition, onl y 57 (8.39%) children were rated as having borderline deviant or clinically significant devi ant displays of total problem behaviors. Due to the number of sociodemographic cont rol variables that were planned to be included in the regression models, as well as the number of mediating and moderating influences to be investigat ed, not enough statistical power was available to explore the impact of maternal depressi on on these categorical classifi cations of risk. A larger sample size is necessary to predic t behavior with small variance. Although assured that information collect ed during study participation would be confidential, data collection of ten occurred in the context of a home interview, with the home visitor having full knowledge of the part icipant’s identity and living situation. Many items ascertained during the home visit were of a sensitive nature including experiences of depression. This scenario may have biased maternal responses. In addition, maternal behaviors may have been altered as a result of data collection procedures. In the case of the Mother-Child Interaction pr ocedure, mothers were fully aware that they were being observed in inte raction with their child. This may have promoted social desirability bias and decrea sed sample variation related to maternal sensitivity. The NICHD SECC study sample is not a na tionally representative sample thus results can only be generalized to the populati on of infants and their related families born in participating hospitals with in the ten sites during 1991. Results cannot be generalized to infants born to mothers under the age 18, t hose not fluent in E nglish, and those with medical conditions including substance abuse. Results also cannot be generalized to

PAGE 416

394 infants that were part of a multiple bi rth, and those that experienced medical complications. Restrictions in the sample to families who resided in neighborhoods considered safe for home visitation, base d on police reports, also may impact the generalizability of findings. Last, infants born in 2005 may be diffe rent from infants born in 1991 further restricting the ge neralizations that can be made from the present study. Last, given the number of dependent variables ascertaining child social competence tested for each research hypothe sis, including total social competence, cooperation, assertion, responsibil ity, self control, and peer competence, caution should be used when interpreting or reporting subscale scores, given variability of results To maximize the generalizability of research findings from this study, it is recommended that total social competence scores, as rated by both mothers and teachers, rather than subscale scores, be used. Study Strengths This study has several strengths. Data were derived from a longitudinal, prospective sample of children and families that began when the child was one month old. The use of longitudinal data allows for powerful inferences of causality to be made, and well as further investigation into the char acteristics of maternal depression that pose the greatest risk to child social developmen t including timing of initial onset, severity of symptoms, and chronicity of symptoms. Non-consideration of the impact of heterogeneity in a mother’s experience with depression on subsequent child development has been a tremendous limitation of previous research on this t opic. The use of longitudinal data in this study enabled the res earcher to separate out the effects of timing of initial onset of maternal depression, as we ll as severity of depressive symptoms along

PAGE 417

395 the trajectory of initial onset. In addition, the availability of l ongitudinal data with multiple assessments of maternal depression during the postpartum period enabled the researcher to more adequately assess the imp act of depression chr onicity on child social competence and display of problem behaviors. Previous studies that investigated the impact of depression chronicity on child outco mes reduced the power of longitudinal data by combining multiple assessments into a sing le count variable indicating the number of times a woman had experienced depression dur ing a given time interval. Not only does this treatment of maternal depression reduce the amount of data available and subsequently decrease statistical power, but it also removes the importance of temporal order of depressive episodes. For example a woman that experienced two consecutive episodes of depression early in the postpartum period would be classified the same as a woman who experienced one episode early in the postpartum period, followed by several episodes of remission, then another episode of depression later in the postpartum period. The impact of depression for these two wo men would most likely not be the same. Further, although the sample used for th is study was a much smaller sample of eligible participants enrolled in Phase I, children and families excluded from this study were found to have more social and demogra phic risk factors than study participants. Thus, the negative effect of ma ternal depression on child so cial competence and display of problem behaviors detected in this study may be an undere stimate. The magnitude of impact of maternal depressi on on child social development is likely much greater in a more socially at-risk diverse sample. Also, even though th e NICHD SECC Phase I sample was not purported to be nationally representative, a comparison of Phase I participants with other indi viduals living in the same census tracts and nationally,

PAGE 418

396 revealed that the NICHD SECC sample was similar to both comparison groups related to household income and ethnicity (NICHD Earl y Child Care Research Network, 2000). The prevalence of maternal depression detect ed in this study was also similar to the prevalence of maternal depression reported elsewhere. This study controlled for several impor tant sociodemographic factors including child sex, child birth order, maternal age, maternal race, maternal ethnicity, maternal education, presence of husband or partner living at home at one month postpartum, family receipt of public assistance at one month postpartum, and maternal rates of breastfeeding at one month postpartum. Previ ous literature has larg ely failed to control for the full range of important social and de mographic factors that might influence the association between maternal depression a nd child social competence and display of problem behaviors. Furthermore, this study i nvestigated the mediati ng effect of maternal sensitivity, and the moderating effects of maternal sensitivity and exposure to nonmaternal care. Few studies have explored more than th e main effect of maternal depression. Last, this study excluded children with depresse d mothers at first grade to reduce the potential for maternal reporting bias in the context of ma ternal depression. Implications for Public Health Maternal depression has received much at tention in the public health and child development literatures documenting a negative effect of child exposure to maternal depression on later child development. Results from this study provide additional evidence that maternal depression increases risk to social development, in particular child social competence and display of problem behaviors.

PAGE 419

397 For example, ever exposure to maternal depression during the first 36 months postpartum was found to be negatively associat ed with maternal ratings of child social competence (total social competence, asse rtion and peer competence), and positively associated with maternal ratings of child display of problem beha viors (total problem behaviors, internalizing problem behaviors, and externalizing pr oblem behaviors). Timing of initial onset of maternal depre ssion, during the first th ree years postpartum, was also found to be significantly associated with child social competence and display of problem behaviors, with earli er onset episodes having the greatest negative impact. For maternal and teacher reported total social competence and peer competence, maternal report of assertion, and teacher report of self-control, the criti cal period of risk for initial onset was six months postpartum. Children w ho were exposed to maternal depression at six months postpartum (among those who had not been exposed at one month postpartum) were rated by mothers as displaying significantly less total social competence, assertion, and peer competen ce; and rated by teachers as displaying significantly less total social competence, self control, and peer competence. Maternal depression that initiated after the sixt h postpartum month (by 15, 24, or 36 months postpartum) did not have statistically signi ficant impacts on child so cial competence at first grade. For child display of total problem behaviors and externalizing problem behaviors at first grade, as re ported by the mother, the critical period of risk for initial onset was one month. Only children exposed to maternal depression th at initiated at one month postpartum scored significantly hi gher in total problem behaviors and externalizing problem behaviors. Depression that initiated at later time points (six, 15, 24, or 36 months) did not significantly imp act maternal report s of child problem

PAGE 420

398 behaviors. Thus, timing of initial onset of maternal depression is important, with episodes initiating at six months postpartum having the most detrimental impact on child social competence, and episodes initiating at one month postpartum having the most detrimental impact on child display of problem behaviors. Furthermore, severity of maternal depre ssion along the trajector y of initial onset was found to be associated with child so cial competence and display of problem behaviors at first grade, with severity of symptoms having the greatest negative impact with earlier onset episodes, for maternal reports of total social competence, self control, peer competence, total problem behaviors, and externalizing problem behaviors; and teacher report of total social competence. For maternal reports of social competence and problem behaviors, depression severity at one month postpartum, only, significantly impacted levels of social competence and pr oblem behaviors. As depression severity increased, mother rated social competence decr eased and mother rated display of problem behaviors increased. Depressi on severity did not significantl y influence maternal rated child social competence or display of probl em behaviors among depressive episodes that initiated after the fi rst postpartum month (six, 15, 24, or 36 months). Comparing these findings with the ones just presented related to the impact of timing of initial onset of depression, one can see that for child social competence, the presence of depression alone at one month postpartum is not enough to warra nt risk, but the severity of depressive symptoms above threshold (<16) does increase ri sk to child social competence. For both total problem behaviors and ex ternalizing problem behaviors, the presence of depression as well as the severity of symptoms above th reshold increases risk to child display of problem behaviors.

PAGE 421

399 For teacher rated social competence, de pression severity at six months among children not previously exposed to depression, significantl y impacted teacher report of peer competence. As depression severity increased, teacher rated peer competence decreased. As compared to findings reported on the impact of initial onset of depression at six months on teacher reported social co mpetence, these results highlight that the presence of depression alone does not increase risk to peer competence as reported by the teacher, but the level of severity of sympto ms above threshold does. Given that some question of a maternal rating bias among moth ers who have previously experienced an episode of depression has been raised due to cognitive distortions, increased negativity, or a decreased tolerance for misbehavior, a si gnificant impact of depression severity on child social competence as reported by the chil d’s teacher is particularly powerful. Depression chronicity was also found to be significantly associated with child social competence and display of problem beha viors at first grade with chronicity of symptoms having the greatest negative impact at later time points. However, the time point at which depression chr onicity had the greatest magnitu de of negative influence on child social competence and display of pr oblem behaviors was 24 months postpartum rather than 36 months postpartum. This tre nd was detected for maternal reports of total social competence, self control, peer compet ence, total problem beha viors, internalizing problem behaviors, and externalizing problem behaviors. These results indicate that chronic depression does in fact pose a serious risk to child social competence and display of problem behaviors, but the negative impact of depression chronic ity does not increase after 24 months postpartum. Note however th at chronic depression was not significantly associated with any aspect of teacher repor ted social competence, calling into question

PAGE 422

400 again the potential for a maternal reporting bi as of mothers with pr evious experience with maternal depression. Recurrent depression, or depression that recu rs after an episode of remittance, was also found to be significantly associated w ith child social compet ence and display of problem behaviors at first grade, with r ecurrent symptoms having the greatest negative impact at later time points (24 or 36 months postpartum). This study also investigated the relative importance of each de pression chronicity dummy vari able within a single model revealing that when more than one chronicity dummy variable was significant, chronic or recurrent depression had the greatest ma gnitude of negative impact, as judged by comparing each variable’s standardized re gression coefficient. Although chronic depression at six and 15 months were found to have the greatest detrimental impact on maternal report of total social competence a nd maternal report of total problem behaviors respectively), at later time assessments ( 24 and 36 months) it was recurrent depression not chronic depression that posed the greatest risk to maternal report of total social competence and maternal report of total problem behaviors. These re sults indicate that both recurrent and chronic depression pose seri ous threats to child social competence and display of problem behaviors. In summary, maternal depression during th e first 36 months postpartum poses risk to child social development at first grad e including social competence and display of problem behaviors, warranting increased public health efforts for early detection and treatment of maternal depression. Experts ha ve advocated that primary care clinicians become more involved in the management of depression (Depression Guideline Panel, 1993), and pediatricians have been identified as an underutilized resource to mothers

PAGE 423

401 experiencing postpartum depression (Hene ghan, Johnson Silver, Bauman, & Stein, 2000). Although pediatricians are the primary care provi ders for children not mothers, pediatric guidelines have emphasized the role of pe diatricians in detecting family problems including maternal depression (Green & Palf rey, 2002). Mothers are often not in regular contact with their own h ealthcare provider, but have fre quent visits to their child’s pediatrician during the first fe w years of the child’s life for well baby visits. It has been suggested that pediatric visits are the most consistent ongoin g contact mothers have with a healthcare provider (Olson et al., 2005), identifying pediatrici ans as a key player in the secondary prevention of maternal depressi on. Healthcare practitioners need to be educated about the negative in fluences of postpartum depression on child development, trained to recognize maternal depression th rough questioning or use of a screening instrument, and advised how best to refer women for treatment. Several screening instruments have been advocated to assi st healthcare providers identify maternal depression including the Edinbur gh Postnatal Depression Scal e (EPDS) (J. Cox et al., 1987), the Zung Self-Rating Depression Scal e (Zung, 1965), the Psychiatric Symptom Index (PSI) (Ilfeld, 1976), and the Center for Epidemiologic Studies Depression Scale (CES-D) (Radloff, 1977). Each of these inst ruments are short, self-administered tools that can be completed by the mother wh ile waiting to be seen by the healthcare professional, and reviewed by the provider durin g the office visit. Other pediatric roles for the promotion of child health in the cont ext of maternal depres sion have included the provision of patient education, supportive listen ing, and practical suggestions to enhance parenting skills and reduce stress, as well as encouraging the mother to seek additional evaluation and treatment from a me ntal health practitioner, if needed (Olson et al., 2005).

PAGE 424

402 These practices including routine screeni ng for maternal depression hold promising potential to combat the negative impact of maternal depression on child social development. With regards to timing of initial onset and sensitive periods of risk, this study found maternal depression initia ting during the first six mont hs postpartum to have the greatest negative impact on child social develo pment. Depression initiating at six months postpartum predicted poorer so cial competence as rated by the mother, and depression initiating at one month postpartum predicted higher displays of child problem behaviors at first grade, also reported by the mother. These findings underscore the importance of early detection and treatment of maternal depression by healthcare practitioners, in particular pediatricians, dur ing the first six months postpartum. Pediatricians have reported time as a critical barrier in their ability to identify ma ternal depression during office visits (Olson et al., 2005), but results from this study highlight the need to incorporate screening methods, at least, during the early pedi atric visits including two to four weeks, two months, four months, and six months, given that initial exposure to maternal depression during the first six mont hs had the strongest detrimental impact on child social development. Emphasis on the pe diatrician’s role in the identification and treatment of maternal depression in this section of the discussion does not intend to preclude to important role of other health care provider s including obstetricians and primary health care providers, but rather to highlight an underused and frequently visited provider. Furthermore, efforts to educate health care providers about maternal depression and encourage routine screening during regular office visits should also emphasize the

PAGE 425

403 important role of symptom severity related to child outcomes. In some instances, this study found that depression alone did not incr ease risk for lower levels of social competence, but severity of symptoms above cut points indicating depression did. Use a structured screening instrument for maternal depression such as the EPDS or the CES-D would assist practitioners with quick assessm ent of the relative severity of symptoms being experienced by a new mother to help de termine and guide the appropriate course of treatment. More stringent follow-up on recommended referrals to mental health counselors, or more in-depth questioning and assessment during subsequent well child visits, may be warranted based on the moth er’s level of depr ession severity. This study also found that both chronic and recurrent depression pose risk to child social competence and display of problem be haviors at first grad e. Thus, although the first six months postpartum was previously priori tized as the critical period to screen for maternal depression, child health promotion e fforts in the context of maternal depression would be optimized if routine screening were conducted at each well baby visit during the first few years postpartum, at least thr ough the child’s second year of life. Results from this study found that although chronic depression was a strong predictor of lower levels of social competence and higher leve ls of problem behavior s, it was depression chronicity at 24 months, not 36 months, that had the greatest magnitude of negative influence, implying that depression chroni city beyond 24 months contributed no added risk to child social development. Practitione rs need to be educat ed on the importance of screening for maternal depression beyond th e first postpartum year, and concessions made to allow for appropriate time to addr ess maternal mental health issues during

PAGE 426

404 routine office visits, given that inadequate ti me has been reported as a common barrier to identification and management of mate rnal depression (Olson et al., 2005). Furthermore, increased public awarene ss on the predictors and symptoms of maternal depression, during both the early postp artum period as well as into the child’s preschool years, would substantially contribut e to positive child development. There are many misperceptions in the community relate d to maternal depression such as (1) confusion between the various types of pos tpartum mood disorders including maternity blues, postpartum psychosis, and postpartum depression; and (2) mi sinterpretation of symptoms such as sleep and appetite distur bances as normal experiences of postpartum women. Educating prenatal a nd postpartum women, as well as other individuals in the woman’s social network, on the prevalen ce, symptoms, and impact of maternal depression on child social development woul d hopefully (1) decrease feelings of shame and stigma associated with mental diso rders including postpartum depression; (2) increase one’s ability to recognize early sign s of depression; and (3) encourage women to not only seek and initiate treatment but conti nue with prescribed interventions. This study also found that the impact of recurrent depr ession, or depression that fluctuated over time, was a severe threat to child social co mpetence and display of problem behaviors, in some cases even having a greater negative im pact on outcome than chronic depression. Thus, adherence to treatment plans and st ability of mood are important factors. Related to mediating effects, maternal sensitivity was only found to mediate the association between maternal depression a nd child social competence and display of problem behaviors when averaged across th e first 36 months postpartum, or when assessed at later time points, particularly 24 or 36 months. Maternal sensitivity at six

PAGE 427

405 months was never detected to mediate the imp act of maternal depres sion on total social competence; and maternal sensitivity at 15 months was only found to partially mediate the association between maternal depression and child social competence in one of the four models tested. On the other hand, ma ternal sensitivity at 24 and 36 months were consistently found to mediate (either partially or fully) the impact of recurrent or chronic depression on child social competence. Th ese finding most likely relate to the various stages of attachment development outlined by Bowlby. According to attachment theory, a true emotional bond between mother and ch ild does not truly form until phase three, which occurs anywhere between the ages of six and 36 months, usually once the child begins locomotion. This is the time point at which child perceptions of maternal sensitivity become internalized. If one accepts that a child’s internalized expectancies of his or her mother’s behavior does not occur until around 24 months, this could explain in part why maternal sensitivity prior to 24 months (six or 15 months) was not found to mediate the association between maternal depression and child social competence and display of problem behaviors. Nevertheless, all significant mediations detected were modest in magnitude explaining only a small per centage of the direct effect of maternal depression indicating that maternal sensit ivity is not the primary causal mechanism through which maternal depression poses risk to child social competence and display of problem behaviors. This study also found moderate support of a buffering effect of high maternal sensitivity in the context of maternal depression, but only at earlier time points (six months). For example, evidence of a modera ting effect of maternal sensitivity at six months was detected on the relationship be tween depression severi ty at six months,

PAGE 428

406 among children not previously exposed to matern al depression, and teac her report of peer competence using a Bonferroni-adjusted p value of <.01 to judge sta tistical significance. Depression severity at six months had a st ronger negative impact on teacher report of peer competence among children with mothers rated lower in maternal sensitivity. Maternal sensitivity at six months was also found to buffer the association between remitted depression at six months (or depres sion present at one month postpartum and not present at six months postpartum) and matern al report of cooperation using a Bonferroni adjusted p value of <.01. Among episodes of depression initiati ng at later time points (15 months), maternal sensitivity was found to be a significant moderator, with higher levels of maternal sensitivity potentiating risk rather than buffering risk. One explanation posited for this finding is that perhaps childre n with less sensitive mothers have learned to compensate for lower levels of maternal responsiveness or have lower expectancies to be met by the mother by 15 months postpartu m, as compared to children with more sensitive mothers who may be confused at the onset of matern al depression. Specific public health implications re lated to the mediating and moderating influences of maternal sensitivity detected in this study include increased assessment of maternal sensitivity behaviors of depressed women by health care providers during routine office visits, as well as education effort s targeting parents, in particular depressed mothers, to promote sensitive and responsive caregiving. Healthcare providers should be trained and encouraged to identify probl ematic maternal behaviors during motherchild interaction such as intrusiveness, withdrawn interactions, or inconsistency. Caregivers identified as displaying insensit ive parenting behaviors may benefit from counseling to identify root causes of the inse nsitive behavior such as a problematic or

PAGE 429

407 abusive relationship with the caregiver’s own parent during childhood. Depressed mothers exhibiting lower levels of maternal sensitivity, as well as those not exhibiting lower levels of maternal sensitivity, may also benefit from instruction promoting motherchild synchrony, maternal skills to soothe and comfort the child, as well as the importance of eye contact and positive vocali zations with the child, given that higher levels of maternal sensitivity were detected in this study to buffer the negative impact of child exposure to maternal de pression (at six months) on later social development. In general, efforts to boost maternal sensitivity during the early postpartum period may offset some of the negative impact of maternal depression on later child social development. Lastly, ever exposure to nonmaternal care by 24 months was found to buffer the negative impact of chronic depression at 24 mo nths on both mother and teacher reports of social competence. For children who had ever been exposed to nonmaternal care, levels of social competence increased for childre n with chronically depressed moms, as compared to children with non-depressed moms. For children who had only been exposed to maternal care, levels of social competence dramatically decreased for children with chronically depressed moms, as compared to children with non-depressed moms. Given that the literature on child exposure to nonmaternal care is much debated and controversial, it is difficult to draw di rect conclusions about the public health implications of these findings. Future rese arch is needed to further understand the buffering effect of ever exposure to nonm aternal care in the context of chronic depression. For example, is the protective e ffect of exposure to nonmaternal care related to providing depressed mothers retreat from parenting responsibilities, exposing the child

PAGE 430

408 to alternative caregivers wher e additional attachment rela tionships and internalized expectations for social intera ctions may form, and/or exposing the child to a more active social environment? Preliminary post-hoc analyses to better unde rstand the mechanisms through which a protective moderating effect was exerted revealed that among mothers who were chronically depressed at 24 mont hs, mothers who had exposed their child to nonmaternal care displayed higher levels of maternal sensitivity (mean=3.43, SD=0.37), as compared to chronically depressed moth ers who had not exposed to their child to nonmaternal care (mean=3.17, SD=0.71). Further research is warranted however to more fully explicate the causal pathways through which exposure to nonmaternal care buffers the negative impact of child exposure to chronic maternal depression. Suggestions for Future Research The fact that ever exposure to maternal depression had no impact on any of the teacher reported aspects of child social comp etence implicates a possible reporting bias of mothers. Although this study purposefully excl uded participants with mothers depressed at first grade to reduce maternal reporting bi as in the context of concurrent depression, results from this study indicate a possible reporting bias of mothers who had ever experienced maternal depression during th e first three postpartum years. Although several studies have documented a maternal reporting bias of child problem behaviors in the context of concurrent depression (Fergusson, Lynskey & Horwood, 1993; LyonsRuth, Easterbrooks & Davidson Cibelli ,1997; Briggs-Gowan, Carter & Schwab-Stone, 1996), and in the context of previous psyc hiatric illness (Chilcoat & Breslau, 1997], no studies have examined the potential reporting bias of mother-rated social competence in the context or concurrent or previous maternal psychopa thology. Further research

PAGE 431

409 investigating a maternal reporting bias of child behaviors, including both social competence and display of problem behaviors, in the context of ever exposure to maternal depression; is needed given that most studies invest igating the impact of maternal depression on long-term child deve lopment rely on maternal rating of child behavior. Further research is needed to more fully understand the mediating and moderating effects of maternal sensitiv ity, and the moderating effects of exposure to nonmaternal care detected in this study. Despite the fact that maternal sensitivity explained a significant portion of the direct effect of maternal depression on child social competence and display of problem behaviors, the magnit ude of mediation was small, necessitating future research endeavors to illuminate th e primary mechanisms through which maternal depression poses risk to child social development. In addition, although positive moderating effects of maternal sensitivity were detected on the associations between depression episodes initiating at six months (high matern al sensitivity buffered the negative impact of maternal depression) on ch ild social competence, higher levels of maternal sensitivity at 15 months were detected to increase risk to development rather than offset the negative impact. Understand ing why this occurred w ould contribute to the knowledge base and help shape child promoti on activities in the c ontext of maternal depression. In addition, more research is needed to further u nderstand the buffering effect of ever exposure to nonmaternal care in the cont ext of chronic depression to illuminate the underlying reasons why this trend was detected.

PAGE 432

410 References Achenbach, T. (1966). The classification of children's psychiatric symptoms: A factoranalytic study. Psychological Monographs, 80 (7), (No. 615). Achenbach, T. (1979). Child Behavior Checklist Bethesda, MD: National Institute of Mental Health. Achenbach, T., McConaughy, SH., & Howell, CT. (1987). Child/adolescent behavioral and emotional problems: Implications of cross-informant correlations for situational specificity. Psychological Bulletin, 101, 212-232. Achenbach, T. (1991). Manual for the Child Behavior Checklist/4-18 and 1991 Profile Burlington, VT: University of Verm ont, Department of Psychiatry. Achenbach, T. (1994). Child Behavior Checkli st and Related Instruments. In M. E. Maruish (Ed.), The Use of Psychological Testing for Treatment Planning and Outcome Assessment (pp. 517-549). Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers. Achenbach, T. (1999). The Child Behavior Chec klist and Related Instruments. In M. E. Maruish (Ed.), The Use of Psychological Testing for Treatment Planning and Outcomes Assessment (2nd ed., pp. 429-466). Mahwah, NJ: Lawrence Erlbaum Associates. Achenbach, T., & Edelbrock, C. (1983). Manual for the child be havior checklist/4-18 and revised child behavior profile Burlington, VT: University of Vermont, Department of Psychiatry. Achenbach, T., Edelbrock, C., & Howell, C. (1987). Empirically based assessment of behavioral/emotional problems of 2and 3-year old children. Journal of Abnormal Child Psychology, 15 629-650. Achenbach, T., & Lewis, M. (1971). A proposed model for clinical research and its application to encopr esis and enuresis. Journal of the American Academy of Child Psychiatry, 10 535-545. Achenbach, T., & Rescorla, L. (2001). Manual for the ASEBA School-Age Forms and Profiles Burlington, VT: University of Vermont, Research Center for Children, Youth and Families. Ainsworth, M. (1963). The development of infant-mother interaction among the Ganda. In B. M. Foss (Ed.), Determinants of Infant Behavior (Vol. 67-104). New York: Wiley.

PAGE 433

411 Ainsworth, M. (1964a). Patterns of attach ment behavior shown by the infant in interaction with his mother. Merrill-Palmer Quarterly, 10 51-58. Ainsworth, M. (1964b). Patterns of attachme nt behaviors shown by an infant in interaction with his mother. Merrill-Palmer Quarterly, 10 51-58. Ainsworth, M. (1969). Maternal Care Scales: system for rating maternal-care behavior Princeton, NJ: Educational Testing Service. Ainsworth, M., & Bell, J. (1970). Attachment, exploration, and separa tion: Illustrated by the behavior of one-year-old s in a strange situation. Child Development, 41 49-67. Ainsworth, M., & Bell, S. M. (1974). Mother-i nfant interaction and the development of competence. In K. Connolly & J. Bruner (Eds.), The Growth of Competence New York, NY: Academic Press. Ainsworth, M., Bell, S. M., & Stayton, D. J. (1971). Individual di fferences in Strange Situation behavior of one year olds. In H. R. Schaffer (Ed.), The origins of human social relations New York: Academic Press. Ainsworth, M., Bell, S. M., & Stayton, D. J. (1974). Infant-mother attachment and social development. In M. P. Richards (Ed.), The introduction of the child into a social world (pp. 99-135). London: Cambridge University Press. Ainsworth, M., Blehar, M. C., Waters, E., & Wall, S. (1978a). Procedures, Patterns of Attachment: A Psychological St udy of the Strange Situation (pp. 31-44). Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers. Ainsworth, M., Blehar, M. C., Waters, E., & Wall, S. (1978b). Relationships Between Infant Behavior in the Strange Situat ion and Maternal Behavior at Home, Patterns of Attachment: A Psychological Study of the Strange Situation (pp. 137-153). Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers. Ainsworth, M., Blehar, M. C., Waters, E., & Wall, S. (1978c). Theoretical Background, Patterns of Attachment: A Psychological Study of the Strange Situation (pp. 328). Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers. Ainsworth, M., & Marvin, R. S. (1995). On the shaping of attachment theory and research: An interview with Mary D.S. Ainsworth. Monographs of the Society for Research in Child Development, 60 (2-3), 3-21. Ainsworth, M., & Wittig, B. A. (1969). Attach ment and the exploratory behaviour of one-year-olds in a st range situation. In B. M. Foss (Ed.), Determinants of Infant Behavior (Vol. 4, pp. 113-136). London: Methuen.

PAGE 434

412 American Psychiatric Association. (1968). Diagnostic and Statistical Manual of Mental Disorders ( 2nd ed.). Washington, DC. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders ( 4th ed.). Washington, DC: Amer ican Psychiatric Association. Appelbaum, M. (1993). Child Care Data Report 1: Recruitment, Catchment, and Sample ( 1). Nashville, TN: Quantitative Systems Laboratory, Vanderbilt University. Appelbaum, M., Batten, D. A., & Wendell, C. (1994). Child Care Data Report 35: Quality of Maternal Care Rating of Mothers in Structured Interaction Situation at 15 Month Home Visit ( 35). Nashville, TN: Quantitative Systems Laboratory, Vanderbilt University. Asher, S., & Taylor, A. (1981). The social outcomes of mainstreaming: Sociometric assessment and beyond. Exceptional Children Quarterly, 1 13-30. Attia, E., Downey, J., & Oberman, M. (1997). Postpartum Psychoses. In L. J. Miller (Ed.), Postpartum Mood Disorders (pp. 99-117). Washington DC: American Psychiatric Press, Inc. Bailey, D. (2000). Foreword. In D. Cryer & T. Harms (Eds.), Infants and Toddlers in Out-of-Home Care (pp. xv-xvi). Baltimore: Paul H Brookes Publishing Company. Barnas, M. V., & Cummings, E. M. (1997). Ca regiver stability and toddlers' attachmentrelated behaviors towards caregivers in day care. Caregiver stability and toddlers' attachment-related behaviors to wards caregivers in day care, Infant Be17havior and Development 171-177. Baumann, B. L., Pelham, W. E., Lang, A. R., Jacob, R. G., & Blumenthal, J. D. (2004). The impact of maternal depressive sy mptomatology on ratings of children with ADHD and child confederates. Journal fo Emotional and Behavioral Disorders, 12 (2), 90-98. Beck, A., Ward, C., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An inventory for measuring depression. Archives of General Psychiatry, 4 561-571. Beck, C. (1992). The lived experience of postpartum depression: A phenomenological study. Nursing Research, 41 (3), 166-170. Beck, C. T. (1996). Postpartum depressed mo thers' experiences interacting with their children. Nursing Research, 45 (2), 96-104.

PAGE 435

413 Beck, C. T. (2002). Postpartum Depression: A Metasynthesis. Qualitative Health Research, 12 (4), 453 472. Bedford, A., & Foulds, G. (1978). Delusions-Symptoms-States Inventory of Anxiety and Depression Windsor, England: National F oundation for Educational Research. Belsky, J., Rosenberger, K., & Crnic, K. ( 1995). The origins of a ttachment security: Classical and contextual determinants. In S. Goldberg & R. Muir & J. Kerr (Eds.), Attachment Theory: Social, Devel opmental and Clinical Perspectives (pp. 153-184). London: Th e Analytic Press. Belsky, J., Rovine, M., & Fish, M. (1989). Th e developing family system. In M. Gunnar (Ed.), Minnesota Symposia on Child Psychology: Vol 22. Systems and Development (pp. 119-166). Hillsdale, NJ: Erlbaum. Belsky, J., Woodworth, S., & Crnic, K. (1996). Trouble in the second year: Three questions about family interaction. Child Development, 67 556-578. Berggren-Clive, K. (1988). Out of the dar kness and into the ligh t: Women's experiences with depression of childbirth. Canadian Journal of Community Mental Health, 17 103-120. Birch, S. H., & Ladd, G. W. (1997). The teacher-child relationship and children's early school adjustment. Journal of School Psychology, 35 61-79. Birtchnell, J. (1988). Depressi on and family relationships. British Journal of Psychiatry, 153 758-769. Birtchnell, J. (1993). Does Recollection of E xposure to Poor Maternal Care in Childhood Affect Later Ability to Relate? British Journal of Psychiatry, 162 335-344. Bland, S., Batten, D. A., Appelbaum, M., & Wendell, C. (1995). Child Care Data Report 78: Quality of Maternal Care Rating of Mother in a Structur ed Play Situation at 24 Month Home Visit ( 78). Nashville, TN: Quantitative Systems Laboratory, Vanderbilt University. Bland, S., Batten, D. A., Appelbaum, M., & Wendell, C. (1996). Child Care Data Report 136: Quality of Maternal Care and Child Behaviors with Mother Rating of Mother and Child in a Structured Play Situation at 36 Months ( 136). Nashville, TN: Quantitative System s Laboratory, Vanderbilt University. Blehar, M. C., Lieberman, A. F., & Ainswo rth, M. D. S. (1977). Early face-to-face interaction and its relation to later infant-mother attachment. Child Development, 48 182-194.

PAGE 436

414 Bowlby, J. (1951). Maternal care and mental health Geneva, Switzerland: World Health Organization Monograph Series. Bowlby, J. (1956). The growth of independence in the young child. Royal Society of Health Journal, 76 587-591. Bowlby, J. (1958). The nature of the child's tie to his mother. International Journal of Psycho-Analysis, 39 350-373. Bowlby, J. (1959). Separation anxiety. International Journal of Psycho-Analysis, 41 89113. Bowlby, J. (1960). Grief and mourni ng in infancy and early childhood. The Psychoanalytic Study of the Child, 15 9-52. Bowlby, J. (1969). Attachment ( 1 ed. Vol. 1). London: Hogarth Press. Bowlby, J. (1973). Separation: Anxiety and Anger ( 1 ed. Vol. 2). London: Hogarth Press. Bowlby, J. (1980). Loss: Sadness and Depression ( 1 ed. Vol. 3). London: Hogarth Press. Bowlby, J. (1988). A Secure Base: Parent-Child Attachment and Healthy Human Development New York: Basic Books, Inc., Publishers. Brennan, P. A., Hammen, C., Andersen, M. J ., & Bor, W. (2000). Chronicity, Severity, and Timing of Maternal Depressive Symptoms: Relationships with Child Outcomes at Age 5. Developmental Psychology, 36 (6), 759-766. Bretherton, I. (1993). Theoretical Contributi ons from Developmental Psychology. In P. G. Boss & W. J. Doherty & R. LaRossa & W. R. Schumm & S. K. Steinmetz (Eds.), Sourcebook of Family Theories and Methods: A Contextual Approach (pp. 275-301). New York: Plenum Press. Bretherton, I. (1995). The Origins of Attachment Theory. In S. Goldberg & R. Muir & J. Kerr (Eds.), Attachment Theory: Social, Developmental, and Clinical Perspectives (pp. 45-84). Hillsdale, NJ: The Analytic Press. Briggs-Gowan, MJ., Carter, A., & Schwab-S tone, M. (1996). Di screpancies among mother, child, and teacher reports: Examining the contributions of maternal depression and anxiety. Journal of Abnormal Child Psychology, 24 (6), 749765. Brockington, I. F. (1996). Motherhood and mental health Oxford: Oxford University Press.

PAGE 437

415 Brown, G., & Harris, T. (1978). Social Origins of Depression: A study of Psychiatric Disorder in Women New York: The Free Press. Burchinal, M., Peisner-Feinber g, E., Bryant, D., & Clifford, R. (2000). Children's social and cognitive development and child care quality: Testing for differential associations related to po verty, gender or ethnicity. Applied Developmental Sciences, 4 149-165. Caldarella, P., & Merrell, K. (1997). Common di mensions of social skills of children and adolescents: A taxonomy of positive behaviors. School Psychology Review, 26 265-279. Campbell, S., & Cohn, J. (1991). Prevalence a nd correlates of postpartum depression in first-time mothers. Journal of Abnormal Psychology, 100 594-599. Campbell, S., & Cohn, J. (1997). The Timing an d Chronicity of Postpartum Depression: Implications for Infant Development. In L. Murray & P. Cooper (Eds.), Postpartum Depression and Child Development (pp. 165197). New York: The Guilford Press. Campbell, S., Cohn, J., Flanagan, C., Poppe r, S., & Meyers, T. (1992). Course and correlates of postpartum depression during the transition to parenthood. Development and Psychopathology, 4 29-27. Campbell, S., Cohn, J., & Meyers, T. (1995). Depression in first-time mothers: motherinfant interaction and depression chronicity. Developmental Psychology, 31 349-357. Caplan, H., Cogill, S., Alexandra, H., Robson, K., & Katz, R. (1989). Maternal depression and the emotional development of the child. British Journal of Psychiatry, 154 818-822. Cappizzano, J., & Adams, G. (2000). The hours that children under five spend in child care: Variation across states. A re port assessing the new federalism. Washington, D.C.: The Urban Institute. Capsi, A., Moffitt, T., Newman, D., & Silva, P. (1996). Behavioral observations at age 3 years predict adult psychiatric disorders. Archives of General Psychiatry, 53 1033-1039. Carlton, M., & Winsler, A. (1999). School readiness: The need for a paradigm shift. School Psychology Review, 28 338-352.

PAGE 438

416 Carter, A., & Briggs-Gowan, M. J. (1993). The Infant-Toddler Soc ial and Emotional Assessment (ITSEA) New Haven, CT: Yale University Department of Psychology. Carter, A. S., Garrity-Rokous, E. F., Chazan-C ohen, R., Little, C., & Briggs-Gowan, M. J. (2001). Maternal Depression and Co morbidity: Predicting Early Parenting, Attachment Security, and Toddler Social-Emotional Problems and Competencies. American Academy of Child and Adolescent Psychiatry, 40 (1), 18-26. Cassidy, J. (1999). The Nature of the Child's Ties. In J. Cassidy & P. R. Shaver (Eds.), Handbook of Attachment: Theory, Re search, and Clinical Application (pp. 320). New York: The Guilford Press. Chilcoat, H., & Breslau, N. (1997). Does psyc hiatric history bias mo thers' reports? An application of a new analytic approach. Journal of the American Academy of Child and Adolescent Psychiatry, 36 (7), 971-979. Chokka, P. (2002). Postpartum Depression: Part 1. The Canadian Journal of Continuing Medical Education, 14 (4), 37-45. Cicchetti, D., & Toth, S. (1991). A devel opmental perspective on internalizing and externalizing disorders. In D. Cicchetti & S. Toth (Eds.), Internalizing and externalizing expressio ns of dysfunction (pp. 1-19). Hillsdale, NJ: Lawrence Erlbaum Associates. Civic, D., & Holt, V. L. (2000). Maternal Depressive Symptoms and Child Behavior Problems in a Nationally Representa tive Normal Birthweight Sample. Maternal and Child Health Journal, 4 (4), 215-221. Clark, R., Gresham, F. M., & Elliott, S. N. (1985). Development and validation of a social skills assessment measure: The TROSS-C. Journal of Psychoeducational Assessment, 4 347-356. Clarke-Stewart, K. A., Vanderstoep, L., & Kill ian, G. (1979). Analysis and replication of mother-child relations at two-years of age. Child Development, 50 (3), 777-793. Coates, D., & Lewis, M. (1984). Early moth er-infant interaction and infant cognitive status as predictors of school performa nce and cognitive behavior in six-yearolds. Child Development, 55 (4), 1219-1230.

PAGE 439

417 Collins, M. E., & Lundgren-Gaveras, L. (1999) Potential Effects of the 1996 Welfare Reform Legislation on the Economic St atus, Health, and Well-being of Lowincome Families. In H. M. Wallace & G. Green & K. Jaros, J. & L. L. Paine & M. Story (Eds.), Health and Welfare for Families in the 21st Century (pp. 153166). Boston: Jones and Bartlett Publlishers. Colllins, N. L., Dunkel-Schetter, C., Lobel, M., & Scrimshaw, S. C. (1993). Social Support in Pregnancy: Psychosocial Correlates of Birth Outcomes and Postpartum Depression. Journal of Personality and Social Psychology, 65 (6), 1243-1258. Conners, C. (1973). Rating scales for use in drug studies with children, Psychopharmacology Bulletin: Pharmacotherapy with children Washington DC: US Government Printing Office. Cooper, P., Campbell, E., Day, A., Kennerley, H., & Bond, A. (1988). Non-Psychotic Psychiatric Disorder After Childbirth: A Prospective Study of Prevalence, Incidence, Course and Nature. British Journal of Psychiatry, 152 799-806. Cooper, P., & Murray, L. (1995). Course and recurrence of pos tnatal depression: Evidence for the specificity of the diagnostic concept. British Journal of Psychiatry, 166 191-195. Cowan, P., Cowan, C., Schulz, M., & Heming, G. (1994). Prebirth to preschool family factors in children's adaptation to kindergarten. In R. Parke (Ed.), Exploring family relationships with other social contexts: Advances in family research (Vol. 4, pp. 75-114). Hillsdale, NJ: Erlbaum. Cowen, E., Pederson, A., Babigan, H., Izzo, L ., & Trost, M. (1973). Long-term follow-up of early detected vulnerable children. Journal of Consulting and Clinical Psychology, 41 (438-446). Cox, A., Puckering, C., Pounds, A., & Mills, M. (1987). The impact of maternal depression on young children. Journal of Child Psychology and Psychiatry, 28 917-928. Cox, J., Holden, J., & Sagovsky, R. (1987). Detection of Postnatal Depression: Development of the 10-item Edinbur gh Postnatal Depression Scale. British Journal of Psychiatry, 150 782-786. Crockenberg, S. C., & Leerkes, E. M. (2003). Parental Acceptance, Postpartum Depression, and Maternal Sensitivity: Mediating and Moderating Processes. Journal of Family Psychology, 17 (1), 80-93. Cronbach, L. (1951). Coefficient alpha a nd the internal structure of tests. Psychometrika, 16 297-334.

PAGE 440

418 Cummings, E., & Cicchetti, D. (1990). Attachment, depression, and the transmission of depression. In M. Greenberg & D. Ci cchetti & E. M. Cummings (Eds.), Attachment in the preschool years: Theory, research and intervention (pp. 339-375). Chicago: University of Chicago Press. Cutrona, C. (1984). Social support and stre ss in the transition to parenthood. Journal of Abnormal Psychology, 93 378-390. Cutrona, C., & Troutman, B. (1987). Social supp ort, infant temperam ent, and parenting self-efficacy: A mediational model of postpartum depression. Child Development, 57 1506-1514. Dawson, G., Ashman, S. B., Panagiotides, H., Hessl, D., Self, J., Yamada, E., & Embry, L. (2003). Preschool Outcomes of Child ren of Depressed Mothers: Role of Maternal Behavior, Contextual Ri sk, and Children's Brain Activity. Child Development, 74 (4), 1158-1175. Depression Guideline Panel. (1993). Depression in Primary Care: Detection, Diagnosis, and Treatment. Quick Reference Guide for Clinicians. No. 5. ( AHCPR Publication No. 93-0552). Rockville, MD : US Department of Health and Human Services, Public Health Service, Agency for Health Care Policy and Research. Derogatis, L. R., Lipman, R. S., & Covi, L. (1973). SCL-90: An outpatient psychiatric scale: Preliminary Report. Psychopharmacology Bulletin, 9 13-27. Derogatis, L. R., & Lynn, L. L. (1999). Psychol ogical Tests in Scr eening for Psychiatric Disorder. In M. E. Maruish (Ed.), The Use of Psychological Testing for Treatment Planning and Outcomes Assessment (2nd ed., pp. 41-80). Mahwah, NJ: Lawrence Erlbaum Associates. Desai, S., Chase-Lansdale, P., Michael, R ., & 1989. (1989). Mother or market? Effects of maternal employment on the intellect ual ability of 4-year-old children. Demography, 26 (545-561). DeWolff, M. S., & van Ijzendoorn, M. H. ( 1997). Sensitivity and Attachment: A MetaAnalysis on Parental Antecedents of Infant Attachment. Child Development, 68 (4), 571-591. Donovan, W. L., Leavitt, L. A., & Walsh, R. O. (1998). Conflict and Depression Predict Maternal Sensitivity to Infant Cries. Infant Behavior and Development, 21 (3), 505-517.

PAGE 441

419 Edhborg, M., Lundh, W., Seimyr, L., & Widstrom A.-M. (2001). The long-term impact of postnatal depressed mood on mother-c hild interaction: a preliminary study. Journal of Reproductive and Infant Psychology, 19 (1), 61 71. Egeland, B., & Farber, E. A. (1984). Infant-M other Attachment: Factors Related to Its Development and Changes over Time. Child Development, 55 753-771. Endicott, J., & Spritzer, R. (1978). A diagnostic interview: The Schedule for Affective Disorders and Schizophrenia. Archives of General Psychiatry, 35 837-844. Erickson, M., Sroufe, A., & Egeland, B. ( 1985). The relationship between quality of attachment and behavior problems in preschool in a high-risk sample. Monographs of the Society for Rese arch in Child Development, 50 (1-2), 146166. Essex MJ, Klein MH, Miech R, & NA, S. (2001). Timing of initial exposure to maternal major depression and children's mental health symptoms in kindergarten. British Journal of Psychiatry, 179 151-156. Feldman, R. S. (1999). An Intr oduction to Child Development, Child Development: A Topical Approach (pp. 3-31). Upper Saddle River, NJ: Prentice Hall. Fellmeth, R. C. (1999). Child Care. In H. M. Wallace & G. Green & K. J. Jaros & L. L. Paine & M. Story (Eds.), Health and Welfare for Families in the 21st Century (Vol. 167-187). Boston: Jones and Bartlett Publishers. Fergusson, D., Lynskey, M., & Horwood, L. (1993) The effect of maternal depression on maternal ratings of child behavior. Journal of Abnormal Child Psychology, 21 245-269. Field, T., Healy, B., Goldstein, S., & Guther tz, M. (1990). Behavior-state matching and synchrony in mother-infant interactions in non-depressed versus depressed dyads. Developmental Psychology, 26 7-14. Field, T., Lang, C., & Martinex, A. (1996). Pres chool Follow-Up of Infants of Dysphoric Mothers. Journal of Clinical Child Psychology, 25 (3), 272-279. Fish, M., Stifter, C. A., & Belsky, J. ( 1993). Early patterns of mother-infant dyadic interaction: Infant, mother, and family demographic antecedents. Infant Behavior and Development (16), 1-18. Frankel, K., & Harmon, R. J. (1996). Depre ssed Mothers: They Don't Always Look as Bad as They Feel. Journal of the American Acad emy of Child and Adolescent Psychiatry, 35 (3), 289-298.

PAGE 442

420 Gagnon, C., Craig, W., Tremblay, R., Zhou, R., & Vitaro, F. (1995). Kindergarten predictors of boys' stable behavior problems at the end of elementary school. Journal of Abnormal Child Psychology, 23 751-766. Gard, P., Handley, S., Parsons, A., & Waldron, G. (1986). A multivariate investigation of postpartum mood disturbance. British Journal of Pharmacology, 70 102-103. Garrison, W., & Earls, F. (1986). Epidemiologi c perspectives on maternal depression and the young child. In T. Field & E. Tronick (Eds.), Maternal Depression and Child Development (pp. 13-30). San Fran cisco: Jossey-Bass. Gerweitz, J. L., & Boyd, E. F. (1977). Does maternal responding imply reduced infant crying? A critique of the 1972 Bell and Ainsworth report. Child Development, 48 1200-1207. Gitlin, M., & Pasnau, R. (1989). Psychiatric sy ndromes linked to reproductive function in women: A review of current knowledge. American Journal of Psychiatry, 146 (11), 1413-1422. Goldberg, D., Cooper, B., Eastwood, M., Kedward, H., & Shepherd, M. (1970). A standardised psychiatri c interview for use in community surveys. British Journal of Preventive Social Medicine, 24 18-23. Goldsmith, H. H., & Alansky, J. A. (1987). Mate rnal and infant predictors of attachment: A meta-analytic review. Journal of Consulting and Clinical Psychology, 55 805-816. Goodman, S., & Brumley, H. (1990). Schizophr enic and depressed mothers: Relational deficits in parenting. Developmental Psychology, 26 32-39. Goodman, S. H., & Gotlib, I. H. (1999). Ri sk for Psychopathology in the Children of Depressed Mothers: A Developmental Model for Understanding Mechanisms of Transmission. Psychological Review, 106 (3), 458-490. Gotlib, I., Whiffen, V., Mount, J., Milne, K ., & Cordy, N. (1989). Prevalence rates and demographic characteristics associated with depression in pregnancy and the postpartum. Journal of Consulting and Clinical Psychology, 57 269-274. Gotlib, I., Whiffen, V. E., Wallace, P., & Mount J. (1991). Prospective investigation of postpartum depression: Factors i nvolved in onset and recovery. Journal of Abnormal Psychology, 100 122-132. Gowen, J. W., & Nebrig, J. E. (1997). Infa nt-Mother Attachment at Risk: How early intervention can help. Infant and Young Children, 9 (4), 62 78.

PAGE 443

421 Green, M., & Palfrey, J. (Eds.). (2002). Bright Futures: Guidelines for Health Supervision of Infants, Children, and Adolescents (2nd ed.). Arlington, VA: National Center for Education in Maternal and Child Health. Gregoire, A., Kumar, R., & Everitt, B. ( 1996). Transdermal oestrogen for treatment of severe postnatal depression. Lancet, 6 (347), 930-933. Gresham, F. M., & Elliott, S. N. (1990). The Social Skills Rating System Manual Circle Pines, MN: American Guidance Service. Haapasalo, J., & Tremblay, R. (1994). Physica lly aggressive boys from ages 6 to 12: Family background, parenting behavi or, and prediction of delinquency. Journal of Consulting and Clinical Psychology, 62 1044-1052. Hamilton, M. (1960). A rati ng scale for depression. Journal of Neurologic Neurosurgical Psychiatry, 23 56-62. Hammen, C., & Brennan, P. A. (2003). Seve rity, Chronicity, and Timing of Maternal Depression and Risk for Adolescent Of fspring Diagnosis in a Community Sample. Archives of General Psychiatry, 60 (3), 253-258. Handler, A. (2004). The Effect of Welfare Reform on Low-Income Pregnant Women's Insurance Status and Prenatal Care Utilization Rockville, MD: U.S. Department of Health and Human Se rvices, Maternal and Child Health Bureau. Handler, A., Zimbeck, M., Chavkin, W., & Ad ams, K. (2003). The Effects of Welfare Reform on the Reproductive Health of Women. In H. M. Wallace & G. Green & K. J. Jaros & L. L. Paine & M. Story (Eds.), Health and Welfare for Families in the 21st Century Sudbury, Massachusetts: Bartlett Publishers. Harnish, J., Dodge, K., & Valente, E. (1995). Mother-child inte raction quality as a partial mediator of the roles of matern al depressive symptomatology and socioeconomic status in the develo pment of child behavior problems. Child Development, 66 (3), 739-753. Harris, B., Johns, S., Fung, H., Thomas, R., Walker, R., Read, G., & Riad-Fahmy, D. (1989). The hormonal environmen t of post-natal depression. British Journal of Psychiatry, 154 660-667. Harris, B., Lovett, L., Newcombe, R., R ead, G., R, W., & Riad-Fahmy, D. (1994). Maternity blues and major endocrine changes: Cardiff puerperal mood and hormone study II. British Medical Journal, 308 (6934), 949-953.

PAGE 444

422 Harter, S. (1985). Manual for the self-perception profile for children. Denver, CO: University of Denver. Hay, D. F. (1997). Postpartum Depression a nd Cognitive Development. In L. Murray & P. Cooper (Eds.), Postpartum Depression and Child Development (pp. 85110). New York: The Guilford Press. Hay, D. F., & Pawlby, S. (2003). Prosocial De velopment in Relation to Children's and Mothers' Psychological Problems. Child Development, 74 (5), 1314-1327. Heidrich, A., Schleyer, M., Spi ngler, H., Albert, P., Knoche, M ., Fritze, J., & Lanczik, M. (1994). Postpartum blues: relationshi p between not-protein bound steroid hormones in plasma and postpartum mood changes. Journal of Affective Disorders, 30 (2), 93-98. Henderson, A., Gregoire, A., Kumar, R., & Studd, J. (1991) Treatment of severe postnatal depression with oe stradiol skin patches. Lancet, 38 816-817. Hendrick, V., & Altshuler, L. L. (1999). Biological Determinants of Postpartum Depression. In L. J. Miller (Ed.), Postpartum Mood Disorders (pp. 65-82). Washington, DC: American Psychiatric Press, Inc. Heneghan, A. M., Johnson Silver, E., Bauman, L. J., & Stein, R. E. (2000). Do Pediatricians Recognize Mothers with Depressive Symptoms? Pediatrics, 106 (6), 1367-1373. Hinde, R. A. (1989). Ethological and rela tionship approaches. In R. Vatsa (Ed.), Annals of child development (Vol. 6, pp. 251-285). Greenwich, CT: JAI Press. Hipwell, A., Goossens, F. A., Melhuish, E., & Kumar, R. (2000). Severe maternal psychopathology and infant-mother attachment. development an psychopathology, 12 157-175. Hock, E., & De Meis, D. (1990). Depression in mothers of infants: The role of maternal employment. Developmental Psychology, 26 285-291. Hogan, A., Scott, K., & Bauer, C. (1992). The Adaptive Social Behavior Inventory (ASBI): A New assessment of social co mpetence in high risk three-year-olds. Journal of Psychoeducational Assessment, 10 230-239. Holden, J. (1991). Postnatal depression: Its nature, effects and identification using the Edinburgh Postnatal Depression Scale. Birth, 18 (4), 211-221. Holmes, J. (1993). Attachment, a nxiety, internal working models, John Bowlby and Attachment Theory (pp. 61-85). London: Routledge.

PAGE 445

423 Holmes, J. (1995). "Something There Is That Doesn't Love a Wall": John Bowlby, Attachment Theory, and Psychoanalysis. In S. Goldberg & R. Muir & J. Kerr (Eds.), Attachment Theory: Social, Developm ental, and Clinical Perspectives Hillsdale, NJ: The Analytic Press. Horowitz, J., & Goodman, H. (in press) The longitundal course of postpartum depression. Research adn Theory for Nursing Practice Howes, C. (1999). Attachment Relationship in the Context of Multiple Caregivers. In J. Cassidy & P. R. Shaver (Eds.), Handbook of Attachment: Theory, Research, and Clinical Applications (pp. 671-687). New York: The Guilford Press. Howes, C. (2000). Social Development, Famil y, and Attachment Relationships of Infants and Toddlers. In D. Cryer & T. Harms (Eds.), Infants and Toddlers in Out-ofHome Care (pp. 87-113). Baltimore: Paul H Brookes Publishing Company. Howes, C., & Ritchie, S. (1998). Changes in child-teacher relationshi ps in a therapeutic preschool program. Early Education and Development, 4 411-422. Howes, C., & Segal, J. (1993). Children's rela tionships with alternat ive caregivers: The special case of maltreated child ren removed from their homes. Journal of Applied Developmental Psychology, 17 71-81. Hubbs-Tait, L., Pond Huges, K., McDonald Culp, A., Osofsky, J. D., Hann, D. M., Eberhart-Wright, A., & Ware, L. M. ( 1996). Children of Adolescent Mothers: Attachment Representation, Maternal Depression, and Later Behavior Problems. American Journal of Orthopsychiatry, 66 (3), 416-426. Ilfeld, F. (1976). PsychologicaFurther valida tion of a psychiatric symptom index in a normal populationl Reports. Psychological Reports, 39 1215-1228. Jacobsen, T. (1999). Effects of postpartum di sorders on parenting and on offspring. In L. J. Miller (Ed.), Postpartum Mood Disorders (Vol. 119-139). Washington DC: American Psychiatric Press, Inc. James, C., Appelbaum, M., Batte n, D. A., & Wendell, C. (1993). Child Care Data Report 8: Quality of Maternal Care Rating of Mothers in Free Play Situation ( 8). Nashville, TN: Quantitative System s Laboratory, Vanderbilt University. Jameson, P. B., Gelfand, D. M., Kulcsar, E., & Teti, D. M. (1997). Mother-toddler interaction patterns associated with maternal depression. Development and Psychopathology, 9 537-550.

PAGE 446

424 Karl, D. (1995). Maternal Responsiveness of Socially High-Risk Mothers to the Elicitation Cues of Their 7-Month-Old Infants. Journal of Pediatric Nursing, 10 (4), 254 263. Karr-Morse, R., & Wiley, M. (1997). Ghosts from the Nursery: Tracing the Roots of Violence New York: Atlantic Monthly Press. Kennerly, R., & Gath, D. (1989). Maternity blues I: Detection and measurement by questionnaire. British Journal of Psychiatry, 155 356-362. Klaus, M. H., Kennell, J. H., & Klaus, P. H. (1995). Developing Ties: The First Days and Weeks, Bonding: Building the Foundations of Secure Attachment and Independence (pp. 103-127). Rading, MA: Addison-Wesley Publishing Company, Inc. Kopp, C. (1992). Emotional distress and control in young children. New Directions in Child Development, 55 41-56. Koren-Karie, N., Oppenheim, D., Dolev, S., Sher, E., & Etzion-Carasso, A. (2002). Mothers' Insightfulness Regarding Their Infants' Internal Experience: Relations with Maternal Sensitivity and Infant Attachment. Developmental Psychology, 38 (4), 534-542. Kumar, R., & Robson, K. (1984). A prosp ective study of emoti onal disorders in childbearing women. British Journal of Psychiatry, 144 35-47. Kurstjens, S., & Wolke, D. (2001). Eff ects of Maternal Depression on Cognitive Development of Children over the First 7 years of Life. Journal of Child Psychology and Psychiatry, 42 (5), 623 636. Ladd, G. W., Birch, S. H., & Buhs, E. S. (1999) Children's social and scholastic lives in kindergarten: Related spheres of influence? Child Development, 70 13731400. Lamb, M. E., Thompson, R. A., Gardne r, W., & Charnov, E. L. (1985a). The Development of Attachment Theory, Infant-Mother Attachment: The Origins and Developmental Significance of Indivi dual Differences in Strange Situation Behavior (pp. 8-26). Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers. Lamb, M. E., Thompson, R. A., Gardner, W., & Charnov, E. L. (1985b). Parent-Infant Interaction adn Subsequent Strange Situation Behavior, Infant-Mother Attachment: The Origins and Developme ntal Significance of Individual Differences in Strange Situation Behavior (pp. 59-97). Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers.

PAGE 447

425 Leathers, S., Kelley, M., & Richman, J. ( 1997). Postpartum depressive symptomatology in new mothers and fathers: Parenting, work, and support. Journal of Nervous and Mental Disease, 185 (3), 129-139. LeCuyer-Maus, E. A. (2000). Maternal Sensitivity and Responsiveness Limit-Setting Style, and Relationship History in the Transition to Toddlerhood. Comprehensive Pediatric Nursing, 23 117 139. Leopold, K., & Zoschnick, L. (Eds.). (1996). Postpartum Depression Lerman, R. I., & Ratcliffe, C. (2001). Are si ngle mothers finding j obs without displacing other workers? Monthly Labor Review Online, 124 (7), 3-12. Levy, T. M., & Orlans, M. (1998). Attachment, Trauma, and Healing Washington, DC: CWLA Press. Lieberman, A. F. (1977). Preschooler's competen ce with a peer: relati ons with attachment and peer experience. Child Development, 48 (1), 277-287. Lieberman, A. F., & Zeanah, C. (1995). Disorders of Attachment in Infancy. Infant Psychiatry, 4 (3), 571-587. Livingood, A. B., & Daen, P. (1983). The Depre ssed Mother as a Source of Stimulation for Her Infant. Journal of Clinical Psychology, 39 (3), 369-375. Locke, L. F., Wyric Spirduso, W., & Silverman, S. J. (2000). Proposals that Work: A Guide for Planning Dissertations and Grant Proposals ( 4 ed.). Thousand Oaks: Sage Publications, Inc. Lorenz, K. (1952). King Solomon's Ring London: Crowell. Luoma, I., Tamminen, T., Kaukonen, P., Laippala, R., Puura, K., Salmelin, R., & Almqvist, F. (2001). Longitudinal Study of Maternal Depressive Symptoms and Child Well-Being. Journal of the Americ an Academy of Child and Adolescent Psychiatry, 40 (12), 1367-1374. Lyons-Ruth, K., Easterbrooks, M. A., & Cibe lli, C. D. (1997). Infant Attachment Strategies, Infant Mental Lag, and Maternal Depressi ve Symptoms: Predictors of Internalizing and Extern alizing Problems at Age 7. Developmental Psychology, 33 (4), 681-692. Lyons-Ruth, K., Zoll, D., Connell, D. B., & Grunebaum, H. U. (1986). The depressed mother and her one-year old infant: Envi ronment, interactions, attachment and infant development. In E. Tronick & T. Fields (Eds.), Maternal depression and infant disturbance (pp. 61-82). San Francisco: Jossey-Bass.

PAGE 448

426 Maccoby, E., & Martin, J. A. (1983). Socializat ion in the context of the family: Parentchild interaction. In P. Mussen (Ed.), Handbook of child psychology (Vol. 4, pp. 1-101). New York: Wiley. Main, M., Kaplan, N., & Cassidy, J. (1985) Security in infancy, childhood, and adulthood: A move to the level of repres entation. In I. Bretherton & E. Waters (Eds.), Growing points in attachment theory and research (pp. 66-104). Chicago, IL: University of Chicago Press for the Society for Research in Child Development. Marrone, M. (1998a). Attachment Theory, Attachment and Interaction (pp. 31-51). Philadelphia: Jessica Kingsley Publishers. Marrone, M. (1998b). On Representational Models, Attachment and Interaction (pp. 7180). London: Jessica Kingsley Publishers. Marvin, R. S. (1977). An ethological-cognitive model for the attenuation of mother-child attachment behavior. In T. M. Allowa y & L. Krames & P. Pliner (Eds.), Advances in the study of communication and affect: The development of social attachment (Vol. 3). New York: Plenum Press. Matas, L., Arend, R., & Sroufe, L. A. (1979). Continuity of adaptation in the second year: the relationship between quality of attachment and later competence. Child Development, 49 547-556. Mauthner, N. (1999). "Feeling low and feeli ng really bad about feeling low": Women's experiences of motherhood and postpartum depression. Canadian Psychology, 40 143-161. Mauthner, N. S. (1995). Postnatal Depre ssion: The Significance of Social Contacts Between Mothers. Women's Studies Inte rnational Forum, 18 (3), 311-323. McCarthy, D. (1972). Manual for the McCarthy scales of children's abilities New York: Psychological Corporation. McIntosh, J. (1993). Postpartum depressi on: women's help-seeking behaviour and perceptions of cause. Journal of Advanced Nursing, 18 178-184. Meins, E. (1997). The origins of Bowlby's attachment theory, Security of Attachment and the Social Development of Cognition East Sussex: Psychology Press. Meredith, P., & Noller, P. (2003). Attachme nt and Infant Difficulties in Postnatal Depression. Journal of Family Issues, 24 (5), 668-686.

PAGE 449

427 Merrell, K. W. (1999a). Assessment of Externalizing Problems, Behavioral, Social, and Emotional Assessment of Children and Adolescents (pp. 211-240). Mahwah, NJ: Lawrence Erlbaum Associates. Merrell, K. W. (1999b). Assessmen t of Internalizing Problems, Behavioral, Social, and Emotional Assessment of Children and Adolescents (pp. 241-274). Mahwah, NJ: Lawrence Erlbaum Associates. Merrell, K. W. (1999c). Assessment of Social Skills and Peer Relationships, Behavioral, Social, and Emotional Assessmen t of Children and Adolescents (pp. 308-336). Mahwah, NJ: Lawrence Erlbaum Associates, Publishers. Merrell, K. W. (1999d). Be havior Rating Scales, Behavioral, Social, and Emotional Assessment of Children and Adolescents (pp. 72-98). Mahway, NJ: Lawrence Erlbaum Associates, Publishers. Miller, L. J. (2002). Postpartum Depression. Journal of the American Medical Association, 287 (6), 762-765. Misri, S. (2002). In postpartum depression, early treatment is key. OBG Management, 14 (4), 50-66. Mueller, D. (1980). Social networks: A promising direction for research on the relationships of the social envi ronment to psychiatric disorder. Social Science and Medicine, 14a 147-161. Murata, A., Nodaoka, T., Morioka, Y., Oiji, A., & Saito, H. (1998). Prevalence and Background Factors of Maternity Blues. Gynecologic and Obstetric Investigation, 46 99-104. Murray, L. (1992). The Impact of Postna tal Depression on Infant Development. Journal of Child Psychology and Psychiatry, 33 (3), 543-561. Murray, L., & Cartwright, W. (1993). The role of obstetric risk factors in postpartum depression. Journal of Reproductive and Infant Psychology, 11 215-219. Murray, L., & Cooper, P. (1997). The role of infant and maternal factors in postpartum depression, mother-infant interactions, a nd infant outcomes. In L. Murray & P. Cooper (Eds.), Postpartum Depression and Child Development (pp. 111-135). New York: Guilford. Murray, L., Fiori-Cowley, A., & Hooper, R. (1 996). The impact of postnatal depression and associated adversity on early mothe r-infant interactions and later infant outcome. Child Development, 67 2512-2526.

PAGE 450

428 Murray, L., Hipwell, A., & Hooper, R. (1996). The cognitive development of 5-year-old children of postnatally depressed mothers. Journal of Child Psychology and Psychiatry, 37 (8), 927-935. Murray, L., Sinclair, D., Cooper, P., Du cournau, P., & Turner, P. (1999). The Socioemotional Development of 5-year-o ld Children of Postnatally Depressed Mothers. Journal of Child Psychology and Psychiatry, 40 (8), 1259-1271. National Mental Health Association. (1996). American Attitudes About Clinical Depression and Its Treatment Alexandria: National Mental Health Association. NICHD Early Child Care Research Network. (1991a). The NICHD Study of Early Child Care and Youth Development Phase I Manual: Administration of the One Month Home Visit Research Triangle Park, NC: National Institute of Child Health and Human Development. NICHD Early Child Care Research Network. (1991b). The NICHD Study of Early Child Care and Youth Development Phase I Manual: Procedure for Videotaping Mother-Infant Intera ction at Six Months Research Triangle Park, NC: National Institute of Child Health and Human Development. NICHD Early Child Care Research Network. (1991c). The NICHD Study of Early Child Care and Youth Development Phase I Manual: Recruitment Procedures Research Triangle Park, NC: National In stitute of Child Health and Human Development. NICHD Early Child Care Research Network. (1992). The NICHD Study of Early Child Care and Youth Development Phase I Manual: Procedures for Videotaping Mother-Child Interaction at 15 Months Research Triangle Park, NC: National Institute of Child Health and Human Development. NICHD Early Child Care Research Network. (1993). The NICHD Study of Early Child Care and Youth Development Phase I Manual: Coding Manual for the 36 Month Lab Research Triangle Park, NC: Nati onal Institute of Child Health and Human Development. NICHD Early Child Care Research Network. (1994). Child Care and Child Development: The NICHD Study of Early Child Care. In S. L. Friedman & H. C. Haywood (Eds.), Developmental Follow-up: Concepts, Domains, and Methods (pp. 377396). San Diego: Academic Press, Inc. NICHD Early Child Care Research Network. (1999a). Child Care and Mother-Child Interaction in the Fi rst 3 Years of Life. Developmental Psychology, 33 (6), 1399-1413.

PAGE 451

429 NICHD Early Child Care Research Network. ( 1999b). Chronicity of Maternal Depressive Symptoms, Maternal Sensitivity, and Child Functioning at 36 Months. Developmental Psychology, 35 (5), 1297-1310. NICHD Early Child Care Research Network. (2000). The Relation of Child Care to Cognitive and Language Development. Child Development, 71 (4), 960-980. NICHD Early Child Care Research Network. (2001). Child-Care and Family Predictors of Preschool Attachment and Stability From Infancy. Developmental Psychology, 37 (6), 847-862. NICHD Early Child Care Research Network. (2002). Early Child Care and Children's Development Prior to School Entry: Results from the NICHD Study of Early Child Care. American Educational Research Journal, 39 (1), 133-164. NICHD Early Child Care Research Network. (2003). Early child care and mother-child interaction from 36 months through first grade. Infant Behavior and Development, 26 345-370. Nott, P. (1982). Psychiatric illness following childbirth in Southhampt on: A case register study. Psychological Medicine, 12 557-561. Nott, P., Franklin, M., Armitage, C., & Geld er, M. (1976). Hormonal changes in mood in the puerperium. British Journal of Psychiatry, 151 523-527. O'Brien Caughy, M., DiPietro, J. A., & Strobino, D. M. (1994). Day-Care Participation as a Protective Factor in the Cognitivie Development of Low-Income Children. Child Development, 65 457-471. O'Hara, M. W. (1995a). Introduction, Postpartum depression: Causes and Consequences (pp. 1-27). New York: Springer-Verlag. O'Hara, M. W. (1995b). Summary and Implications, Postpartum Depression: Causes and Consequences (pp. 168-194). New York: Springer-Verlag. O'Hara, M. W. (1997). Introduction to Postpart um Depressive Disorders. In L. Murray & P. Cooper (Eds.), Postpartum Depression and Child Development (pp. 3-34). New York: The Guilford Press. O'Hara, M. W., Neunaber, D., & Zekoski, E. (1984). A prospectiv e study of postpartum depression: Prevalence, cour se, and predictive factors. Journal of Abnormal Psychology, 93 158-171.

PAGE 452

430 O'Hara, M. W., Schlechte, J., & Lewis, D. (1991). Prospective study of postpartum blues: Biologic and psychosocial factors. Archives of General Psychiatry, 48 801806. O'Hara, M. W., Schlechte, J., Lewis, D., & Varner, M. (1991). Controlled prospective study of postpartum mood disorders: psychological, environmental, and hormonal variables. Journal of Abnormal Psychology, 100 (1), 63-73. O'Hara, M. W., & Swain, A. (1996). Rates and Risk of Postpartum Depression: A MetaAnalysis. International Review of Psychiatry, 8 (37-54). O'Hara, M. W., Zekoski, E., Philipps, L., & Wright, E. (1990). A controlled prospective study of postpartum mood disorders: Comparison of childbearing and nonchildbearing women. Journal of Abnormal Psychology, 99 3-15. Olson, A. L., Kemper, K. J., Kelleher, K. J., Hammond, C. S., Zuckerman, B. S., & Dietrich, A. J. (2005). Primary Care Pediatricians' Roles and Perceived Responsibilities in the Id entification and Management of Maternal Depression. Pediatrics, 110 (6), 1169-1176. Olson, S., Bates, J., & Bayles, K. (1984). Mo ther-infant interaction and the development of individual differences in ch ildren's cognitive competence. Developmental Psychology, 201 (1), 166-179. Offord, D., Boyle, M., & Racine, Y. (1989). Ontario child health st udy: correlates of disorder. Journal of the American Acade my of Child and Adolescent Psychiatry, 28 856-860. Paykel, E., Emms, E., Fletcher, J., & Rassaby, E. (1980). Life events and social support in puerperal depression. British Journal of Psychiatry, 136 339-346. Pederson, D. R., Gleason, K. E., Moran, G., & Bento, S. (1998). Maternal Attachment Representations, Maternal Sensitivity, and the In fant-Mother Attachment Relationship. Developmental psychology, 34 (5), 925-933. Pederson, D. R., & Moran, G. (1995). A cate gorical description of infant-mother relationships in the home and its relation to Q-wort measures of infant-mother interaction. Monographs of the Society for Re search in Child Development, 60 (2-3), 111-145. Peduzzi, P., Concato, J., Kemper, E., Holford, T., & Feinstein, A. (1996). A stimulation study of the number of events per vari able in logistic regression analysis. Journal of Clinical Epidemiology, 49 (12), 1373-1379.

PAGE 453

431 Petterson, S. M., & Burke Albers, A. (2001). E ffects of Poverty and Maternal Depression on Early Child Development. Child Development, 72 (6), 1794-1813. Philipps, L., & O'Hara, M. W. (1991). Prospe ctive study of postpartum depression: 4 1/2year follow-up of women and children. Journal of Abnormal Psychology, 100 151-155. Phillipps, L. H., & O'Hara, M. W. (1991). Pr ospective Study of Postpartum Depression: 4 1/2-Year Follow-Up of Women and Children. Journal of Abnormal Psychology, 100 (2), 151-155. Pitt, B. (1968). "Atypical" depression following childbirth. British Journal of Psychiatry, 114 1325-1335. Popper, S. D., Ross, S., Cohn, J. F., & Campbell, S. B. (1993, March 1993). Social and Object Mastery Play in 12-Month Ol ds with Depressed and Non-Depressed Mothers: Developmental Changes and Correlates. Paper presented at the Society for Research in Child Development, New Orleans, LA. Pyszczynski, T., & Greenberg, J. (1987). De pression, self-focused attention, and selfregulatory preservation. In C. R. Snyder & C. E. Ford (Eds.), Coping with negative life events: C linical and social psychological perspectives (Vol. 105129). New York: Plenum Press. Quay, H., & Peterson, D. (1983). Interim Manual for the Re vised Behavior Problem Checklist Coral Gables, FL: University of Miami, Applied Social Sciences. Radke-Yarrow, M., Cummings, E ., Kuczynski, L., & Chapman, M. (1985). Patterns of attachment in twoand three-year-olds in normal families and families with parental depression. Child Development, 56 884-892. Radloff, L. S. (1977). The CES-D scale: a self -report depression scale for research in the general population. Applied Psychological Measurement, 1 385-401. Raikes, H. (1993). Relationship duration in infa nt care: Time with a high ability teacher and infant-teacher attachment. Early Childhood Research Quarterly, 8 309325. Raskin, A., Schulterbrandt, J., Reatin g, N., & Rice, C. (1967). Factors of psychopathology in interview, ward be havior, and self-re port ratings of hospitalized depressives. Journal of Consulting Psychology, 31 (3), 270-278. Research Triangle In stitute. (1999a). The NICHD Study of Early Child Care Enrollment Process: Technical Note 204 Research Triangle Park, NC: Research Triangle Institute.

PAGE 454

432 Research Triangle Institute. (1999b). NICHD Study of Early Child Care Phase I Instrument Document Research Triangle Park, NC: Research Triangle Institute. Resnick, M. B., Gomatam, S. V., Carter, R. L., Ariet, M., Roth, J., Kilgore, K. L., Bucciarelli, R. L., Mahan, C. S., Cu rran, J. S., & Eitzman, D. (1998). Educational Disabilities of Neona tal Intensive Care Graduates. Pediatrics, 102 (2), 308-314. Resnick, M. B., Gueorguieva, R. V., Cart er, R. L., Ariet, M., Sun, Y., Roth, J., Bucciarelli, R. L., Curran, J. S., & Ma han, C. S. (1999). The impact of low birth weight, perinatal conditions and sociodemographic factors on educational outcomes in kindergarten. Pediatrics, 104 (6), 74-84. Richman, N., & Graham, P. (1971). A behaviour al screening questionnaire for use with three-year old children: preliminary findings. Journal of Child Psychology and Psychiatry, 12 5-33. Riecher-Rossler, A., & Hofecker, F., M. ( 2003). Postpartum Depression: Do We Still Need This Diagnostic Term? Acta Psychiatrica Scandinavica, 108 (Suppl 418), 51-56. Robertson, J., & Bowlby, J. (1952). Responses of young children to separation from their mothers. Courrier Centre, Internationale Enfance, Paris, 2 131-140. Rockliff, B. (1971). A brief rating sc ale for anti-depressant drug trials. Comprehensive Psychiatry, 12 122-135. Roff, M. (1963). Childhood social in teractions and young adult psychosis. Journal of Clinical Psychology, 19 152-157. Roff, M., Sells, B., & Golden, M. (1972). Social adjustment and personality development in children Minneapolis: University of Minnesota Press. Roff, M., & Sells, S. (1968). Juvenile delinquen cy in relation to peer acceptance-rejection and sociometric status. Psychology in the Schools, 5 3-18. Ruddy, M., & Bornstein, M. (1982). Cognitive corre lates of infant attention and maternal stimulation over the fi rst year of life. Child Development, 53 (1), 183-188. Rutter, M., Tizard, J., & Whitmore, K. (1970). Parental psychiatric disorder: Effects on children. Psychological Medicine, 14 853-880. Scott, D. (1992). Early Iden tification of Maternal Depre ssion as a Strategy in the Prevention of Child Abuse. Child Abuse and Neglect, 16 345-358.

PAGE 455

433 Seguin, L., Potvin, L., St-Denis, M., & Loiselle J. (1999a). Depressi ve Symptoms in the Late Postpartum Among Low Socioeconomic Status Women. Birth, 26 (3), 157-163. Seguin, L., Potvin, L., St-Denis, M., & Loise lle, J. (1999b). Socio-Environmental Factors and Postnatal Depressive Symp tomatology: A Longitudinal Study. Women and Health, 29 (1), 57-72. Seidman, D. (1998). Postpartum Psychiatric Illness: The Role of the Pediatrician. Pediatrics in Review, 19 128-131. Shapiro, S., Skinner, E., Kessler, L., von Kro ff, M., German, P., Tischler, G., Leaf, P., Benham, L., Cottler, L., & Regier, D. ( 1984). Utilization of health and mental health services. Archives of General Psychiatry, 41 971-982. Shaw, D. S., & Vondra, J. I. (1995). Infant A ttachment Security and Maternal Predictors of Early Behavior Problems: A Longit udinal Study of Low-Income Families. Journal of Abnormal Child Psychology, 23 (3), 335-357. Smither, P. B., & Pederson, D. R. (1988). Mate rnal sensitivity and patterns of infantmother attachment. Child Development, 59 1097-1101. Snow, C. (1989). The developm ent of definitional skill. Journal of Child Language, 17 697-710. Spinelli, M. (1998). Antepart um and postpartum depression. Journal of Gender-Specific Medicine, 1 (2), 33-36. Spitzer, R., Endicott, J., & Robins, E. (1978) Research Diagnostic Criteria: Rationale and Reliability. Archives of General Psychiatry, 36 773-782. Spitzer, R., Williams, J., Gibbon, M., & First, M. (1989). Structured Clinical Interview for DSM-III-R (with psychotic screen) New York: New York State Psychiatric Institute. Spitzer, R., Williams, J., Gibbon, M., & First, M. (1992). The Structured Clinical Interview for DSM-III-R (SCID), I: History, Rationale, and Description. Archives of General Psychiatry, 49 624-629. Stein, J. A., Lu, M. C., & Gelbery, L. (2000). Severity of homelessness and adverse birth outcomes. Health Psychology, 19 (6), 524-534. Stephens, T. (1978). Social skills in the classroom Columbus, OH: Cedars Press.

PAGE 456

434 Stern, G., & Kruckman, L. (1983). Multi-Dis ciplinary Perspectives on Postpartum Depression: An Anthropological Critique. Social Science and Medicine, 17 (15), 1027-1041. Stipek, D., Feiler, R., Daniels, D., & Milburn, S. (1995). Effects of di fferent instructional approaches on young children's achievement and motivation. Child Development, 66 209-223. Storr, A. (1992). John Bowlby Munks Roll, pp. 49. Stover, A., & Marnejon, J. (1995). Postpartum Care. American Family Physician, 52 (5), 1465-1472. Stuart, S., Couser, G., Schilder, K., O'Hara M. W., & Gorman, L. (1998). Postpartum Anxiety and Depression: Onset and Co morbidity in a Community Sample. Journal of Nervous and Mental Disease, 186 (7), 420-424. Taylor, V. (1996). Introduction, Rock-A-By Baby (pp. 1-23). New York: Routledge. Terry, D., Mayocchi, L., & Hynes, G. ( 1996). Depressive symptomatology in new mothers: A stress and coping perspective. Journal of Abnormal Psychology, 105 220-231. Teti, D. M., Gelfand, D. M., Messinger, D. S., & Isabella, R. (1995). Maternal Depression and the Quality of Early Att achment: An Examination of Infants, Preschoolers, and Their Mothers. Developmental Psychology, 31 (3), 364-376. Thomas, A., Chess, S., & Korn, S. (1982). The Reality of Difficult Temperament. Merrill-Palmer Quarterly, 28 (1-20). Thompson, R. A. (1999). Early At tachment and Later Development. In J. Cassidy & P. R. Shaver (Eds.), Handbook of Attachment: Theory, Research, and Clinical Applications (pp. 265-286). New York: Guilford Press. Tremblay, R., Ragani-Kurtz, L., Masse, L., Vitaro, F., & Pihl, R. (1995). A bimodal preventive intervention for disruptive kindergarten boys: Its impact through mid-adolescence. Journal of Consulting and Clinical Psychology, 63 560-568. U.S. Bureau of the Census. (1992). Statistical abstraact of the United States Washington, D.C.: U.S. Government Printing Office. Ugarriza, D. (1995). A descriptive study of postpartum depression. Perpectives in Psychiatric Care, 31 (3), 25-29.

PAGE 457

435 Ugarriza, D. N., & Robinson, M. K. (1997). Assessment of Postpartum Depression: A Review of the Research. The Online Journal of Knowledge Synthesis for Nursing, 4 (6), 1-10. Vandell, D., Owen, M., Wilson, K., & Henders on, V. (1988). Social development in infant twins: Peer and mo ther-child relationships. Child Development, 59 168169. Vandell, D. L. (1979). Effects of playgr oup experience on mother-son and father-son interaction. Developmental Psychology, 15 379-385. Walker, H., & Hops, H. (1976). Increasing a cademic achievement by reinforcing direct academic performance and/or f acilitating nonacademic responses. Journal of Educational Psychology, 68 218-225. Walther, V. N. (1997). Postpart um Depression: A Review for Perinatal Social Workers. Fundamentals of Perinatal Social Work : A Guide for Clinical Practice, 24 (3/4), 99-111. Watson, J., Elliott, S., Rugg, A., & Brough, D. (1984). Psychiatric disorder in pregnancy and the first postnatal year. British Journal of Psychiatry, 144 453-462. Weinfield, S. N., Sroufe, L. A., Egeland, B., & Carlson, E. (1999). The Nature of Individual Differences in Infant-Caregiv er Attachment. In J. Cassidy & P. R. Shaver (Eds.), Handbook of Attachment: Theory, Research, and Clinical Applications New York: Guilford Press. Weiss, R. (1982). Attachment in adult life. In C. M. Parkes & J. Stevenson-Hinde (Eds.), The Place of Attachment in Human Behavior London: Routledge. Weissman, M., Sholomaskas, D., Pottenger, M., Prusoff, B., & Locke, B. (1977). Assessing depressive symptoms in five psychiatric populations: a validation study. American Journal of Epidemiology, 106 (3), 203-214. West, J., Denton, K., & Germino-Hausken, E. (2000). America's kindergarteners Washington, D.C.: National Cent er for Education Statistics. Whiffen, V. E. (1988). Vulnerability to postpartum depression: A prospective multivariate study. Journal of Abnormal Psychology, 97 467-474. Whiffen, V. E., & Gotlib, I. (1993). Co mparison of postpartum and nonpostpartum depression: Clinical presentation, ps ychiatric history, and psychosocial functioning. Journal of Consulting and Clinical Psychology, 61 485-494.

PAGE 458

436 Whiffen, V. E., & Johnson, S. M. (1998). An Attachment Theory Framework for the Treatment of Childbearing Depression. Clinical Psychology: Science and Practice, 5 (4), 478-493. Wile, J., & Arechiga, M. (1999). Sociocultura l Aspects of postpartum depression. In L. Miller (Ed.), Postpartum Mood Disorders (pp. 83-98). Washington DC: American Psychiatric Press, Inc. Zahn-Waxler, C., Iannotti, R., Cummings, E. M., & Denham, S. (1990). Antecedents of problem behaviors in childr en of depressed mothers. Development and Psychopathology, 2 271-291. Zeanah, C., Boris, N., & Scheeringa M. (1997). Psychopathology in infancy. Journal of Child Psychology and Psychiatry and Allied Disciplines, 38 81-99. Zung, W. (1965). A self-rat ing depression scale. Archives of General Psychiatry, 12 63-70.

PAGE 459

437 Appendix A: Pearson Correlation Matrices for Maternal Depression Over Time, Maternal Sensitivity Over Time, and Maternal Depression and Maternal Sensitivity Over Time

PAGE 460

438 Table A1: Pearson Correlation Matrices fo r Maternal Depression Over Time, Maternal Sensitivity Over Time, and Maternal Depr ession and Maternal Sens itivity Over Time Maternal Depression Maternal Depression 1 Month 6 Months 15 Months 24 Months 36 Months 1 Month -6 Months 0.38* -15 Months 0.39* 0.51* -24 Months 0.33* 0.46* 0.41* -36 Months 0.38* 0.41* 0.42* 0.46* -Maternal Sensitivity Maternal Sensitivity 1 Month 6 Months 15 Months 24 Months 36 Months 1 Month -6 Months --15 Months -0.32* -24 Months -0.22* 0.35* -36 Months -0.34* 0.31* 0.38* -Maternal Sensitivity Maternal Depression 1 Month 6 Months 15 Months 24 Months 36 Months 1 Month --0.11* -0.14* -0.11* -0.15* 6 Months --0.11* -0.16* -0.11* -0.18* 15 Months ---0.15* -0.09* -0.13* 24 Months ----0.18* -0.24* 36 Months -----0.17* *Indicates statistical significance at p<.05.

PAGE 461

439 Appendix B: Partial Model Results from OL S Regression Testing the Impact of Ever Depressed on Child Social Competence and Disp lay of Problem Behavi ors at First Grade: Significant Main Effects of Ever Depressed

PAGE 462

440 Table B1: Partial Model Results Testing the Effect of Ever De pressed on Maternal Report of Socia l Competence Including (1) Tot al Social Competence; (2) Assertion; and (3) Assertion Total Social Competence Assertion Peer Competence b coefficient t value (b=0) b coefficient b coefficient b coefficient t value (b=0) Ever Depressed No 0.00 -0.00 -0.00 -Yes -2.93 -2.57* -0.46 -2.55 -0.52 -2.51* Model Fit Statistics N 679 679 679 R2 0.11 0.06 0.07 Overall Model F 5.27* 2.85* 3.09* *Indicates statistical significance at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeedi ng at one month postpartum.

PAGE 463

441 Table B2: Partial Model Results Testing the Effect of Ever De pressed on Maternal Report of Child Display of Problem Behaviors Including (1) Total Problem Behaviors; (2) Inter nalizing Problems; and (3) Externalizing Problems Total Problem Behaviors Internalizing Problems Externalizing Problems b coefficient t value (b=0) b coefficient t value (b=0) b coefficient t value (b=0) Ever Depressed No 0.00 -0.00 -0.00 -Yes 2.19 2.92* 1.91 2.76* 1.96 2.58* Model Fit Statistics N 679 679 679 R2 0.05 0.04 0.05 Overall Model F 2.32* 2.05* 2.55* *Indicates statistical significance at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeedi ng at one month postpartum.

PAGE 464

442 Appendix C: Partial Model Results from OLS Regression Testing the Impact of Timing of Initial Onset of Maternal Depression on Child Social Competence and Display of Problem Behaviors at First Grade: Significan t Main Effects of Timing of Initial Onset

PAGE 465

443 Table C1: Partial Model Results from OLS Regression Testing th e Impact of Timing of Initial Onset of Ma ternal Depression on Maternal Report of Total Social Competence Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depressed at One Month No 0.00 (--) na na na na Yes -2.16 (-1.50) na na na na Depressed at Six Months No na 0.00 (--) na na na Yes na -4.26 (-2.01)* na na na Depressed at 15 Months No na na 0.00 (--) na na Yes na na -2.09 (-0.70) na na Continued on the next page

PAGE 466

444 Table C1 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depressed at 24 Months No na na na 0.00 (--) na Yes na na na -2.72 (-1.15) na Depressed at 36 Months No na na na na 0.00 (--) Yes na na na na 0.06 (0.02) Model Fit Statistics N 679 558 510 486 447 R2 0.10 0.12 0.10 0.11 0.11 Overall Model F 4.95* 4.81* 3.84* 3.79* 3.64* *Indicates statistical significance at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum ..

PAGE 467

445 Table C2: Partial Model Results from OLS Regression Testing th e Impact of Timing of Initial Onset of Ma ternal Depression on Maternal Report of Assertion Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depressed at One Month No 0.00 (--) na na na na Yes -0.37 (-1.61) na na na na Depressed at Six Months No na 0.00 (--) na na na Yes na -0.67 (-1.99)* na na na Depressed at 15 Months No na na 0.00 (--) na na Yes na na -0.53 (-1.14) na na Continued on the next page

PAGE 468

446 Table C2 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depressed at 24 Months No na na na 0.00 (--) na Yes na na na -0.09 (-0.23) na Depressed at 36 Months No na na na na 0.00 (--) Yes na na na na -0.39 (-0.81) Model Fit Statistics N 679 558 510 486 447 R2 0.06 0.06 0.05 0.04 0.04 Overall Model F 2.57* 2.20* 1.87* 1.47 1.22 *Indicates statistical significance at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum.

PAGE 469

447 TableC3: Partial Model Results from OLS Regression Testing the Impact of Timing of Initial Onset of Ma ternal Depression on Maternal Report of Peer Competence Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Depressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depressed at One Month No 0.00 (--) na na na na Yes -0.49 (-1.86) na na na na Depressed at Six Months No na 0.00 (--) na na na Yes na -0.82 (-2.13)* na na na Depressed at 15 Months No na na 0.00 (--) na na Yes na na -0.16 (-0.30) na na Continued on the next page

PAGE 470

448 Table C3 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Depressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depressed at 24 Months No na na na 0.00 (--) na Yes na na na -0.25 (-0.59) na Depressed at 36 Months No na na na na 0.00 (--) Yes na na na na -0.22 (-0.39) Model Fit Statistics N 679 558 510 486 447 R2 0.06 0.07 0.07 0.07 0.06 Overall Model F 2.88* 2.70* 2.31* 2.19* 1.70* *Indicates statistical significance at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum .

PAGE 471

449 Table C4: Partial Model Results from OLS Regression Testing th e Impact of Timing of Initial Onset of Ma ternal Depression on Teacher Report of Total Social Competence Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depressed at One Month No 0.00 (--) na na na na Yes -0.26 (-0.20) na na na na Depressed at Six Months No na 0.00 (--) na na na Yes na -4.07 (-2.03)* na na na Depressed at 15 Months No na na 0.00 (--) na na Yes na na -0.14 (-0.05) na na Continued on the next page

PAGE 472

450 Table C4 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depressed at 24 Months No na na na 0.00 (--) na Yes na na na 0.93 (0.42) na Depressed at 36 Months No na na na Na 0.00 (--) Yes na na na Na 2.83 (0.98) Model Fit Statistics N 679 558 510 486 447 R2 0.07 0.07 0.05 0.04 0.06 Overall Model F 3.21* 2.53* 1.56 1.40 1.86* *Indicates statistical significance at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum .

PAGE 473

451 Table C5: Partial Model Results from OLS Regression Testing th e Impact of Timing of Initial Onset of Ma ternal Depression on Teacher Report of Self Control Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depressed at One Month No 0.00 (--) na na na na Yes -0.18 (-0.52) na na na na Depressed at Six Months No na 0.00 (--) na na na Yes na -1.32 (-2.52)* na na na Depressed at 15 Months No na na 0.00 (--) na na Yes na na 1.17 (1.65) na na Continued on the next page

PAGE 474

452 Table C5 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depressed at 24 Months No na na na 0.00 (--) na Yes na na na -0.16 (-0.27) na Depressed at 36 Months No na na na na 0.00 (--) Yes na na na na 0.90 (1.21) Model Fit Statistics N 679 558 510 486 447 R2 0.07 0.07 0.07 0.06 0.08 Overall Model F 3.43* 2.80* 2.38* 2.17* 2.48* *Indicates statistical significance at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum.

PAGE 475

453 Table C6: Partial Model Results from OLS Regression Testing th e Impact of Timing of Initial Onset of Ma ternal Depression on Teacher Report of Peer Competence Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depressed at One Month No 0.00 (--) na na na na Yes -0.12 (-0.35) na na na na Depressed at Six Months No na 0.00 (--) na na na Yes na -1.67 (-3.23)* na na na Depressed at 15 Months No na na 0.00 (--) na na Yes na na 0.27 (0.38) na na Continued on the next page

PAGE 476

454 Table C6 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depressed at 24 Months No na na na 0.00 (--) na Yes na na na -0.21 (-0.36) na Depressed at 36 Months No na na na na 0.00 (--) Yes na na na na 1.03 (1.42) Model Fit Statistics N 679 558 510 486 447 R2 0.07 0.08 0.06 0.07 0.09 Overall Model F 3.31* 3.05* 2.23* 2.23* 2.80* *Indicates statistical significance at p<.05. Note: Controlling for all sociodemographic contro l variables including child sex, child birth order, maternal age, maternal rac e, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum .

PAGE 477

455 Table C7: Partial Model Results from OLS Regression Testing th e Impact of Timing of Initial Onset of Ma ternal Depression on Maternal Report of Total Problem Behaviors Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depressed at One Month No 0.00 (--) na na na na Yes 2.39 (2.53)* na na na na Depressed at Six Months No na 0.00 (--) na na na Yes na 2.53 (1.82) na na na Depressed at 15 Months No na na 0.00 (--) na na Yes na na 0.60 (0.32) na na Continued on the next page

PAGE 478

456 Table C7 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depressed at 24 Months No na na na 0.00 (--) na Yes na na na 2.14 (1.42) na Depressed at 36 Months No na na na na 0.00 (--) Yes na na na na -0.33 (-0.17) Model Fit Statistics N 679 558 510 486 447 R2 0.05 0.04 0.06 0.06 0.06 Overall Model F 2.17* 1.63 2.19* 2.15* 1.79* *Indicates statistical significance at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum .

PAGE 479

457 Table C8: Partial Model Results from OLS Regression Testing th e Impact of Timing of Initial Onset of Ma ternal Depression on Maternal Report of Externalizing Problem Behaviors Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depressed at One Month No 0.00 (--) na na na na Yes 2.15 (2.25)* na na na na Depressed at Six Months No na 0.00 (--) na na na Yes na 2.25 (1.60) na na na Depressed at 15 Months No na na 0.00 (--) na na Yes na na 0.29 (0.15) na na Continued on the next page

PAGE 480

458 Table C8 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depressed at 24 Months No na na na 0.00 (--) na Yes na na na 1.67 (1.11) na Depressed at 36 Months No na na na na 0.00 (--) Yes na na na na 0.16 (0.08) Model Fit Statistics N 679 558 510 486 447 R2 0.05 0.05 0.07 0.07 0.07 Overall Model F 2.44* 1.86* 2.39* 2.44* 2.19* *Indicates statistical significance at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum .

PAGE 481

459 Appendix D: Partial Model Results from OLS Regression Testing the Impact of Depression Severity Along the Trajectory of Initial Onset on Child Social Competence and Display of Problem Behaviors at First Grade: Significant Main Effects of Depression Severity

PAGE 482

460 Table D1: Partial Model Results from OLS Regression Testing the Impact of Severity of Maternal Depression Along the Trajectory of Initial Onset on Maternal Report of Total Social Competence Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depression Severity at One Month -0.31 (-2.30)* na na na na Depression Severity at Six Months na -0.20 (-0.93) na na na Depression Severity at 15 Mont hs na na -0.42 (-1.20) na na Depression Severity at 24 Mont hs na na na -0.06 (-0.24) na Depression Severity at 36 M onths na na na na 0.13 (0.37) Model Fit Statistics N 679 558 510 486 447 R2 0.10 0.11 0.11 0.11 0.11 Overall Model F 5.17* 4.57* 3.91* 3.70* 3.65* *Indicates statistical significance at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum .

PAGE 483

461 Table D2: Partial Model Results from OLS Regression Testing the Impact of Severity of Maternal Depression Along the Trajectory of Initial Onset on Maternal Report of Self Control Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depression Severity at One Month -0.06 (-2.10)* na na na na Depression Severity at Six Months na -0.03 (-0.70) na na na Depression Severity at 15 Mont hs na na -0.15 (-1.92) na na Depression Severity at 24 Mont hs na na na -0.04 (-0.63) na Depression Severity at 36 M onths na na na na 0.12 (1.42) Model Fit Statistics N 679 558 510 486 447 R2 0.09 0.09 0.10 0.09 0.10 Overall Model F 4.52* 3.78* 3.58* 3.08* 3.30* *Indicates statistical significance at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum .

PAGE 485

463 Table D3: Partial Model Results from OLS Regression Testing the Impact of Severity of Maternal Depression Along the Trajectory of Initial Onset on Maternal Report of Peer Competence Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depression Severity at One Month -0.06 (-2.44)* na na na na Depression Severity at 6 Mont hs na -0.04 (-0.92) na na na Depression Severity at 15 M onths na na 0.01 (0.08) na na Depression Severity at 24 Mont hs na na na -0.01 (-0.31) na Depression Severity at 36 M onths na na na na 0.03 (0.39) Model Fit Statistics N 679 558 510 486 447 R2 0.06 0.06 0.07 0.06 0.06 Overall Model F 3.06* 2.44* 2.30* 2.17* 1.70* *Indicates statistical significance at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum .

PAGE 486

464 Table D4: Partial Model Results from OLS Regression Testing the Impact of Severity of Maternal Depression Along the Trajectory of Initial Onset on Teacher Report of Peer Competence Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depression Severity at One Month -0.03 (-0.78) na na na na Depression Severity at 6 Months na -0.11 (-2.22)* na na na Depression Severity at 15 M onths na na 0.05 (0.57) na na Depression Severity at 24 M onths na na na 0.01 (0.19) na Depression Severity at 36 M onths na na na na 0.02 (0.24) Model Fit Statistics N 679 558 510 486 447 R2 0.07 0.07 0.06 0.07 0.08 Overall Model F 3.35* 2.66* 2.24* 2.22* 2.66* *Indicates statistical significance at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum .

PAGE 487

465 Table D5: Partial Model Results from OLS Regression Testing the Impact of Severity of Maternal Depression Along the Trajectory of Initial Onset on Maternal Report of Total Problem Behaviors Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depression Severity at One Month 0.20 (2.26)* na na na na Depression Severity at 6 Mont hs na 0.19 (1.40) na na na Depression Severity at 15 M onths na na 0.01 (0.02) na na Depression Severity at 24 Mont hs na na na -0.02 (-0.10) na Depression Severity at 36 Mont hs na na na na -0.31 (-1.31) Model Fit Statistics N 679 558 510 486 447 R2 0.05 0.04 0.06 0.06 0.06 Overall Model F 2.09* 1.54 2.18* 2.01* 1.91* *Indicates statistical significance at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum .

PAGE 488

466 Table D6: Partial Model Results from OLS Regression Testing the Impact of Severity of Maternal Depression Along the Trajectory of Initial Onset on Maternal Report of Externalizing Problem Behaviors1 Model 1 Model 2 Model 3 Model 4 Model 5 Month of Initiation 1 6 15 24 36 Population All Never Depressed Never Dep ressed Never Depressed Never Depressed Independent Variables b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) b coefficient (t-value) Depression at One Month 0.21 (2.31)* na na na na Depression at 6 Months na 0.23 (1.63) na na na Depression at 15 Months na na 0.02 (0.11) na na Depression at 24 Months na na na -0.03 (-0.20) na Depression at 36 Months na na na na -0.31 (-1.31) Model Fit Statistics N 679 558 510 486 447 R2 0.05 0.05 0.07 0.07 0.07 Overall Model F 2.46* 1.87* 2.39* 2.36 2.31* *Indicates statistical significance at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum .

PAGE 489

467 Appendix E: Partial Model Results from OLS Regression Testi ng the Impact of Depression Chronicity on Child Social Comp etence and Display of Problem Behaviors at First Grade: Significant Main E ffects of Depression Chronicity

PAGE 490

468 Table E1: Partial Model Results from OLS Regression Testing the Impact of Depressi on Chronicity on Maternal Report of Total Social Competence Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at One Month Never 0.00/0.00 (--) na na na na New -2.16/-0.06 (-1.50) na na na na Depression Chronicity at Six Months Never na 0.00/0.00 (--) na na na New na -4.30/-0.08 (-2.02)* na na na Remitted na -1.27/-0.03 (-0.75) na na na Chronic na -5.54/-0.09 (-2.31)* na na na Depression Chronicity at 15 Months Never na na 0.00/0.00 (--) na na New na na -2.34/-0.03 (-0.80) na na Remitted na na -2.47/-0.64 (-1.71) na na Recurrent na na -2.38/-0.87 (-0.03) na na Chronic na na -9.03/-2.80 (-0.11)* na na Continued on the next page

PAGE 491

469 Table E1 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at 24 Months Never na na na 0.00/0.00 (--) na New na na na -2.96/-0.05 (-1.26) na Remitted na na na -2.63/-0.08/-1.96 na Recurrent na na na -4.34/-0.06/-1.67 na Chronic na na na -12.89/-0.10 (-2.73)* na Depression Chronicity at 36 Months Never na na na na 0.00/0.00 (--) New na na na na 0.23/0.00 (0.08) Remitted na na na na -1.79/-0.05 (-1.39) Recurrent na na na na -7.15/-0.12 (-3.27)* Chronic na na na na -12.79/-0.10 (-2.72)* Model Fit Statistics N 679 679 679 679 679 R2 0.10 0.11 0.11 0.11 0.12 Overall Model F 4.95* 4.77* 4.60* 4.74* 5.05* *Indicates statistical si gnificant at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum .

PAGE 492

470 Table E2: Partial Model Results from OL S Regression Testing the Impact of Depr ession Chronicity on Maternal Report of Cooperation Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at One Month Never 0.00/0.00 (--) na na na na New -0.23/-0.03 (-0.74) na na na na Depression Chronicity at Six Months Never na 0.00/0.00 (--) na na na New na -0.70/-0.06 (-1.56) na na na Remitted na -0.29/-0.03 (-0.81) na na na Chronic na -0.32/-0.02 (-0.61) na na na Depression Chronicity at 15 Months Never na na 0.00/0.00 (--) na na New na na 0.09/0.01 (0.14) na na Remitted na na -0.34/-0.04 (-1.08) na na Recurrent na na -0.37/-0.02 (-0.62) na na Chronic na na -0.98/-0.05 (-1.40) na na Continued on the next page

PAGE 493

471 Table E2 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at 24 Months Never na na na 0.00/0.00 (--) na New na na na -0.67-0.05/ (-1.33) na Remitted na na na -0.27/-0.04 (-0.93) na Recurrent na na na -0.56/-0.04 (-1.00) na Chronic na na na -2.44/-0.09 (-2.39) na Depression Chronicity at 36 Months Never na na na na 0.00/0.00 (--) New na na na na -0.23/-0.01 (-0.36) Remitted na na na na -0.37/-0.05 (-1.34) Recurrent na na na na -0.47/-0.04 (-1.00) Chronic na na na na -2.44/-0.09 (-2.40) Model Fit Statistics N 679 679 679 679 679 R2 0.03 0.04 0.04 0.05 0.04 Overall Model F 1.59 1.54 1.46 1.74* 1.71* *Indicates statistical significant at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum .

PAGE 494

472 Table E3: Partial Model Results from OLS Regression Testing the Impact of Depression Chronicity on Maternal Report of Assertio n Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at One Month Never 0.00/0.00 (--) na na na na New -0.37/-0.06 (-1.61) na na na na Depression Chronicity at Six Months Never na 0.00/0.00 (--) na na na New na -0.66/-0.08 (-1.95) na na na Remitted na -0.22/-0.03 (-0.82) na na na Chronic na -0.92/-0.09 (-2.39)* na na na Depression Chronicity at 15 Months Never na na 0.00/0.00 (--) na na New na na -0.52/-0.04 (-1.11) na na Remitted na na -0.54/-0.09 (-2.34)* na na Recurrent na na -0.36/-0.03 (-0.83) na na Chronic na na -0.68/-0.05 (-1.31) na na Continued on the next page

PAGE 495

473 Table E3 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at 24 Months Never na na na 0.00/0.00 (--) na New na na na -0.10/-0.01 (-0.27) na Remitted na na na -0.45/-0.08(-2.13)* na Recurrent na na na -0.84/-0.08 (-2.02)* na Chronic na na na -0.84/-0.04 (-1.11) na Depression Chronicity at 36 Months Never na na na na 0.00/0.00 (--) New na na na na -0.34/-0.03 (-0.70) Remitted na na na na -0.36/-0.07 (-1.75) Recurrent na na na na -0.85/-0.09 (-2.43)* Chronic na na na na -0.84/-0.04 (-1.11) Model Fit Statistics N 679 679 679 679 679 R2 0.06 0.06 0.06 0.06 0.06 Overall Model F 2.57* 2.65* 2.42* 2.46* 2.48* *Indicates statistical significant at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum .

PAGE 496

474 Table E4: Partial Model Results from OL S Regression Testing the Impact of Depr ession Chronicity on Maternal Report of Responsibility Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at One Month Never 0.00/0.00 (--) na na na na New -0.15/-0.02 (-0.57) na na na na Depression Chronicity at Six Months Never na 0.00/0.00 (--) na na na New na -0.55/-0.05 (-1.41) na na na Remitted na 0.12/0.01 (0.38) na na na Chronic na -0.93/-0.08 (-2.09)* na na na Depression Chronicity at 15 Months Never na na 0.00/0.00 (--) na na New na na -0.57/-0.04 (-1.06) na na Remitted na na -0.20/-0.03 (-0.75) na na Recurrent na na -0.08/-0.01 (-0.15) na na Chronic na na -1.54/-0.10 (-2.59)* na na Continued on the next page

PAGE 497

475 Table E4 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at 24 Months Never na na na 0.00/0.00 (--) na New na na na -0.40/-0.04 (-0.92) na Remitted na na na -0.25/-0.04 (-1.00) na Recurrent na na na -0.96/-0.08 (-1.99)* na Chronic na na na -1.09/-0.05 (-1.24) na Depression Chronicity at 36 Months Never na na na na 0.00/0.00 (--) New na na na na -0.05/-0.00 (-0.10) Remitted na na na na -0.11/-0.02 (-0.47) Recurrent na na na na -1.37/-0.13 (-3.39)* Chronic na na na na -1.06/-0.05 (-1.22) Model Fit Statistics N 679 679 679 679 679 R2 0.08 0.09 0.09 0.09 0.10 Overall Model F 3.96* 3.87* 3.73* 3.63* 4.03* *Indicates statistical significant at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum .

PAGE 498

476 Table E5: Partial Model Results from OLS Regression Testing the Impact of Depression Severity on Maternal Report of Self Contr ol Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at One Month Never 0.00/0.00 (--) na na na na New -0.49/-0.06 (-1.50) na na na na Depression Chronicity at Six Months Never na 0.00/0.00 (--) na na na New na -0.50/-0.04 (-1.01) na na na Remitted na -0.34/-0.03 (-0.89) na na na Chronic na -0.99/-0.07 (-1.80) na na na Depression Chronicity at 15 Months Never na na 0.00/0.00 (--) na na New na na -0.67/-0.04 (-1.00) na na Remitted na na -0.29/-0.03 (-0.87) na na Recurrent na na -0.65/-0.04 (-1.05) na na Chronic na na -2.19/-0.11 (-2.97)* na na Continued on the next page

PAGE 499

477 Table E5 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at 24 Months Never na na na 0.00/0.00 (--) na New na na na -0.72/-0.05 (-1.34) na Remitted na na na -0.54/-0.07 (-1.77) na Recurrent na na na -0.35/-0.02 (-0.58) na Chronic na na na -3.48/-0.12 (-3.22)* na Depression Chronicity at 36 Months Never na na na na 0.00/0.00 (--) New na na na na 0.80/0.04 (1.15) Remitted na na na na -0.22/-0.03 (-0.77) Recurrent na na na na -1.57/-0.12 (-3.15)* Chronic na na na na -3.44/-0.12 (-3.20)* Model Fit Statistics N 679 679 679 679 679 R2 0.09 0.09 0.10 0.11 0.12 Overall Model F 4.36* 3.97* 4.10* 4.30* 4.77* *Indicates statistical significant at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum.

PAGE 500

478 Table E6: Partial Model Results from OLS Regression Testing the Impact of Depressi on Chronicity on Maternal Report of Peer Competence Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at One Month Never 0.00/0.00 (--) na na na na New -0.49/-0.07 (0.06) na na na na Depression Chronicity at Six Months Never na 0.00/0.00 (--) na na na New na -0.79/-0.08 (-2.05)* na na na Remitted na -0.28/-0.04 (-0.92) na na na Chronic na -1.20/-0.11 (-2.75)* na na na Depression Chronicity at 15 Months Never na na 0.00/0.00 (--) na na New na na -0.16/-0.01 (-0.30) na na Remitted na na -0.53/-0.08 (-2.02)* na na Recurrent na na -0.63/-0.05 (-1.28) na na Chronic na na -1.31/-0.09 (-2.23)* na na Continued on the next page

PAGE 501

479 Table E6 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at 24 Months Never na na na 0.00/0.00 (--) na New na na na -0.31/-0.03 (-0.74) na Remitted na na na -0.42/-0.07 (-1.72) na Recurrent na na na -1.05/-0.09 (-2.22)* na Chronic na na na -2.29/-0.10 (-2.66)* na Depression Chronicity at 36 Months Never na na na na 0.00/0.00 (--) New na na na na -0.16/-0.01 (-0.30) Remitted na na na na -0.39/-0.07 (-1.66) Recurrent na na na na -0.90/-0.09 (-2.26)* Chronic na na na na -2.27/-0.10 (-2.64)* Model Fit Statistics N 679 679 679 679 679 R2 0.06 0.07 0.07 0.05 0.07 Overall Model F 2.88* 3.00* 2.72* 2.94* 2.93* *Indicates statistical significant at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum .

PAGE 502

480 Table E7: Partial Model Results from OL S Regression Testing the Impact of Depres sion Chronicity on Teacher Report of Total Social Competence Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at One Month Never 0.00/0.00 (--) na na na na New -0.26/-0.01 (-0.20) na na na na Depression Chronicity at Six Months Never na 0.00/0.00 (--) na na na New na -3.81/-0.07 (-1.92) na na na Remitted na -0.35/-0.01 (-0.22) na na na Chronic na -1.31/-0.02 (-0.59) na na na Depression Chronicity at 15 Months Never na na 0.00/0.00 (--) na na New na na -0.18/-0.00 (-0.06) na na Remitted na na -1.18/-0.03 (-0.88) na na Recurrent na na -5.04/-0.08 (-1.99)* na na Chronic na na 0.90/0.01 (0.30) na na Continued on the next page

PAGE 503

481 Table E7 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at 24 Months Never na na na 0.00/0.00 (--) na New na na na 1.33/0.02 (0.61) na Remitted na na na -1.40/-0.04 (-1.12) na Recurrent na na na -1.82/-0.03 (-0.75) na Chronic na na na 3.19/0.03 (0.72) na Depression Chronicity at 36 Months Never na na na na 0.00/0.00 (--) New na na na na 3.84/0.05 (1.35) Remitted na na na na -1.28/-0.04 (-1.07) Recurrent na na na na 1.16/0.02 (0.56) Chronic na na na na 3.42/0.03 (0.78) Model Fit Statistics N 679 679 679 679 679 R2 0.07 0.07 0.07 0.07 0.07 Overall Model F 3.21* 3.06* 2.94* 2.84* 2.93* *Indicates statistical significant at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeedi ng at one month postpartum.

PAGE 504

482 Table E8: Partial Model Results from OL S Regression Testing the Impact of Depres sion Chronicity on Teacher Report of Self Control Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at One Month Never 0.00/0.00 (--) na na na na New -0.18/-0.02 (-0.52) na na na na Depression Chronicity at Six Months Never na 0.00/0.00 (--) na na na New na -1.23/-0.09 (-2.33)* na na na Remitted na -0.40/-0.04 (-0.94) na na na Chronic na -0.12/0.01 (-0.21) na na na Depression Chronicity at 15 Months Never na na 0.00/0.00 (--) na na New na na 1.16/0.06 (1.60) na na Remitted na na -0.35/-0.04 (-0.97) na na Recurrent na na 1.72/0.10 (2.55)* na na Chronic na na 0.11/0.01 (0.13) na na Continued on the next page

PAGE 505

483 Table E8 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at 24 Months -----Never na na na 0.00/0.00 (--) na New na na na -0.05/0.00 (-0.09) na Remitted na na na -0.23/-0.03 (-0.70) na Recurrent na na na 0.76/0.05 (1.18) na Chronic na na na 0.09/0.00 (0.08) na Depression Chronicity at 36 Months -----Never na na na na 0.00/0.00 (--) New na na na na 1.18/0.06 (1.55) Remitted na na na na -0.42/-0.05 (-1.30) Recurrent na na na na 0.55/0.04 (1.00) Chronic na na na na 0.17/0.01 (0.14) Model Fit Statistics N 679 679 679 679 679 R2 0.07 0.08 0.09 0.07 0.08 Overall Model F 3.43* 3.37* 3.45* 2.93* 3.19* *Indicates statistical significant at p<.05. Note: Controlling for all sociodemographic contro l variables including child sex, child birth order, maternal age, maternal rac e, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeedi ng at one month postpartum.

PAGE 506

484 Table E9: Partial Model Results from OL S Regression Testing the Impact of Depres sion Chronicity on Teacher Report of Peer Competence Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at One Month Never 0.00/0.00 (--) na na na na New -0.12/-0.01 (-0.35) na na na na Depression Chronicity at Six Months Never na 0.00/0.00 (--) na na na New na -1.55/0.12 (-3.01)* na na na Remitted na -0.33/-0.03 (-0.81) na na na Chronic na -0.16/-0.01 (-0.28) na na na Depression Chronicity at 15 Months Never na na 0.00/0.00 (--) na na New na na 0.26/0.01 (0.37) na na Remitted na na -0.48/-0.05 (-1.38) na na Recurrent na na -1.85/-0.11 (-2.79*) na na Chronic na na 0.13/0.01 (0.17) na na Continued on the next page

PAGE 507

485 Table E9 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at 24 Months Never na na na 0.00/0.00 (--) na New na na na -0.12/-0.01 (-0.21) na Remitted na na na -0.50/-0.06 (-1.52) na Recurrent na na na -1.02/-0.06 (-1.61) na Chronic na na na 0.68/0.02 (0.59) na Depression Chronicity at 36 Months Never na na na na 0.00/0.00 (--) New na na na na 1.34/0.07 (1.81) Remitted na na na na -0.57/-0.07 (-1.82) Recurrent na na na na 0.08/0.00 (0.15) Chronic na na na na 0.77/0.03 (0.67) Model Fit Statistics N 679 679 679 679 679 R2 0.07 0.08 0.08 0.08 0.08 Overall Model F 3.31* 3.49* 3.31* 3.03* 3.23* *Indicates statistical significant at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum .

PAGE 508

486 Table E10: Partial Model Results from OL S Regression Testing the Impact of Depressi on Chronicity on Maternal Report of Total Problem Behaviors Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at One Month Never 0.00/0.00 (--) na na na na New 2.40/0.10 (2.53)* na na na na Depression Chronicity at Six Months Never na 0.00/0.00 (--) na na na New na 2.36/0.07 (1.69) na na na Remitted na 2.17/0.08 (1.95) na na na Chronic na 3.67/0.09 (2.33)* na na na Depression Chronicity at 15 Months Never na na 0.00/0.00 (--) na na New na na 0.82/0.02 (0.42) na na Remitted na na 1.53/0.06 (1.62) na na Recurrent na na 4.41/0.10 (2.47)* na na Chronic na na 6.80/0.12 (3.22)* na na Continued on the next page

PAGE 509

487 Table E10 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at 24 Months Never na na na 0.00/0.00 (--) na New na na na 2.37/0.06 (1.55) na Remitted na na na 1.52/0.07 (1.73) na Recurrent na na na 5.94/0.14 (3.50)* na Chronic na na na 9.40/0.12 (3.05)* na Depression Chronicity at 36 Months Never na na na na 0.00/0.00 (--) New na na na na -0.69/-0.01 (-0.35) Remitted na na na na 1.55/0.07 (1.85) Recurrent na na na na 4.860.13 (3.38)* Chronic na na na na 9.200.11 (2.98)* Model Fit Statistics N 679 679 679 679 679 R2 0.05 0.05 0.06 0.07 0.07 Overall Model F 2.17* 2.13* 2.37* 2.72* 2.64* *Indicates statistical significant at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum .

PAGE 510

488 Table E11:Partial Model Results from OL S Regression Testing the Impact of Depr ession Chronicity on Maternal Report of Internalizing Problem Behaviors Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at One Month Never 0.00/0.00 (--) na na na na New 1.69/0.08 (1.93) na na na na Depression Chronicity at Six Months Never na 0.00/0.00 (--) na na na New na 2.26/0.07 (1.75) na na na Remitted na 1.78/0.07 (1.74) na na na Chronic na 2.22/0.06 (1.52) na na na Depression Chronicity at 15 Months Never na na 0.00/0.00 (--) na na New na na 1.94/0.04 (1.09) na na Remitted na na 1.02/0.05 (1.16) na na Recurrent na na 5.62/0.13 (3.41)* na na Chronic na na 4.21/0.08 (2.16)* na na Continued on the next page

PAGE 511

489 Table E11 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at 24 Months Never na na na 0.00/0.00 (--) na New na na na 1.02/0.03 (0.72) na Remitted na na na 1.22/0.06 (1.51) na Recurrent na na na 6.24/0.16 (3.98)* na Chronic na na na 4.69/0.06 (1.64) na Depression Chronicity at 36 Months Never na na na na 0.00/0.00 (--) New na na na na 0.91/0.02 (0.49) Remitted na na na na 1.24/0.06 (1.59) Recurrent na na na na 4.48/0.13 (3.36)* Chronic na na na na 4.54/0.06 (1.58) Model Fit Statistics N 679 679 679 679 679 R2 0.04 0.04 0.06 0.06 0.05 Overall Model F 1.79* 1.76* 2.19* 2.33* 2.08* *Indicates statistical significant at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum .

PAGE 512

490 Table E12: Partial Model Results from OLS Regression Testing the Impact of Depr ession Chronicity on Maternal Report of Externalizing Problem Behaviors Model 1 Model 2 Model 3 Model 4 Model 5 Month of Assessment 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at One Month Never 0.00/0.00 (--) na na na na New 2.15/0.09 (2.25)* na na na na Depression Chronicity at Six Months Never na 0.00/0.00 (--) na na na New na 2.16/0.06 (1.52) na na na Remitted na 2.12/0.07 (1.88) na na na Chronic na 2.94/0.07 (1.84) na na na Depression Chronicity at 15 Months Never na na 0.00/0.00 (--) na na New na na 0.47/0.01 (0.24) na na Remitted na na 1.37/0.06 (1.42) na na Recurrent na na 3.62/0.08 (2.00)* na na Chronic na na 6.71/0.12 (3.13)* na na Continued on the next page

PAGE 513

491 Table E12 (Continued) Model 1 Model 2 Model 3 Model 4 Model 5 Month 1 6 15 24 36 Independent Variables b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) b coefficient/ Beta coefficient (t-value) Depression Chronicity at 24 Months Never na na na 0.00/0.00 (--) na New na na na 1.87/0.05 (1.20) na Remitted na na na 1.57/0.07 (1.75) na Recurrent na na na 3.51/0.08 (2.03)* na Chronic na na na 10.26/0.12 (3.26)* na Depression Chronicity at 36 Months Never na na na na 0.00/0.00 (--) New na na na na -0.11/-0.00 (-0.06) Remitted na na na na 1.51/0.07 (1.77) Recurrent na na na na 3.28/0.09 (2.24)* Chronic na na na na 10.17/0.12 (3.24)* Model Fit Statistics N 679 679 679 679 679 R2 0.05 0.06 0.06 0.07 0.07 Overall Model F 2.44* 2.30* 2.52* 2.66* 2.67* *Indicates statistical significant at p<.05. Note: Controlling for all sociodemographic cont rol variables includi ng child sex, child birth order, maternal age, maternal ra ce, maternal ethnicity, maternal education, pr esence of partner/husband at home at one m onth postpartum, family receipt of public assistance at one month postpar tum, and maternal breastfeed ing at one month postpartum.

PAGE 514

About the Author Lauren Zapata received a Bachel or’s Degree in Biology from Emory University in 1997 and a M.S.P.H. from the Co llege of Public Health, University of South Florida in 1999. Following the completi on of her M.S.P.H., she began the Ph.D. program in the College of Public Health, University of South Florida. Lauren has two children, Paolo, age three, and Noah, age one. She and her family currently live in Buford, Georgia. Af ter graduation, Lauren will begin a two-year post-doctorate fellowship at the Centers for Di sease Control and Prevention in Atlanta as an Epidemic Intelligence Service officer in the commissioned corps of the Public Health Service.


xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001680948
003 fts
005 20060215071021.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 051130s2005 flu sbm s000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001174
035
(OCoLC)62385567
SFE0001174
040
FHM
c FHM
049
FHMM
090
RA425 (Online)
1 100
Zapata, Lauren B.
0 245
Associations between maternal depression and child social competence and display of problem behaviors
h [electronic resource] :
b a longitudinal investigation of direct, indirect and moderating effects /
by Lauren B. Zapata.
260
[Tampa, Fla.] :
University of South Florida,
2005.
502
Thesis (Ph.D.)--University of South Florida, 2005.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 514 pages.
Includes vita.
520
ABSTRACT: Postpartum depression is a significant public health problem facing women, children, and families in the United States with an estimated 10-15% of U.S. mothers experiencing an episode of non-psychotic depression within six months of delivery. The purpose of this study was to investigate the associations between maternal depression during the first three years postpartum and child social competence and display of problem behaviors at first grade. The impact of several characteristics of maternal depression were examined including general exposure, timing of initial onset in the postpartum period, severity of symptoms along the trajectory of initial onset, and chronicity of symptoms. This study also explored the mediating and moderating influences of maternal sensitivity, as well as the moderating influence of exposure to nonmaternal care.This study was based upon secondary analysis of data from the National Institute of Child Health and Human Development Study of Early Child Care (NICHD SECC), a multi-site, prospective, three-phase longitudinal study. The sample consisted of 679 infants and their families that participated in both Phase I and Phase II of the NICHD SECC, had compete data on all variables of interest in the study, and had non-depressed mothers at first grade. Results identified the first six months postpartum as a sensitive period of risk for depression initiation. Severity of symptoms was also found to be important. In some instances depression alone did not increase risk for lower levels of social competence, but severity of symptoms above cut points indicating depression did. Chronic depression at 24 months rather than 36 months postpartum was found to pose the greatest magnitude of negative influence on outcome.
590
Adviser: Melinda S. Forthofer, PhD.
653
Postpartum.
Mental health.
Child development.
Maternal sensitivity.
Attachment theory.
690
Dissertations, Academic
z USF
x Public Health
Doctoral.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.1174