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 22 Ka 4500
controlfield tag 007 cr-bnu---uuuuu
008 s2010 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0004598
Literacy and behavior in early childhood :
b exploring the factors that impact achievement
h [electronic resource] /
by Melissa Todd.
[Tampa, Fla] :
University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains X pages.
Dissertation (PHD)--University of South Florida, 2010.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
ABSTRACT: Academic achievement has been the focal point in education for decades. In 2001, an Act of Congress was proposed to improve individual outcomes in education through evidenced based research using measurable goals, higher standards, and accountability. This federal legislation, known as the No Child Left Behind Act of 2001, mandates that all teachers be highly qualified by 2006 and that all students become proficient by the 2013/14 school year, specifically in the area of literacy. Consequently, kindergarten readiness has become an area of concern, thus placing preschool teachers under pressure to prepare children for school. The purpose of this study was to examine multiple factors that have been identified in the literature as impacting achievement in elementary and secondary education to ascertain their contribution toward literacy development in preschool children. Such factors included child (gender, race, home SES, attendance, behavior) and childcare site (teacher education, teacher experience, class size, site SES, class environment). Additionally, within-child protective factors were examined for their role in literacy development for children with and without challenging behaviors. To examine early literacy and behavior in preschool children, hierarchical linear modeling (HLM) was conducted with literacy skills (expressive language and phonemic awareness) assessed at four points in time though the Individual Development and Growth Indicators (IGDI). A significant relationship was found between expressive language skills and race, attendance, classroom environment and class size. Phonemic awareness was significantly related to gender, home SES, and teacher education. Within-child protective factors positively impacted phonemic awareness skills for children in the non-challenging behavior group only. An in-depth description of the findings and limitations are discussed within this document. Overall, this study suggests that many of the factors impacting achievement in elementary and secondary education also impact literacy development in preschool children. These findings support the use of early intervention and preventative services for this population as a means to promote kindergarten readiness and future achievement.
Advisor: George Batsche, Ed.D.
x Psychological & Social Foundations
t USF Electronic Theses and Dissertations.
Literacy and Behavior in Early Childhood: Exploring the Factors t hat Impact Achievement by Melissa Farino Todd A dissertation proposal submitted in partial fulfillment of the requirement for the degree Doctor of Philosophy Department of School Psychology College of Education University of South Florida Co-Major Professor: George Batsche, Ed.D. Co-Major Professor: Kath leen Hague-Armstrong, Ph.D. Linda Rafaelle Mendez, Ph.D. John Ferron, Ph.D. Date of Approval: July 10, 2010 Keywords: NCLB, preschool, IGD I, DECA, protective factors Copyright 2010, Melissa Farino Todd
Dedication I would like to dedicate this dissertat ion to my amazing daughters, Kaitlyn and Olivia, who have demonstrated pati ence beyond their years during the many hours Mommy was writing (a.k.a., doing Â“h omeworkÂ”). I can only hope that through this process I have taught t hem the importance of education and perseverance. Girls, never stop reaching for the stars or let anything stand in the way of your dreams. Mommy did it, and so can you. I love you both more than you can imagine. I would like to thank my husband, Shannon, for his unwavering love and support as I pursued my doctorate. Thr oughout that decade, you celebrated my achievements and helped me through the chal lenges. I am certain that this document would not have been completed wi thout your support. You have been my rockÂ….I love you. Thank you, also, to my parents, Frank and Dianna Farino, who have been my cheerleaders since birth. Y ou have always been on the sidelines encouraging me to do my personal best. Words cannot express my gratitude not only for all that you have taught me but for the love and support that you have given me throughout my life. Finally, I would like to dedicate this dissertation to my late father-in-law, James C. Todd, Ph.D. Our frequent c onversations about academia and your genuine interest in my scholastic endeavors have been a gr eat source of motivation. I miss you tremendously and ta ke comfort in knowing that you are celebrating this achievement in heaven. I hope that I have made you proud.
i Table of Contents LIST OF TABLES iv LIST OF FIGURES vi ABSTRACT vii CHAPTER ONE: INTRODUCTION 1 Deve lopmental Trajectories: Im plications for Academic Achievement 3 Implications of Children At-Risk 4 Behavioral Co mpetencies and Academic Achievement 8 Academic Achievement and Behavioral Co mpetencies: A Circular Relationship 10 Stat ement of the Problem 11 Pu rpose of the Study 13 Research Questions 13 Definition of Terms 14 CHAPTER TWO: REVIEW OF THE LITERATURE 16 Risk Fact ors Associated with Academic and Social-Emotional Development 19 Ec ological Systems Theory 22 Microsystem 22 Effects of gender on achievement 22 Effe cts of gender on behavior 24 Effects of minor ity status on achievement 25 Effects of minority status on behavior 28 Effects of family soci oeconomic status on achievement 29 Effects of family socioeconomic status on behavior 33 Effects of class size on achievement 35 Effects of class size on behavior 37 Effects of attendance on achievement 37 Effects of teac her experience and education on achievement and behavior 38 Mesosystem 39 Exosystem 40 Neighborhood di sadvantage and achievement 40 Neighbor hood disadvantage and behavior 42
ii Child care setting and academics 44 Child care setting and behavior 45 Asse ssing Change over Time 46 Summary 47 CHAPTER THREE: METHOD 49 Research Design 49 Descripti on of the Original Study 49 ELO Participants 51 Early Childhood Teachers 51 Presc hool Student Participants 53 Research Variables 54 Predi ctor/Independent Variables 54 Outc ome/Dependent Variable 57 Measures 57 Dem ographic Sheet 58 Individual Growth and Development Indicators Â– Preschool Version 58 Deveraux Early Childhood Assessment 61 Early Litera cy Observation Checklist 63 Procedur es for Original Study 66 Cohort One 69 Cohort Two 73 Cohort Three 73 Procedures for Selection, Review and Analysis of Archival Data in Current Study 75 Archival Data 75 Resear ch Design and Data Analysis 76 CHAPTER FOUR: RESULTS 83 De scriptive Statistics 83 Time One 93 Time Two 94 Time Three 94 Time Four 95 Hiera rchical Linear Modeling 95 Intr aclass Correlations 100 Picture Naming 112 Alliteration 115 Rhyming 116 Summary 120 CHAPTER FIVE: DISCUSSION 122 Summary 122 Responses to Research Questions 122 Research Question #1 122
iii Research Question #2 124 Research Question #3 128 Summary 129 Implications for School Psychology 129 Limitations 132 Future Research 135 REFERENCES 137 APPENDICES 151 Appendix A. A pplication and Agreement Form for ELO Teacher Participation 152 Appendix B. Demographic Information Sheet 154 Appendix C. Attendance Information Sheet 155 Appendix D. Teacher Consent Form 156 Appendix E. Parental Assent Form 159 Appendix F. Parental Information Letter 164 Appendix G. IGDI Recording Form 165 Appendi x H. HLM Path Diagram 166 ABOUT THE AUTHOR End Page
iv List of Tables Table 1. Definition of Terms 14 Table 2. Teacher Participants in Cohorts One, Two and Three 52 Table 3. Descriptive Information on Teacher Partici pants in Cohorts One and Two by Condition 53 Table 4. Descriptive Information on the Eight Teacher Participants in Cohort Three by Site 53 Table 5. Demographic Information on Student Participants in Cohort Three 54 Table 6. Measures Used to Assess Research Variables 57 Table 7. Summary of ELO Grant Procedures and Timeline for Cohorts One, Two and Three 68 Table 8. Description of Measurement Data for Cohort Three 76 Table 9. Descriptive Statistics related to Child and Teacher Demographic Characteristics 84 Table 10. Descriptive Statistics for Demographic Variables 85 Table 11. Skewness and Kurtosis Values for Dependent Measures 86 Table 12. Means and Standard Devi ations for Dependent Variables 87 Table 13. Correlations between Pr edictors and Literacy Outcomes at TimeOne 96 Table 14. Correlations between Pr edictors and Literacy Outcomes at Time Two 97 Table 15. Correlations between Pr edictors and Literacy Outcomes at Time Three 98
v Table 16. Correlations between Pr edictors and Literacy Outcomes at Time Four 99 Table 17. Intraclass Co rrelation Coefficients 100 Table 18. Linear Model of Literacy Growth 101 Table 19. Linear Model of Literacy Growth including Child Characteristics 105 Table 20. Linear Model of Literacy Growth including Child Characteristics: Variance Estimates 107 Table 21. Linear Model of Literacy Growth including Child and Classroom Characteristics 110 Table 22. Linear Model of Lite racy Growth including Child and Classroom Characteristics: Variance Estimates 111 Table 23. Linear Model of Literacy Growth including Child Characteristics 112 Table 24. Linear Model of Literacy Growth including Child Characteristics: Variance Estimates 117 Table 25. Linear Model of Literacy Growth including Child and Classroom Characteristics 118 Table 26. Linear Model of Lite racy Growth including Child and Classroom Characteristics: Variance Estimates 120
vi List of Figures Figure 1. Behavior and Picture Naming Scores 88 Figure 2. Behavior and Alliteration Scores 89 Figure 3. Behavior and Rhyming Scores 90 Figure 4. Relationship between Behavior and Race on Picture Naming Scores 91 Figure 5. Relationship between B ehavior and Race on Alliteration Scores 92 Figure 6. Relationship between Behavior and Race on Rhyming Scores 93 Figure 7. Growth Over Time for a Random Sele ction of Student Participants on the Picture Naming Subtest 102 Figure 8. Growth Over Time fo r a Random Select ion of Student Participants on the Alliteration Subtest 103 Figure 9. Growth Over Time fo r a Random Select ion of Student Participants on the Rhyming Subtest 104 Figure 10. Relationship between With in-child Protective Factors and the Picture Naming Slope 115 Figure 11. The Moderating Effect of Within-child Protective Factors on the Relationship between Behavior and Alliteration 116
vii Literacy and Behavior in Early Childhood: Exploring the Factors t hat Impact Achievement Melissa Farino Todd ABSTRACT Academic achievement has been the foca l point in education for decades. In 2001, an Act of Congress was proposed to improve individual outcomes in education through evidenced based research using measurable goals, higher standards, and accountability. This federal legislation, known as the No Child Left Behind Act of 2001, mandates that all teachers be highly qualified by 2006 and that all students become proficient by the 2013/14 school year, specifically in the area of literacy. Consequently, kindergarten r eadiness has become an area of concern, thus placing preschool t eachers under pressure to prepare children for school. The purpose of this study was to examine multiple factors that have been identified in the literature as impacting achievement in elementary and secondary education to ascertain their c ontribution toward lit eracy development in preschool children. Such factor s included child (gender, race, home SES, attendance, behavior) and childcare site (t eacher education, teacher experience, class size, site SES, class environment). Additionally, within-child protective factors were examined for t heir role in literacy develop ment for children with and without challenging behaviors.
viii To examine early literacy and behavior in preschool children, hierarchical linear modeling (HLM) was conducted with literacy skills (expressive language and phonemic awareness) assessed at four points in time though the Individual Development and Growth Indicators (IGDI) A significant relationship was found between expressive language skills and race, attendance, classroom environment and class size. Phonemic aw areness was significantly related to gender, home SES, and teacher education. Within-child protective factors positively impacted phonemic awareness skills for children in the non-challenging behavior group only. An in-depth description of the findings and limitations are discussed within this document. Overall, this study suggests that many of the factors impacting achievement in elementary and secondar y education also impact literacy development in preschool children. T hese findings support the use of early intervention and preventative services for this population as a means to promote kindergarten readiness and future achievement.
1 CHAPTER ONE INTRODUCTION Academic achievement, specifically in the area of literacy development, has been the focus of national concerns about effective schooling since the 1980Â’s (A Nation At Risk, 1984). Since t hat time, state and f ederal legislation has placed increasingly higher expectations on public schools to improve student achievement. In 2001, the United States Congress re-authorized the Elementary and Secondary Education Act (ESEA) as t he No Child Left Behind (NCLB) Act. The goal of this Act is to improve t he performance of elementary and secondary schools by requiring schools to make Adequate Yearly Progress (AYP), with all students meeting proficiency (as set by the individual state) by the 2013-14 academic year (12-Year Timeline). Statewide accountably through annual assessment also was mandated, requiring disaggregated results (i.e., poverty, race, ethnicity, disability, and limited Englis h proficiency) to measure the schoolsÂ’ effectiveness in teaching all children. The 12-year timeline set by the NCLB Act was established to enable states and school districts to conform to the legislation and raise student achievement to predetermined benchmarks. In an effort to monitor progress over time, an expected trajectory was mapped ou t, thus providing a slope depicting the start and goal points of st udent proficiency. This slope shifts based on yearly
2 assessment outcomes and represents the rate of change over ti me (how far the school is from meeting the goal). W hen the NCLB Act was passed in 2001, the slope illustrated a gradual yearly pr ogression toward the goal of student proficiency. In the 2007-08 academic year, 28.1% of schools in the United States did not maintain AYP. This is in comparison to 25.8% and 26% of schools failing to make AYP in the 2005-06 and 2006 -07 years respectively. As the requirement for the percentage of student s making proficiency increases, the difficulty of the task al so increases. As children enter kindergarten, they demonstrate variable levels of readiness that are dependent upo n childhood experiences during the preschool years. Some groups (e.g., low SES) are more vulnerable. Early childhood educators (preschool teachers of three and four year old ch ildren) are, therefore, under pressure to ensure that students ar e Â’readyÂ’ for school. The term Â‘readyÂ’ as it relates to education typically is def ined as the specific set of cognitive, linguistic, social, and motor skills that enable the child to assimilate the curriculum (Lewit & Baker, 1995). In rec ent years, there has been an increased interest in improving student readiness fo r kindergarten. In 2002, a constitutional amendment was passed in Florida, and subsequently signed by Governor Bush in 2005, requiring a free and voluntary pres chool program for all four-year-old children. This Voluntary Pre-kindergart en Education Program (VPK) is designed to prepare children for school by enhanc ing their pre-re ading, pre-math, language, and social skills. There ar e approximately 220,000 four-year-old
3 children who are eligible for VPK each y ear. As of February 2008, 123,857 four year olds were enrolled. Developmental Trajectories: Implic ations for Academic Achievement Academic and social-emotional compet encies are key contributors to healthy development and subsequent success in society. According to Ramey and Ramey (1998), a childÂ’s competencies increase steadily throughout his or her life to produce a pattern of typical development. This was depicted through their model illustrating the trajectories of children based on the quality of cognitive and social development. The basic premise is t hat the trajectory changes to illustrate a developmentally delayed course when cognitive and social competencies are deficient. As time passes, the gap between the typical and delayed trajectories increase, known as the zone of modifiability (Ramey & Ramey, 1998), or the area w here remediation attempts are implemented. The significance of this model is the author sÂ’ theory that exper iences in early childhood may alter childrenÂ’s competencies over time, therefore supporting the need for appropriate early preventi on and intervention services. Although intervention at any point in t he trajectory is beneficial to the child, it is the first five years of life that are critic al to development. Early experiences during this time fuel the neural connecti ons that lay the f oundation for language, reasoning, problem-solving, behavior, and emotional health (Getting Ready, 2004). Children are actively learning fr om the moment of birth through the various types of experiences the infant has with caregivers, which are ultimately related to all aspects of development (Ramey & Ramey, 2004). Research has
4 shown that children with developmental delays learn and benefit when they enroll in school; however, the rate of learning is not suffici ent enough to compensate for the entrylevel gaps, which often are in excess of 2 or more years (Ramey & Ramey, 2004). Efforts to close this gap and promote posit ive child outcomes must include several influences such as contributions of the family, neighborhood, and childcare setting (Getting Ready, 2004). Unfortunately, low student achievement tends to be persistent over time (Snow, Burns, and Griffin, 1998). Pers istent poor academic achievement has been identified as one of the primary fact ors leading to school drop out in a review conducted by the National Research Council (2001). It is important to begin intervention early, often prior to t he typical start of school for students atrisk. This has been especially notable for children coming from economically disadvantaged families. These children tend to begin kindergarten lacking readiness skills (Getting Ready, 2004; Ramey & Ramey, 2004). Intervention prior to the entrance of elementar y school addresses the maladaptive developmental trajectory, a trajectory that threatens future academic achievement. Implications of Children At-Risk Approximately 250,000 child ren between the ages of birth to three were identified having a developmental delay in 2001 (U.S. Department of Education) and consequently received Part C services ( early intervention services for infants and toddlers with disabilities provided under Individuals with Disabilities Education Act). Many children en ter the school system unprepared for the
5 demands and expectations set for them (Ramey & Ramey; 1998). Without the appropriate early intervention, t hese children become at-risk for low achievement, high retention rates, s pecial education placement, and drop-out (Ramey & Ramey, 1998; Dodge, Petit, and Bates, 1994). Additionally, the probability that these children will experience teen pregnancy, delinquency, unemployment, and social depe ndency later in life increases (Barrera, et al., 2002; Ramey & Ramey; 1998). There are num erous factors that contribute to a childÂ’s success or failure in school. These include parental involvement in education, family socioeconomic status self-regulation, and appropriateness of the school curriculum in relation to the childÂ’s needs, all of which impact the preschool and early elementary sc hool years (Stipek, 2001). The literature often refers to the aforementioned fa ctors in terms of risk and protective factors. Risk, in social science, refers to the likelihood of adversities occurring to an individual or a group based on the presence of one or more factors (Garmezy, 1994; Werner, 1992) For example, a child may be at risk for reading difficulties if the parents are illiterate and provide no enrichment activities in the home. However, if t he same child is enrolled in a preschool program with a strong emphasis on reading ac tivities and accompanies his or her cousin to the local library each Saturday, the risk is decreased. The latter scenario refers to protective factors, which serve as safeguards promoting adaptation and enabling the individual to resist the adver sity. Risk and protective factors are best consider ed within an ecological framework (accounting for
6 family, peer, home, school etc.) to apprec iate the various factors affecting a childÂ’s development. Academic achievement is a promi nent outcome measure utilized in examining the impact of risk and protective factor s. The NCLB Act has emphasized the importance of literacy development at the elementary school level. This has led to an increase in research on the development of readiness skills and early literacy development in pr eschool children. It has become clear that early literacy skills such as vo cabulary, letter recognition and sound/letter correspondence are good predictors of ch ildrenÂ’s reading abilities throughout their education (Getting Ready, 2004; Na tional Reading P anel, 2000; Snow, Burns, and Griffin, 1998). However, Ramey and Ramey (2004) reported that nearly one third of children entering kindergarten were Â“ not readyÂ” for the typical kindergarten curriculum. A sc hool district in west cent ral Florida reported that 30.3% of children who enrolled in kinder garten were Â‘not readyÂ’ to begin the kindergarten curriculum (Pinellas C ounty Schools Kinder garten Readiness Standards Report for 2002). This percent age increased to a range of 38% to 66.7% for the children in the school who were designated as Â“low socioeconomic studentsÂ” through their enrollm ent in subsidized child care programs (e.g., Head Start) and supplementa l educational services (e.g., Title 1). At a statewide level, the result s of the 2006-07 Florida Kindergarten Readiness Screener (FLKRS) illustrate the difference in readiness between children who attended VPK and children who di d not. Ninety-three percent of the children who completed the VPK program scored Â“ReadyÂ” for the kindergarten
7 curriculum as measured by the Early Chil dhood Observation System. This is in comparison to 84% of children who di d not participate in the program. When measured on the Dynamic Indicators of Basic Early Literacy Skills (DIBELS), 84% of the children who completed the VPK program scored Â“ReadyÂ” on the Letter Naming Fluency component, compared to 64% of children who did not participate. Initial Sound Fluency is another area of the DIBELS that is used in the kindergarten screening process. Sev enty-two percent of the children who completed the program scored Â“ReadyÂ” on this measure as opposed to 62% of children who did not participate in VPK. Not surprising, early literacy skills also tend to be more developed in young children who have been read to on a r egular basis by their caregiver and have been linked to increased academic achievement and later success in school (Child Trends, 2004, Ramey & Ra mey, 2004). Unfortunately, the occurrence of this daily, beneficial parent-c hild interaction was reported as being slightly over 50% for children birth thr ough five years of age, with 21% of children under the age of three being read to twice weekly or less. It is for these reasons that Ramey & Ramey (1998) argue that ear ly intervention is imperative in the efforts to prevent poor intellectual dev elopment for children who do not receive adequate stimulation at home. Academic-based tasks such as identifying letters and numbers are important when assessing student readiness (Ramey 2004). However, the academic behavioral competencies (managing emotions and behaviors, attending to the task, etc.) of the child are often of equal or greater significance
8 (Lin & Lawrence, 2003; Webster-Stratt on & Reid, 2004) when determining level of readiness for school. Johnson, Gallagher, Cook and Wong (1995) examined the views of kindergarten teachers regar ding the skills deemed critical for success in their classrooms and found that academic skills were not as high of a priority as originally hypothesized. Rat her, the skills highlight ed as high priority were self-help skills, understanding, follo wing classroom rules and routines, and working independently. Overall, 22 skills were listed, of which only 4 were academically oriented. The highest ranking developmental domain was the social domain with language ranked as a close second. Behavioral Competencies and Academic Achievement The relationship between appropri ate classroom behavior and student achievement is well established (Patters on et al., 1982; Frick et al., 1991; Hindshaw, 1992; Arnold, 1997, Arnold, e t. al, 1999; DSM-IV-TR, 2000; Squires, 2000; Nelson et al., 2003). Academic behav ioral competencies such as selfcontrol, attending to, and remaining on, task and following directions are associated with high academic achiev ement. Poorly developed academic behavioral competencies may compromise academic achievement and lead to subsequent antisocial behavior (Child Trends, 2004). Conroy and Brown (2004) reported the prevalence of significant social/emotional delays in preschool childr en. Twelve to sixteen percent of 1 and 2 year olds (37% of these children c ontinuing along a maladaptive trajectory throughout their preschool years) and 25% in 2-3 year olds (50% of this group remaining on the maladaptive track) demons trate these delays. Data suggest
9 that developmental delays in the social /emotional domain are widely associated with problematic transition (difficult y adapting to the expectations and boundaries) into the school setting (Ri mm-Kaufman, Planta, and Cox, 2000). The Kindergarten Readiness Standar ds Report (2002) reported that a significant number of children who were enrolled in publicly funded childcare centers experienced difficulties in lit eracy development and demonstrated delays in social/emotional developm ent. This report indicated that 30% of the children were unable to follow classroom rules, 25% were unable to handle a problem acceptably, and 15% did not interact appropriately with peers or adults. The need for augmented focus on childrenÂ’s literacy and social and emotional development is clear. The co-occurrence of poor academic achievement and behavior problems often adversely impacts student achievem ent in reading (Farmer and Bierman, 2002; Hindshaw, 1992). Arnold et al. (1999) concluded t hat the more severe the behavior problem, the poorer the liter acy achievement. Kamps (2000) and Kauffman (2001) reported that 60% of ch ildren who exhibited behavior problems also had academic difficulties, predom inately in the area of reading. Furthermore, children who did not develop basic literacy skills before they entered kindergarten were 3 to 4 times more likely to drop out of school in later years (Kamp, 2003). Kamps (2003) r eported that ther e was an increased occurrence of disruptive behaviors negativ ely impacting instruction and student learning as well as an increased number of students who failed to acquire competent levels in reading. Although the relationship between deficits in
10 reading achievement and externalizi ng behavior problems has been well established (Torgeson, 2000; Hinshaw, 1992; Frick et al., 1991), no clear directionality has been determined. W hat is clear is that academic underachievement and behavior problem s become less responsive to interventions over time ( Hindshaw, 1992; Kazkin, 1987). Academic Achievement and Behavioral Com petencies: A Circular Relationship The relationship between conduct pr oblems and academic achievement is circular in nature meaning that it is difficult to ascertain where the problem begins. Arnold (1997), Arnold et al. ( 1999) and Stipek (2001) suggest that conduct problems limit the childÂ’s opportunities to learn. For example, if a child is either engaged in or being reprimanded for inappropriate behavior, the amount of academic engaged time is subsequently reduced. A cycle develops whereby continual behavior problems contribute to an increase in negative perceptions regarding school, decreasing motivation, wh ich then augments the childÂ’s poor achievement, ultimately fueling the behavio r problems. This pattern typically becomes stable over time, making the cycl e less responsive to interventions. The second perspective examines the presence of poor academic skills in preschool or early elementary school, wh ich consequently exacerbates behavior problems (Stipek, 2001). In this scenar io, a child may engage in inappropriate behavior to mask the academic difficulty or to express frustrat ion with the task. Teachers often contribute to the circular relationship by providing fewer learning opportunities (i.e., less likely to call on, question or provide information) to children who display behavior problems. Th is reinforces the child for avoiding
11 the aversive academic tasks, while at the same time limits the much needed instruction to increase skills This cycle has been found in preschool environments as well, resulting in ch ildren learning to become disengaged from the academic environment prio r to entering formal schooling (Arnold et. al, 1999). Arnold (1997) reported that externa lizing behaviors predicted academic skills and vice versa, with the relati onship between the two strengthening with age. Increased levels of externalizing behaviors were reported for children who experienced early reading difficulties. According to Hindshaw (1992), the appearance of the behavior changes over time, pairing inattention and hyperactivity to childhood underachi evement and antisocial behavior and delinquency to adolescent underachievemen t. When controlling for prosocial behavior, Caprara (2000) found that early academic achievement did not predict later achievement; rather prosocial behav ior strongly predicted subsequent levels of achievement when holding early ac hievement constant. In summary, academic achievement is associated with the academic behavioral competencies that complement learning (Raver and Kn itzer, 2002). The childÂ’s academic achievement and experiences with success or failure influence the foundation for future behavior and subsequent achievement as they affect the childÂ’s conduct and motivation. Statement of the Problem Educators across America have been challenged with the task of increasing the effectiveness of schools thr ough the provisions of the NCLB Act. Early Reading First is a nationwide effo rt developed to improve the effectiveness
12 of instruction in early literacy in early childhood centers that serve low-income families. Scientifically based reading res earch was used to develop instruction to enhance language and cognitive skills and to improve the early reading foundation that prepares children for kindergarten and beyond (U.S. Department of Education, 2008). Children with low academic skills are at risk for later academic difficulties (Ramey & Ramey, 2004; Stipek, 2001; Sn ow, Burns & Griffin, 1998), and early emergent behavior problems in presc hoolers are likely to continue on a maladaptive trajectory (Hindshaw, 1992; Pa tterson, et al., 1992). However, these children are not predes tined for failure. Rather the research clearly supports the need for systems change in early education pertaining to policy and practice in an effort to circumvent t he maladaptive trajectory. However, a substantial void remains with regard to which developmental domain should be the focal point. That is, while the case for early intervention is provided, it remains unclear as to which risk factors emerge first, behavior issues or poor achievement. Support has been established fo r the circular rela tionship between the two, with primary attention on el ementary aged children and adolescents. Several limitations are evident in the lit erature to date. First, the research examining the relationsh ip between academic achievement and behavioral competencies has not focused on preschool children. Second, many studies quantify academic achievement by obtaini ng normative scores on standardized measures. A more appropriate measure of academic achievement is curriculumbased measurement (CBM), which is a more sensitive method for gathering
13 information regarding student performance based on the childÂ’s curriculum. Nichols et. al (2004) supports the noti on that the use of CBM data to guide instruction results in greater growth in phonemic awareness skills despite gender, socioeconomic status (SES), preschool ex perience and race, a characteristic that is especially applicable in this line of re search. Third, whil e the literature has addressed the role of an individual medi ating factor (SES), there does not seem to be a line of research examining multip le factors and their potential roles in the achievement-behavior relationship. Purpose of the Study The purpose of this study is twofold. First, it will exam ine the relationship between literacy development and behavior diffi culties in preschool children. Second, the role of within-child protecti ve factors in literacy development will be explored. It is anticipated that data gleane d from this study will contribute to the literature as well as provide relevant information regarding the potential avenues for early intervention services. Research Questions 1. How does positive and negative classr oom behavior contribute to the rate of literacy development am ong preschool children? 2. What factors (i.e., gender, race, SES, teacher experience, classroom environment, class size) contribute to the rate and levels of literacy development for children i dentified with typical or challenging behaviors? 3. What differences are there between literacy development in children with challenging behaviors who have hi gh scores measuring within-child
14 protective factors in comparison to children with chal lenging behaviors who have low scores measuring withi n-child protecti ve factors? Definition of Terms The terminology and concepts utilized in the current study are presented in Table 1. The purpose of this table is to ensure the reader becomes familiar with terms used in the Early Lear ning Opportunity (ELO) grant. Table 1. Definition of Terms Concept/Term Definition Teacher Participant Early childhood educators who participated in the program evaluation com ponent of the ELO grant. Student Participant Preschool children who were taught by the teacher participants and participated in the program evaluation com ponent of the ELO grant. ELO Head Evaluators Three doctoral candidates from the University of South Florida who were hired to collect and manage the data obtained from the ELO evaluation activities. Home Socioeconomic Status (SES) The median household income within a geographical region (based on zip code) in which the child resides. Site SES The median household income within a geographical region in which the childcare site is located. Classroom Environment Represents the litera cy-related environment (variety of books and writing materials easily accessible to the child) and teacher-child interaction (use of open-ended questions) within the preschool classroom. These data are based on a classroom observation checklist (e.g., ELOC) used in the program evaluation component of the ELO grant. Early Literacy Development Represents preschool achievement in Expressive Language and Phonemic Awareness as measured by the three subt ests of the Individual Growth and Development Indicators (IGDI).
15 Table 1. (Continued) Definition of Terms Within-child protective factors Characteristics that enhance resiliency and discourage adverse outcomes in preschool children. These are represented by the Total score on the teacher-compl eted Devereux Early Childhood Assessment (DECA) questionnaire, which is comprised of three subtests (i.e., Initiative, Self-Control and Attachment). Behavior The academic behavioral competencies (e.g., self-control, attending to and remaining on, task, and following directions) t hat typically aid in academic achievement. The Behavior Concerns score on the DECA represents high or low levels of behavioral competencies of the child participant.
16 CHAPTER TWO REVIEW OF THE LITERATURE According to Stipek (2001) and Arnold e t. al (1999), childrenÂ’s long term academic success is highly predicted by their academic skills as they begin school, with academic development beginning long before they enter formal schooling. Stipek (2001) linked academic achievement in first grade to high school completion, suggesting that lo w academic performance in the earlier grades leads to low performance in s ubsequent grades. Howse, Calkins, Anastopoulos, Keane, and Shelton (2003) stated that childrenÂ’sÂ’ academic performance remains Â“extremely stableÂ” after the first grade. Specific to reading, Al Otaiba & Fuchs (2002) reported in a re view of the literat ure that children experiencing reading difficulties in firs t grade remained poor readers in fourth grade, with the gap between these childr en and their fluent peers widening over time. Specific skills associated with th is link included receptive and expressive language ability, both of wh ich have been correlated wit h reading ability in the first few years of elementary school (Pikulski and Tobin, 1989; Scarborough, 1989). To expand on the importanc e of early academic su ccess, Stipek (2001) reported that the performanc e of elementary children se gregated in low-skilled reading groups is substantially less t han those children placed in the high-skilled groups. This difference in achievement gains was explained by the teaching
17 methods employed within each of the two groups. It was noted that the focus in the lower-skilled group c entered on decoding whereas the higher-skilled group was involved in more meaningful questi ons and opportunities to connect reading to personal experiences. The former group will likely have difficulty catching up to the latter group due to this ability group placement method (Stipek, 2001). Another issue regarding children enter ing kindergarten with low skills is the common practice of grade retention. Although geared to assist children in Â“catching up,Â” the research has clearly pr oven the aversive affects, including a higher drop out rate. Mo re specific, retention for one year leads to a 50% likelihood of dropout while retention for a second year has a dropout rate of 90% (Baker, et al., 2001). Despite this st atistic and lack of research supporting retention (Jimerson, 2001), student s continue to be retained based on inadequate academic progress. Furthermo re, in 2002, the Florida legislature mandated that unless they meet Â“good c ause exemption,Â” third graders who obtain a level 1 on the Florida Comprehens ive Achievement Test (FCAT) would not be promoted to the fourth grade. This demonstrat es that grade retention continues to be a widely used technique and further supports the view that children who enter school with poor academ ic skills are at a disadvantage. Children demonstrating behavioral di fficulties also are placed at a disadvantage in comparison to their more typically adjusted peers. In one study, Howse et al. (2003) explored the relati onship between emoti onal regulation in early childhood and emotional/behavio ral self-regulation and academic achievement in kindergarten. Results s uggest that children who have difficulty
18 with emotional regulation in early child hood continue to experience challenges with regulation in the kindergarten classr oom. Furthermore, children with higher emotional and behavioral regulation demonst rated higher achievement scores in literacy, math and listening comprehension. Howse et al. (2003) cited Blair (2002) in explaining these results, i ndicating that children who experience difficulty with emotional regulation ar e not able to simultaneously engage in problem solving tasks and may withdraw in response to anxiety evoking situations, thus interfering with their ab ility to remain academically engaged in the classroom. Although a link has been established between academic achievement and behavior difficulties in elementary aged ch ildren and adolescents, there is little research examining these variables am ong preschool children. For example, reviews of the literature conducted by Al Otaiba & Fuchs (2002) and Nelson, Benner, & Gonzalez (2003) revealed evidence supporting the link between reading deficiencies and behavio r problems in children. According to Nelson et. al (2003), sixty to 100 percent of childr en with behavior disorders also have poor reading performance, with thr ee out of four espousing l anguage deficits specific to phonological processing. Both of thes e statistics are known to be stable or increase over time. Lane, Wehby, Menz ies, Gregg, Doukas, and Munton (2002) examined the effectiveness of early litera cy instruction in first grade students who were at risk for anti-soci al behavior. Participants were selected based on their documented resistance to previous school-wide intervention attempts, which included a literacy and behavioral component. Children in the study participated
19 in 30 small group lessons, yielding a tota l of 15 hours of intervention across nine weeks. Effect sizes from the study i ndicated strong progress in decoding skills for all students, with increases in oral reading fluency for three of the seven children. Effect sizes also revealed si gnificant decreases in disruptive behavior within the classroom setting. Torgesen et. al (1999) st ated that the two best pr edictors of a childÂ’s response to intervention in relation to reading achievement is the home environment (i.e., family income and par ent education) and beha vior problems. Torgesen et. Al (1999) and Torgesen (2000) have found that behavioral issues impede the childÂ’s ability to benefit from intervention, even in a one to one setting. Given that the ma jority of instruction is presented in a whole group format, it appears that these children ar e often not profitin g from the muchneeded academic material. However, Barre ra et al. (2002) reported that the implementation of comprehensive intervent ions to prevent behavioral problems have a favorable impact on social development when it includes an academically-based instructional com ponent. With the data stacking up in support of the achievement-behavior linkage, the focus of this literature review now turns to the underlying factors that influence development in these areas deemed essential for success in society. Risk Factors Associated with Academic and Social-Emotional Development The literature supports the aversive e ffects of risk factors in relation to child development (Atzaba-Poria, Pike and Deater-Deckard, 2004; Schulz & Shaw, 2003; Stipek, 2001; Ziegler & St yfco, 2001), indicating that 50% of
20 kindergarteners in the United States are from families with one or more risk factors for school failure (Landry, 2002). Thirty-three percent of these children who espouse only one risk fact or will obtain reading score s in the bottom quartile (Landry, 2002). Researchers examining mu ltiple risk factors have developed an argument pertaining to the ef fects of each risk, debating whether cumulative risks adhere to a threshold (Rutter, 1979) or li near (Sameroff et. al 1998) model. More specific, a threshold model implies that as risk factors are added the effect multiplies as opposed to a linear model where an increase results in a steady increase in problematic outcomes. A recent exploration of these models indicates a substantial increas e in problem behaviors as a result of exposure to increased risk factors, particularly if ex posure was at an early age; however, a threshold effect was not support ed (Appleyard, et. al, 2005). Flanagan, Bierman and Kam (2003) espous ed a slightly different model, although maintained a similar construc t. These authors suggested that aggression, hyperactivity-inattention, and social skills deficits are developmentally intertwined, with elevati ons in any one increasing the probability of elevations in the remaining two. The authors found that these characteristics assessed at school entry predicted later school difficulties in the behavioral, academic and social adjustment domains with increased levels of maladjustment being contingent upon the presence of more than one behavioral issue. In a more specific study regarding risk factors relating to academic and behavioral development, Kamps (2003) cond ucted a 3-year longitudinal study examining the literacy growth and risk fact ors of young children, the majority of
21 whom were from culturally di verse and economically disadvantaged backgrounds. Screenings were conducted us ing the Dynamic Indicators of Basic Early Literacy Skills (DIBELS) to asse ss literacy skills and the Systematic Screening for Behavior Disorders (SSBD) to tap into early behavior problems. Utilizing hierarchical linear modeling (HL M) the researchers found that early risk influenced studentsÂ’ progress in reading fluency over time. Of greater importance is that children who demonstr ated the greatest difficulty becoming fluent readers were those who initially possessed both academic and behavioral risks (Kamps, 2003). The next group to demonstrate reading difficulties was the students with academic risks only, follow ed by those students with behavioral risks only. In comparison to the general population, Kamps (2003) found that children in early elementary school fac ed with a single risk factor, behavior or academic, performed lower in reading fluency assessments. Kamps (2003) concludes that children who begin school with lower skill levels are least ready and demonstrate slower progress over time; however, effective curricula and frequent progress monitoring can fa cilitate literacy success. Implications of these research studies clearly indicate an intense need for early intervention services geared toward the reduction of risk factors in early childhood years in an effort to curb the potential for maladaptive pathways. Several risk factors appear prominent durin g this critical developmental phase including family-based, community-based and childcare centerbased. For the purposes of this review, the latte r two areas will be discussed.
22 Ecological Systems Theory Bronfenbrenner (1979) fo rmulated the ecological approach to human development, reasoning that children are in fluenced by their family, with the wellbeing of the family, in turn, influenced by the community in which they reside. Following Bronfenbrenner Â’s Ecological Systems Theory where the child is viewed as developing within a comp lex system of relationships affected by multiple levels of the surrounding environment, T he Child Mental Health Foundations and Agencies Network (FAN) reported risk factors associated with academic and behavioral problems at the beginning of school This nested system includes the microsystem, mesosystem and exosystem per taining to the childÂ’s dynamic and ever-changing environment. The discussion below will categorize each of the risk factors based on their pl ace within this theory. Microsystem The innermost layer, the Microsyst em, encompasses the activities and interaction patterns in the childÂ’s imm ediate surroundings, with all relationships being bi-directional and reciprocal. In appl ying this layer to young children there are two distinct areas including Family /Peers and Childcare, in addition to the general identifying characte ristics of the child. Effects of gender on achievement In their extensive review of the literature, Diamond and Onwuegbuzie (2001) reported the trend of gender differences over time in the area of readi ng. More specific, the authors revealed that females obtained higher levels of reading achievement than their male counterparts while also espousing a more po sitive attitude toward the activity of
23 reading, particularly recreational. In addition, fewer females were referred for leaning problems in reading. Accord ing to Diamond and Onwuegbuzie (2001) the gender difference gap in reading achiev ement has been consistently found to be stable over time with differences in reading attitudes widen ing with age, both favoring females. Although gender diffe rences were not found when children with diagnosed learning di sabilities were examined, disparities were noted when children were informally identified by t heir teacher as struggling academically (Morgan and Dunn, 1988). Notably, girls and boys were described as Â“invisibleÂ” and Â“visibleÂ” respectively, based on their response to academic problems. Morgan and Dunn (1988) sugges ted that the former gro up presents themselves as less noticeable and more shy, often in an attempt to hide th eir uncertainty. This type of behavior does not disrupt the classroom environment, unlike the boy-specific Â“visibleÂ” behaviors (i.e., dem anding teacher attention, acting out), therefore increasingly the likelihood of girlÂ’s being under-identified as experiencing learning difficult ies. In examining this premise further, Stowe, Arnold and Ortiz (2000) conduc ted a study of the rela tionship between language development and disruptive behaviors in preschool children based on gender. Their findings support the notion of Â“invis ibilityÂ” in that preschool boys with deficient language skills were more likely to be disruptive than girls possessing similar deficiencies. In terms of referrals for special services, the findings indicated that perceptions of problemat ic behaviors rather than their language development predicted whether the child would be referred, which typically was for speech and language ther apy. Quantitatively, Stowe, Arnold, and Ortiz
24 (2000) noted that children who had deficient language scores but demonstrated typical behaviors had a 2.3% chance of being referred whereas children who fell within the mean range for language but espoused behaviors 2 standard deviations above the norm had a 50.5% chance of being referred. Girls with early academic difficultly, then, are at a disadvantage when it comes to early intervention. Effects of gender on behavior Similar discrepancies have been documented with gender and behavior issu es, with males demonstrating more behavior difficulties within the educational se tting. A review conducted by Green et al. (1996) suggests that alt hough the prevalence rate for childhood psychological disorders is the same for gi rls as it is boys, the latter population tends to receive more mental health serv ices. Further research has illustrated the differences in topography between ma le and females behavior, noting that females characteristically display more internalizing issues whereas males are typically more externalizing (Gr een, Clopton, and Pope, 1996; MacMillan, Gresham, Lopez and Bocian, 1996; Atzaba-Poria, Pike and Deater-Deckard, 2004). Nelson et al. (2003) revealed that elementary aged girls received higher ratings than did boys on teacher report forms when items tapped into physical symptoms and fears. Ov erall, behavioral problems among girls are apparent and, in some cases, significant; howev er it has been repeatedly suggested that maladaptive behaviors must be present to a greater degree than boys in order to be appropriately identified and granted serv ices (Nelson et al., 2003). Green et al. (1996) found that teachers based refe rrals on type of behavior such that boys
25 or girls with externalizing behaviors were viewed as needed services than either sex demonstrating internalizing behaviors. Furthermore, this study revealed that teachers were significantly less likely to refer children for behavioral issues if the student is doing well academically, a pattern also more prevalent in girls. This is consistent with other research (e.g., MacM illan et al., 1996) indicating that males routinely demonstrate severe exter nalizing behaviors combined with poor academic achievement, a trend not seen in their female counterparts, whose achievement is not compounded with behav ioral issues. It, then, may be hypothesized that gender differences exis t due to the higher incidence of internalizing behaviors and academic com petence among girls in comparison to boys. Effects of minority status on achievement Disparities between minority and white students have been noted for decades and have been found as far back as the 1967 Report of the US Commission requested by President Johnson regarding racial isolation in the schools. T he findings of this research clearly link economic status to achievement with chil dren from low SES families exhibiting greater difficulties in core areas such as reading and mathematics. Although the children within these two groups demonstr ated interest and aspirations in high achievement, the children in high SES environm ents were able to attain this goal due to increased opportunities and s upport (US Commission on Civil Rights, 1967). More recent, Diamond and On wuegbuzie (2001) conducted a study examining ethnic differences as it relate s to reading achievement. Although the researchers noted that overall reading achievement for Black students have
26 improved since the 1970Â’s (whereas Whit e students remained stable), they continue to exhibit significantly lower performance levels. According to the NCES (2000), sc hools in America are becoming increasingly heterogeneous when consider ing demographic information such as race and ethnicity. Unfortunately, the schools are not prepared to handle such changes, leading to inequalities in academ ic achievement (Meece and KurtzCostes, 2001). In a review of the liter ature regarding ethnic minority children, Meece and Kurtz-Costes (2001) indicat ed that the prominent focus has been the difficulties minority groups have experi enced in conforming to the school environment that tends to favor the main stream culture. More specific, the authors noted that in an effort to resi st the mainstream value of education minority groups often reject achievemen t, devaluing the importance of academic success. Locally, data collected by the Florida Department of Education pertaining to the minority status of those children affected by FloridaÂ’s third grade retention mandate indicates a gross overrepres entation of Blacks and Hispanics. Quantitatively, Black students constitut ed 24% of the thir d grade population in 2002-03, while Hispanic children account ed for 22% of the population. The retention status for these two minorit y groups was 39% and 29% respectively. This is in comparison to their White peers who made up 49% of the third grade population out of which only 28% were reta ined. These data clearly supports the differences in reading achievem ent in minority students.
27 Differences also were revealed in preschool aged children with reference to child care enrollment. More specific, Magnuson an d Waldfogel (2005) noted that Black and Hispanic children are more likely to be enrolled in center care programs offering subsidies than white children, with such programs often providing lower-quality care. As such the achievement gap is maintained. Disparities have been revealed outside of the formal child care environment, with Black and Hispanic children typically co ming from homes with fewer reading and other educationally relevant materials, and being read to less frequently (BrooksGunn and Markman, 2005). Relating this to future performance, the authorÂ’s reported the achievement gap between blac k and white children is reduced by 3 to 9 points when parenting is accounted for. Referring back to BrofenbrennerÂ’s theory, it is obvious that factors and Â‘systemsÂ’ cross each otherÂ’s boundaries, making clear associations between variables difficult. In re sponse to the challenges portrayed in the research, Meece and Kurtz-Costes (2001) provided seve ral limitations and future directions for conducting research with the minority population. First, t hey point out the entangled issue of SES and ethnicity, repor ting that many samples representing minority families also are categorized as economically disadvantaged. In these situations it is challenging to ascertain whether the minority status variable is significant or whether t he mediating effects of SES are contributing to the findings. Second, using the White populat ion as the norm with which to compare other groups does not allow fo r cultural beliefs. In ot her words, is the child behaving in a manner that is consistent with his/her cultural /ethnic background in
28 which the behavior is acceptable? Thir d, the authors noted t hat there is less emphasis on protective factors leadi ng to academic success when examining minority status. Fourth, t here appears to be a lack of focus on outside contexts (other ecological factors) that are important to child Â’s success. Finally, Meece and Kurtz-Costes (2001) discuss the fa ilure to consider developmental perspective of the child in re lation to their achievement. Effects of minority status on behavior Several research studies have been conducted examining the effects of race and behavior problems in the school setting. While they all take a diffe rent path in explori ng this issue, the results all indicate a clear difference in the trajectory between minority children and their Caucasian peers. In detail, it was revealed that Black children display more externalizing behaviors than Wh ite children. (MacMillan et al., 1996; Epstein, March, Conners, and Jackson, 1998), with teachers rating Black students as demonstrating more behaviora l difficulties accompanied by less behavioral competencies during the first two years of formal school as compared to Caucasian children (Sbarra and Piant a, 2001). All studies appreciate the influence of socioeconomic disadvantage as playing a mediating role in the negative trajectories, addressi ng the high correlation betw een minority status and SES. However, McLeod and Nonnemaker (2000) found that although Black children were rated as displaying more delinquent behaviors t han White children, the effects of poverty increases the risks of such behaviors to a greater degree in White children. Regardless, it is cl ear that minority st udents are at-risk for
29 maladaptive social-emotional development, a course that has been proven to become stable over time. Effects of family socioeconomic status on achievement Poverty is a major risk factor for future school failure, as it may affe ct the rate of learning, which then influences intelligence and acade mic success (Sattler, 1990). More specific, children exhibit few differences in intellectual functions during the first two years of life related to race or social class; however, it is at ages 3-4 that the differences not only become apparent but remain stable through school years (Sattler, 1990). The 1999 Census data reported that 19% of children under 18 were growing up in a family below the pov erty line. The statistics for minority groups such as African Americans and Hispanics yield numbers of 35% as compared to 14% of Whites who were living in poverty. Academic achievement was discussed by Stipek (2001) in relation to socioeconomic status, supporting previous research findings that children from economically disadvantaged families, on average, begin school with poorer academic skills then their economically adv antaged peers. It is important to reiterate, however, that SES does not direct ly effect achievement, rather it serves as a mediator between achievement and t hose factors that are associated with low SES (i.e., parental involvement, stre ss, expectations, availability of resources, stimulating environment), with these effects as more pronounced during early childhood (Baydar et al., 1993). Nichols, Rupley, Rickelman and Algozzi ne (2004) found that children who came from a low socioeconomic ba ckground, had little or no preschool
30 experience and who were of Hispanic ethnici ty were at a greater risk for not developing phonemic awareness and concep ts of print in kindergarten as compared to their peers. Ov erall, it was noted that a ll children made gains with children of low SES backgrounds achieving lower scores on curriculum based assessments. Orr (2003) suggests that wealth im pacts academic achievement due to levels of financial and human capital, thus providing a rationale as to why the gap in test scores between African Americans and Caucasians exist. The argument lies on the theory that wealth span s beyond income and encompasses other assets, including interests and dividends that enable families to positively contribute to their childÂ’s achievement. More specific, Orr (2003) and Entwisle et al (1997) hypothesized that families ar e more likely to devote earned income to educationally rich items (i.e., books, com puters, private schools, tutors, and visits to museum/zoo/concerts) when they have a financial back-up. These additional resources then increase the childÂ’s learni ng time and aids in the development of academic skills. The analysis of data derived from the National Longitudinal Survey of Youth supported this hypothesis; therefore solidifying the differences in academic achievement despite comparable levels of income. Differences between students from economically disadvantaged and advantaged families were further demonstrated when achievement was measured after a summer break when schoo l was not in session. To preface this, it was noted that bot h groups of students exhibited similar achievement gains during the school year. Heyns (as cited in Burkam et. al, 2004) suggested
31 that schools act as mediators duri ng the school year, thus providing disadvantaged students with the cognitive ex periences they lack from their home environment. However, access to these ex periences is limited for these children during the summer months, placing thei r counterparts at a critical advantage (Burkam et. al, 2004; Ramey and Ramey, 2004; Alexander et al., 2001, Snow, Burns and Griffin, 1998). In addition, the quality and content of parental interactions appear to play a role in that higher SES parents tend to engage in more cognitive thinking sk ills (Burkam et al., 2004). Another argument provided by Stipek (2001) regarding SES and achievement is the skewed expectati ons held by teachers in the lower elementary grades. In this line of res earch it is posited that teachers based educational decisions and learning opport unities according to their perceived notion of the childÂ’s ability level. Initia lly, this has been found to be the case in relation to low achieving students, wher eby teachers placed children in lower Â“tracksÂ” or provided them with easier/less work. Unfortunately, a trend also has been noted illustrating lower expectations for children considered as economically disadvantaged. Part of this circular model includes the stability of teacher expectations in that attention may only be given to evidence confirming the teacherÂ’s belief (Entwisle & Hayduk, 1982 as cited in Stipek, 2001). An example provided to demonstrate this is a case where the teacher does not adjust the academic assignment according to student performance (i.e., doesnÂ’t realize that the child is reading below skill level because he/s he does not provide a more challenging book to read). The consequence in these situations is the
32 attenuation or restriction in learning opportunities and progress in developing skills (Stipek, 2001), which has been previously discussed as a possible precursor in the development of behavior problems. Providing a different perspective of the relationship between SES and achievement, Gregory, Williams, Baker and St reet (2004) explored the roles of social class, home-based resources and classroom approach on early literacy success for preschool children. In thei r longitudinal study, the authors followed classrooms from three Brit ain schools whose socio-econ omic composition varied extensively and found that progress wa s a product of several interacting components. For example, the demogr aphics at one school were described as low SES, low percentage of minority students, low levels of parent involvement, and low home-based resources. This, coupled with a child-centered teaching approach focusing on socialization of the children as opposed to academic performance, led to the poor literacy succe ss at the end of the school year. In contrast, a school deemed as rich in home-based resources, parental involvement, and SES excelled in litera cy development. Although the teacher approach in this school also was child-c entered, the families supplemented the insufficient literacy curricu lum through extensive parental involvement and tutors. In this situation the children were essentia lly bringing skills to school to Â“practiceÂ” them as opposed to learning new skills The literacy development of these children far exceeded the national aver age as they mastered the alphabet and began reading books. The third school wa s characterized as consisting of predominantly minority families with the lowe st levels of economic resources as
33 compared to the previous two schools. The classroom approach was described as having a strong academic focus with hi gh expectations. Interestingly, this school exceeded the literacy success of t he first school, attaining scores that were comparable to the national average. Noteworthy is the performance of the second and third schools given their successe s in light of the varying degrees of resources and academic supports. These fi ndings indicate that while low SES can impact the academic development of children, literacy success can be evident through a strong school-based cu rriculum and through parental support of weak curriculaÂ’s. Effects of family socioeconomic status on behavior One of the most commonly identified demographic family va riables that is related to behavior problems in children is low socioeconomic status (SES). According to studies reported by Huaqing Qi (2003), the pr evalence of behavior problems has been noted at 3-6% in the general child populat ion, whereas a 30% incident rate has been associated with low-income presc hool children. Although low SES does not cause severe behavior problems, numerous studies (Huaqing Qi, 2003; Frick et al., 1989; Haddad et al., 1991; Rutte r, 1985; Behar & Stewart, 1984) have found that this characteristic is associated with its occurrence. As always, it is important to note that it is not low SES alone, but low SES in combination with other variables such as maternal antisoc ial personality, low family cohesion, and high family conflict (Frick et al., 1989; Hindshaw, 1992; Schultz & Shaw, 2003) that is associated with the development of disruptive behavior disorders. This finding suggests that low SES may be a medi ating variable in that socioeconomic
34 disadvantage places the child at hi gher risk for the development of behavior disorders when low SES is combined with ot her variables (e.g., parental discord, aversive parent-child interactions). Due to the strong interconnected relationship between these variables, a causal relationship between SES and childhood behavior disorders cannot be assumed. According to Gautheir (2003), physical aggression is found more often in children who were raised in low SES environments. On the same note, we k now that not all children from poor families develop aggressive tendencies, leading to additional research revealing that family characteristics account for t he majority of the variance (53% versus 3% variance in low and high SES families respectively). Additionally, environmental conditions can play an integr al role in the childÂ’s tendency to engage in destructive injurious behavior (B erk, 2000). This hostility is found more often in environments that are tense and competit ive rather than friendly and cooperative. Further, these types of environments are more common in poverty-stricken neighborhoods with a wide range of stressors (poor quality schools, limited recreational and empl oyment opportunities, negative adult role models). Relatedly, children raised in these environments have greater access to deviant peers and enrollment into gangs Among low-income, ethnic minority children, such neighborhoods predict aggression beyond family influences (Kupersmidt, as cited in Berk, 2000) and place children at risk for severe emotional stress, deficits in mora l reasoning and behavior problems. Social class differences also are noted in the way parent raise their children. More specific, Berk (2000) r eports that lower-income families tend to
35 use physical punishments and harsh command whereas middle-class families use more warmth, explanation, and verbal praise. In addition, parents who work in skilled/semi-skilled occupations tend to place a high value on external characteristics such as obedience, neatness and cleanliness. This is in contrast to parents in professional occupations who tend to emphasize inner psychological traits including curiosit y, happiness and self control. These differences are hypothesized to be a result of the different lif e situations that these parents encounter. For example, low income families may feel a certain degree of lack of power outside the home w here they are required to follow the rules and obey people in authority. They then duplicate this relationship at home with them as the figure in power. Additional stresso rs of meeting basic needs due to finances (i.e., food, shel ter, clothing) also play a role in the amount of energy and attention the parent is able to devote to the child. Effects of class size on achievement Children who attend smaller schools have been found to demonstrate an increase in student achievement (Lee & Loeb, 2000), especially in schools with a high enrollment of minority students (Lee & Loeb, 2000; Ready, Lee, & Welner, 2004; Finn, Gerber, & BoydZaharias, 2005). Nye, Hedges and Kons tantopoulos (1999) s uggested that small classes had immediate effects on academ ic achievement as well as lasting benefits, according to their 5 year follow up study. In examining the reading and math scores of students over time, River a-Batiz & Marti (1995) revealed that there was approximately a 2% to 9% difference between performances on proficiency tests of children in overcr owded, low-income schools as compared to
36 non-overcrowded, low income schools, with the former obtaining more failing scores. According to Achilles (2005), th is can be explained through the teacherÂ’s ability to use good pedagogy as well as appropr iately address diversity, inclusion and assessment within the classroom. Blatchford et. al (2003) examined the ef fect of class size in the UK through a longitudinal analysis of children as t hey were followed from their kindergarten year through the end of second grade. No t surprising, the study suggests that smaller class size is related to an incr ease in teaching time and a greater quality of interactions between teachers and st udents. Literacy development appears to be the academic area most affected by cl ass size, possibly due to the level of support the teacher is able to provi de. Teacher read-aloud and individual support during independent reading occurs to a greater ex tent in smaller classes. Blatchford et al (2003) concludes that ch ildren who are most in need with respect to literacy development will benefit the most from smaller class size. Similarly, this classroom composition has positive effects on children of low ability or who come from economically disadvantaged fa milies (Cooper, 1989, Achilles, 2005). National Institute of Child Health and Human Development (NICHD) Early Child Care Research Network revealed simi lar findings linking smaller class size to increased educational and emotional support and increased literacy skills. However, as stated by Snow, Burns and Griffin (1998), it is the quantity and quality of teacher student interactions t hat are crucial in student achievement, and although a large class size may hi nder these interactions, they do not necessarily improve as cla ss size is reduced.
37 Effects of class size on behavior Similarly, behavior management within the classroom setting also has been found as positively affected by low teacherpupil ratio. The NICHD, in addition to a review from Achi lles (2005), suggest that improvement in studentsÂ’ behavior when enrolled in smaller classes may be attributed to the increased opportunities to participate in Â“learning communities,Â” where they develop important prosocial skills and are more actively engaged. Effects of attendance on achievement. A longitudinal study examining the relationship between school absences in elementary school and reading achievement was conducted by Easton & Engelhard (1982), revealing a negative correlation. That is, as student absenc e rates increase, the performance on both teacher assigned reading grades and standardi zed test scores decrease. These findings were based on a regression analysis, which controlled for variables such as previous achievement. More recently Moonie, Sterling, Figgs, Castro (2008), reported a negative impact of absenteeism on standardized tests. Although this study focused on children with asthma, t he analyses controlled for this health issue revealing no overall difference between children with and without the condition. Utilizing a two stage least s quares format, Gottfried (2009) also explored the relationship between a ttendance and achievement, confirming the aforementioned findings. Despite the long standing interest with the impact of student attendance on academic achievement, focus has been on the elementary school years and upward, thus leaving a question regarding the link in preschool.
38 Effects of teacher experience and education on achievement and behavior. In examining the educ ational level of teachers within low versus high poverty schools, as measured by percent age of free/reduced lunch status, the National Center for Educ ational Statistics (2000) revealed that teachers employed in the former schools are less likely to have a masterÂ’s degree than teachers in the latter school. Further, Darling-Hammond (1999) indicated that certification and higher degree levels are significantly and positively correlated with student outcomes. ChardÂ’s (2004) re view of the literature supports this notion in his summative statement that Â“teac her quality has a significant effect of student academic achievementÂ” (p. 175). A study conducted by Ascher and Fruchter (2001) analyzed New York Cit yÂ’s schools, examining the teacher experience and quality on studentsÂ’ academic achievement. Findings suggest that there were a greater percentage of teachers who were temporarily placed, had less than five years experience, did not posses an advanced degree, and were fully licensed in lower performing schools as opposed to the highperforming schools. Additionally, it was revealed that there was a 10% higher absentee rate in the former schools. Overal l, the study indicat ed that as teacher qualifications increased, reading achievemen t increased. An important caveat to these findings is the high percentage of ec onomically disadvantaged and minority population in the low performing schools (93% and 98% respectively) as compared to the high performing school s (37% and 52%). The authorÂ’s addressed this by suggesting these variables serve as Â“systemat ic mediators of a less professional teachi ng staffÂ” (p. 213).
39 Child Trends (2004) indicated that higher quality childcare as defined by smaller teacher-child ratio and caregiver training and education, predicts positive outcomes in relation to cognitive, languag e and social development for at-risk children. In comparing the type of child care setting (i.e., center versus home based), they reported center-based care as housing staff with more education and training, in addition to providing mo re structured activities with greater access to child-oriented toys. Although the benefit of home-based care is the lower teacher-child ratio, center-bas ed care has been reported as leading to better cognitive, language, and social outcomes (Child Trends, 2004). In the Early Childhood Collaborative (2002) disseminated by the local county referred to throughout this document there was a 59% turnover rate for the 2001 year, along with low wages and benefits. This trend has been consistently noted despite the increase in childcare costs, which seemingly serves as a catalyst for high attriti on rates and difficulty attracting qualified personnel. With regard to trainings, t he aforementioned annual report stated that while training workshops are offered, t here is no organized protocol or followup/support provided for the training received. Furthermore, the training that is offered is through individual trainer s as opposed to those professionals associated with research-based in stitutions in the field of early education or early childhood mental health. Mesosystem The mesosystem encompasses the relationships and avenues of communication between the various micro systems involved in the childÂ’s life.
40 Collaboration between home and child ca re setting, for example, is deemed crucial in the academic and behavioral development due to the benefits of consistency and support. This part nership focuses on the roles and responsibilities of each party as th ey promote the social and academic development of the child. Accordi ng to Christenson, Rounds and Franklin (1992), it is the recognition that two systems working together can accomplish more than either could s eparately. The authors su mmarize four literature reviews concluding that parent involvem ent in education has several promising effects. More specific, higher student achievement as well as higher noncognitive behavior (i.e., attendance, atti tudes, self-concept, school behavior) are positively correlated with parent involvem ent, in addition to an increase in educational programs and schools that are deemed more successful and effective. The effects on achievement appear to be most significant and long lasting when such involvement and colla boration is initiated at an early age (Christenson et al., 1992). Exosystem Exosystem is the setting that does not contain the child directly but affects their experiences. This includes par entsÂ’ workplace, welfare services, community resources etc. Neighborhood disadvantage and achievement Neighborhood poverty has been linked to poorer developm ental outcomes including def icits in verbal ability, reading recognition and achievement scores (Child Trends, 2004). According to the data presented in Child Trends (2004) for the 1999 calendar year, 22% of
41 children under the age of five resided in neighborhoods in which 20% of the population was categorized as being below the poverty line. Four percent of children were living in nei ghborhoods with poverty perce ntages exceeding 40%. Unfortunately, low SES participates in a dow nward spiral of aversive trends. For example, low achieving children from economically disadvantaged families and neighborhoods tend to enroll in schools that have deprived resources (Child Trends, 2004; Stipek, 2001) and whose principals have a difficult time hiring qualified teachers (National Center for Education Stat istics [NCES], 1998c). In addition, these teachers are often not cert ified in the area they teach (NCES, 1998b) and spend less time engaged in instruction due to their report of attending frequently to classroom managemen t and discipline activities (NCES, 1998c). Monetarily, the schools that se rve economically disadvantaged students have lower per-pupil expenditures ( NCES, 1998a), thus facilitating the aforementioned spiral. An additional component of neighborhood po verty is the perceived level of safety as reported by parents. More spec ific, if a parent views the community as unsafe they will be unwilling to utilize existing resources such as libraries, parks, and childrenÂ’s programs (Child Trends, 2004) Additionally, these fears tend to isolate children and reduce their exposu re and interaction, thus negatively impacting their ability to learn and succeed (Child Trends, 2004). This impediment to academic success is more pronounced when considering community or school SES as opposed to fam ily SES, suggesting that a child from a low SES family is at lower risk when enrolled in a moderate/high SES school,
42 thus further confirming the detrim ental effects of neighborhood disadvantage (White, as cited in Snow, Burns, and Gr iffin, 1998). Statistically, White (1982) reported average correlations of 68 between SES at the school level and achievement in a meta-analysis, in c ontrast to average co rrelations of .23 between achievement and SES at the indivi dual level, supporting the use of school SES in research studies. Neighborhood disadvantage and behavior Available community resources are deemed challenging for low income families living in neighborhoods that are higher in crime and lower in public services. In examining socioeconomic status as a ri sk factor contributing to the development of behavior disorders, Mc Gee and Williams (1999) sugges ted several potential trajectories. First, they suggested that the persistent poverty experienced by low SES families places an extraordinary amount of stress on parents, resulting in an interference in parenting skills. Relatedly, Haddad (1991) noted that the parental values of low SES families might contri bute to the high incidence of aversive behaviors among their children. In co mparing high SES parents to low SES parents, Haddad noted that the former emphasized an internalized system of self-direction whereas the latter emphasiz ed conformity to externally imposed rules. These differences in disciplinary styl es are significant in the acquisition of values and behavior. Second, the lack of a significant income limits a familyÂ’s access to health care, which hinders t he probability of receiving effective treatment. Lastly, children from low SES homes are more likely to be exposed to unsafe or unhealthy environments. Such environments may include a range of
43 negative situations, from witnessing ph ysical violence at home or in the community to lack of supervision and parental support. Exposure to such violence may hinder the childÂ’s ability to develop autonomy, security and trust, as well as facilitate the need for self-pro tective behaviors (Child Trends, 2004). Overall, socioeconomic disadvantage plac es increased levels of stress on the family coupled with fewer resources, thus facilitating the likelihood of caregivers responding in a more hostile and negativ e manner towards their child(ren) (Schultz & Shaw, 2003). Research also has explored the im plications of neighborhood economic disadvantage on the social/emotional out comes of children over and above family socioeconomic status. Schneiders, Drukker, van der Ende, Verhulst, van Os, and Nicolson (2003) as well as Kalff, Kroes, Vles, Hendriksen, Feron, Steyaert, et al. (2001) found that increas ed levels of behavior problems were present at a statistically significant level despite controlling for gender, age and family SES, therefore sugges ting that living in such environments serves as an independent risk factor for children. Po ssible hypotheses for this conclusion as provided by Schneiders, et al. (2003) incl ude, (1) perceived danger leading to anxiety, (2) exposure to inappropriate peers and adults, (3) low levels of neighborhood cohesion, and (4) informal social control and collective efficacy. In keeping perspective, however, it is import ant to realize that familial SES plays a larger role in the presence of child hood behaviors than neighborhood SES (Boyle and Lipman, 2002). This is, in part, due to the immediate exposur es of the family environment.
44 Neighborhood risk factors also were studied by Shaw, Owens, Giovannelli, and Winslow (2001), supporting t he aforementioned findings that the community environment serves as an in fluence on the childÂ’s behavioral repertoire. More specific the findings corroborate prev ious research revealing that children with externalizing behaviors are more likely than typical children to be exposed to maladaptive parenting prac tices and coercive interactions. Supplementing this well-established resear ch trend is the significant role of the neighborhood in which the fa mily resides. Shaw et al. (2001) concluded that children with disruptive behaviors have a higher tendency to live in more dangerous neighborhoods as co mpared to both typical children and children with ADHD. These findings were based on a longitudinal study that further documented the continued presence of externalizing problems throughout the preschool years. Child care setting and academics Childcare settings vary drastically and thus may impact the influence it has on the childÂ’s development and subsequent readiness for formal schooling. Accord ing to Magnuson and Waldfogel (2005), structural quality indicators are used to gauge the level of care provided to children. What they f ound was that 86% of school -based preschool teachers possessed a four-year degree, t wice that of center care and Head Start teachers. Teacher salary also was higher for the former group and was in-line with elementary teachers. Ov erall, Magnuson and Waldf ogel (2005) reported that preschool programs provide comparatively high quality care, particularly when meeting or exceeding the recommendations of the Na tional Association for the
45 Education of Young Children (NAEYC) r egarding class size and child-caregiver ratio. In terms of academ ic outcomes, the authorÂ’s referred to previous studies pertaining to the positive outcomes of pr ograms such as the Infant Health and Development Program (IHDP) and the Ca rolina Abecedarian Project; however, their interests lied in other types of pr ograms, given that not all children go to model programs such as the ones menti oned above. Not surprising, the findings suggest that children who attended presch ool (center care or school-based) fared better on measures of achievem ent skills than did their peers with no preschool experience (incl uding parental child care). These effects were significant for three and four year olds; however the link to academic performance was not observed when child care was extended downward from birth to three (Magnuson and Waldfogel, 2005) As with previous research, the largest effects were noted fo r disadvantaged groups. Child care setting and behavior Finally, under the ex osystem, is the issue of social-emotional adjustment as it relates to the amount of time spent in a child care setting. This topic is included here, as opposed to the microsystem (which is more closely tied to child care se tting) due to the implications of reduced amount of caregiver/child time. In other words, risk factors as described within the exosystem pertain to circumstances in which the quality/quantity of caregiver time is compromised due to stress, employ ment, etc. As noted by the Bureau of Labor Statistics (2000), the rate s of maternal re-employm ent prior to the childÂ’s first birthday have steadily increased from 27% in 1970 to 58% in 1999. As such, there are more and more children entering non-maternal childcare settings during
46 their infant, toddler and presc hool years. According to a review of the literature, as well as an in-depth longitudinal study, conducted by the National Institute of Child Health and Human Development Ea rly Child Care Research Network (NICHD, 2003) there is a significant lin k between the amount of time a child spends in a non-maternal care setting and the presence/extent of externalizing problems exhibited at 54 m onths of age and during kinder garten. These findings remained stable despite controlling for variables such as quality, type, and instability of the childcare center. Magnuson and Waldfogel (2005) reported similar findings, adding that children who received parental care with no formal preschool entered kindergar ten with better behavior and self control, even when other child and family characteri stics were controlled for. Assessing Change over Time Behavioral and Social sciences often seek to measure individual change over time; however this undertaking is frequently challenged by inadequacies in conceptualization, measur ement and design (Bryk & Raudenbush, 1987, p.147). To elaborate, Bryk and Raudenbush (1987) noted that tests used to measure change typically compare individuals based on a fixed point in time, thus failing to address the rate of change among those indi viduals. Further, the design of many studies focus on data pertaining to observa tions at two points in time (pre/post test) which, according to Bryk and Raudenbush (1987) provide an inadequate basis for studying change. The application of hierarchical linear models (HLMs) presents an alternative to ot her methods by creating an integrated approach to examining the structure of i ndividual growth. That is, growth trajectories and the
47 various characteristics that impact grow th can be examined individually while holding different levels of influence const ant. A detailed description of HLM can be found in chapter three. For the moment, a brief revi ew of research supporting the use of HLM will be presented. Recent studies (Cusumano, et al, 2006; Taylor et al; 2005, Armstrong, Dedrick and Greenbaum, 2001, Stipek & Miles, 2008) support the use of Hierarchical Linear Modeling (HLM) as a method to demonstrate change in educational research. Specifically, Cusum ano et al (2006) explored the impact of early childhood educator training and c oaching on literacy acquisition of preschool children. A three level model was structured, ex amining within-child differences (reading scores), child characte ristics (age, race, etc.) and classroom characteristics (treatment intensity, etc) Taylor et al (2005) utilized HLM to analyze the impact of school and classroom level characteristics on the reading growth of elementary school students. Armstrong, Dedrick, & Greenbaum (2003) applied HLM to examine factors associat ed with community adjustment of young adults with serious emotional disturbanc es, whereas Stipek and Miles (2008) tested three hypotheses explaining t he association between aggression and achievement through HLM. All studies we re able to investigate change over time while holding variables with potential effect s constant, thus providing more clear results. Summary In summary, the research thus far supports the relationship between poor academic achievement and social-emoti onal maladjustment in school aged children. More importantly, a host of risk factors have been identified as
48 underlying mechanisms that exacer bate the presence of a negative developmental trajectory, which remains stable and resistant over time. Although the above review of theoretical risk factors pertaining to academic and social/emotional development is compartmen talized, it goes without saying that attention to all levels of the ecologica l system be given when intervening with a child. As Finn, Gerber, and Boyd-Zaharias (2005) suggest, it is the culmination of experiences, often beginning in the early years, that lead to maladaptive outcomes. Given this unequivocal need for prevent ion and early intervention, the current study attempts to explore t he achievement-behavior relationship among preschool children, with specific emphasis on early literacy skills. A more indepth analysis will address t he particular pathways or behavioral profiles that affect the rate and levels of literacy ac quisition. Additional analyses will examine the literacy development of children wit h challenging behaviors who experience varying levels of within-child protective factors.
49 CHAPTER THREE METHOD The purpose of this chapter is to pr esent the research design, procedures and statistical analyses for the current study A description of the archival study, of which this study is an extension, is provided, including participants, measures and procedures. Research Design The current study used archival data gathered from a quasi-experimental study that examined the impact of an early learning opportunity project on the literacy skills and social-emotional devel opment of preschool children. The original study consisted of three c ohorts of teachers who received the independent variable. Data were analyzed in that original study on two of the three cohorts. The present study added the data from the third cohort and addressed research questions for that cohort. Description of t he Original Study This section describes the original st udy from which the Cohort Three data were obtained. The Pinellas County School Reading Coalition so ught to improve levels of literacy, reading readiness and social-emotional functioning in children from birth to five years through the im plementation of a community collaboration project. The Coalition designed the Pinellas Early Literacy Learning Community Project (LCP) to provide early litera cy training and coaching for teachers and
50 child care professionals across a variety of preschool education settings (e.g., Head Start, subsidized child care, Earl y Intervention programs, faith-based programs). The LCP provi ded age-appropriate early experiences that supported development in the language and social do mains, both of which are known to contribute to improved literacy outcomes (Barrera, et al., 2002; Kamps, 2003). The grant, entitled Early Lear ning Opportunities (ELO ), was funded through a collaborative effort of four agencies includ ing Coordinated Child Care of Pinellas County, Directions for Mental Health, Flori da Mental Health Institute, and Pinellas County Childcare Lice nsing Board. The project consisted of two primary acti vities: direct training for teachers in early literacy development skills and coaching support to teachers. Saint Petersburg Community College provided the literacy training to the early childhood professionals through a c ourse entitled, Â“LAE 2000: Language Development for Young Children.Â” The c ourse, designed specifically for the study participants, used HeadsUp! Readi ng (HUR), a researched-based, early literacy distance-learning curriculum (Nati onal Head Start Association -NHSA) to enhance the early literacy skills of the parti cipants. In addition, literacy coaches were provided to facilitate the transfer of skills from the co llege classroom to the preschool classroom. Data were colle cted to examine the effects of the intervention (training and coaching) on t he literacy development of the preschool children.
51 ELO Participants Early Childhood Teachers A total of 48 early childhood teachers from Pinellas County participated in one of the Language Development in Children courses offered in the spring and summe r of 2004. Recruitment of teacher participants initially consisted of three activities: an invitational package mailed to the 1,400 licensed childcare facilit ies registered through the Pinellas County License Board for ChildrenÂ’s Centers and Family Day Care Homes; a project description in the local college course ca talog; and a project description in the newsletters from the Coordi nated Child Care of Pinellas and the Licensing Board. Interested teachers completed an applicat ion to participate in the grant (see Appendix A). Participation in the project was limit ed to early childhood teachers of children between the ages of three and five employed at a childcare site, private pre-kindergarten, or Head Start program. Teachers who were employed at a family or home-based childcare setting we re excluded in Cohorts One and Three. The 48 teachers who participated in the st udy were divided into three Cohorts. Cohort One was comprised of 12 teac hers who received both training and coaching at the same time (Concurrent), 10 teachers who received training and no coaching (Delayed) and 19 teachers wh o received no training or coaching (Control). Cohort Two was comprised of 26 teachers, 19 teachers from the Control group of Cohort One and an additional seven teachers from family and/or home-based childcare centers. Sev enteen teachers in Cohort Two received training and coaching (Concurrent) and ni ne received training and no coaching
52 (Delayed). Cohort Two did not have a Control group. Cohort Three was comprised of eight of the nine teachers in the Delayed group of Cohort Two. Four of the teachers in Cohort Three were assigned to the teaching and coaching group (Concurrent) and four were a ssigned to the teaching and no coaching group (Delayed). Since all eight teacher s completed the training as part of Cohort Two, the sole difference between the two groups in Cohort Three was, theoretically, the coaching component (one group was scheduled to receive coaching and the other group was not to receive coaching). However, due to time constraints and organizational issues coaching was not provided to either group. Assignment into Concurrent, De layed, and Control groups differed based on cohort and will be described in the appropriate section. Table 2 is a summary of the teac her participants in each cohort by treatment condition. Table 2. Teacher Participants in Cohorts One, Two and Three Number of Teachers in Cohort Cohort One Cohort Two Cohort Three Concurrent: Training/Coaching 12 17 4 Delayed: Training/No Coaching 10 9 4 Control: No Training/ No Coaching 19 n/a n/a Total Sample 41 26 8 Data collected on the ELO teacher participants included number of children assigned to the classroom, year s of teaching experience, and highest level of education. See Table 3 for a summary of these data on Cohorts One and Two.
53 Table 3. Descriptive Information on Teacher Part icipants in Cohorts One and Two by Condition Number of Student Participants Avg. Experience (in Years) Highest Level of Education H.S. Some College AA 4 Yr Degree Cohort One Concurrent 165 8.24 8 2 1 1 Delayed 106 13.59 6 0 0 4 Control 115 7.99 1 9 5 4 Total Sample 386 9.68 15 11 6 9 Cohort Two Concurrent 107 6.42 9 5 0 3 Delayed 54 11.22 1 1 0 7 Total Sample 161 8.16 10 6 0 10 Table 4 represents data obtained on the each of the teacher participants for the current study (Cohort Three). Table 4. Descriptive Information on the Eight Teache r Participants in Cohort Three by Site Condition Type of Childcare Site Number of Student Participants Years of Teaching Experience Highest Level of Education* Site # 1 Concurrent Faith Based 13 8 A.A. Site # 2 Concurrent Private Center5 3 High School Site # 3 Delayed Faith Based 3 1 A.A. Site # 4 Concurrent Priv ate Center19 20 B.A. Site # 5 Delayed Faith Based 12 20 Some College Site # 6 Concurrent Priv ate Center14 16 B.A. Site # 7 Delayed Private Center4 Some College Site # 8 Delayed Faith Based 13 1 Some College Note. A.A. = AssociateÂ’s Degr ee, B.A. = BachelorÂ’s Degree Preschool Student Participants A total of 630 preschool children participated in Cohorts One, Two and Thr ee of the program evaluation, with 386,
54 161, and 83 students respectively in each Cohort. Participation was limited to children who met the followin g criteria: (1) between three and five years of age, (2) English as the primary language, (3) no diagnosed cognitive delays, and (4) no hearing or visual disabilities. One additi onal criterion (i.e., has not previously been a student of a teacher who participated in the ELO grant) was included for Cohort Three. The focus of the current study is on the literacy development of preschool children within the Third Cohort. An overview of the student demographic data for Cohort Three is depicted in Table 5 below. Table 5. Demographic Information on Student Participants in Cohort Three Number of Student Participants Age in YearsGender Racial Distribution 3 4 5 Male FemaleWhiteA.A.Hisp. AsianOther Site # 1 1 12 0 7 6 11 1 1 0 0 Site # 2 2 3 0 2 3 0 5 0 0 0 Site # 3 2 1 0 1 2 0 3 0 0 0 Site # 4 0 13 6 10 9 19 0 0 0 0 Site # 5 3 9 0 9 3 9 2 1 0 0 Site # 6 0 14 0 9 5 12 0 0 0 2 Site # 7 3 1 0 3 1 1 2 0 1 0 Site # 8 0 13 0 3 10 7 0 4 1 1 Total 11 66 6 44 39 59 13 6 2 3 Note. A.A. = African American, Hisp. = Hispanic Research Variables Predictor/Independent Variables The independent variables in this study are categorized into two groups: child ( demographics, within-child protective factors and behavior) and childcare site. The child variables previously found to influence academic achievement include race, gender, age, attendance, home SES, presence of within-child protective factors (e.g., In itiative, Self-Control, Atta chment) and student behavior.
55 Demographic data (i.e., race, gender and age) for each child participating in the project were obtained from teacher participants via a data information sheet (See Appendix B) developed specifically for th is project. Teacher participants also provided the number of days (attendance) each child was absent during the semester of interest to quantify expos ure to the litera cy program. This information was documented on an attendanc e data sheet (See Appendix C) completed by the teacher. Home socioecon omic status for each child participant was assigned by obtaining the median household annual income of the neighborhood in which the child resides. Income levels were grouped into equal increments and were labeled for data ent ry (e.g., 1=$10,000 -19,999; 2=$20,00029,999). This method was employed due to the barrier in obtaining income levels for the child participants from the childcare providers. Using an internetbased GIS Mapping system, the zip codes we re entered to obtain an indicator of median neighborhood SES. The m apping system was developed by the Pinellas County Economic Development department as a tool for linking geographic locations with demographic indicators such as racial distributions, home values, and median household incomes ( http://www.siliconbay.org/gis3/gis_content.cfm ) in which the data are sorted by census trac ts, municipalities, and zip codes. This method has become increasingly popular as a means to ascertain SES, particularly in mental health research (Krieger, Williams & Moss, 1997; Krieger, 1992). Krieger (1992) stated t hat it is a Â“valid and useful approach to overcoming the absence of socioeconomic data in mo st US medical re cordsÂ” (p. 709). Although using aggregate geogr aphic data to proxy socioeconomic status does
56 not come without limitations, it is an alternative approach to obtaining such information (Krieger et al, 2002; Soobader, et al, 2001). The level of within-child protective factors was determined by the sum of scores on three subtests (Initiative, Self -control and Attachment) of the Deveraux Early Childhood Assessment (DECA, LeBu ffe & Naglieri, 1999). Protective factors refer to those char acteristics (both individual and environmental) that buffer negative events or stressors and result in more positive psychological and behavioral outcomes (Masten & Garmezy, 1985). Children who demonstrate these characteristics are considered re silient while those children who lack protective factors are at a greater risk for developi ng behavioral and emotional problems (LeBuffe & Naglieri, 1999) The student behavior score was determined by using the 10-item subscale score of the Behavior Concerns Scale on the DECA. Childcare site is the second ind ependent variable previously found to impact academic achievement. Site vari ables examined in the study include teacher education, teacher experience, school SES, and classroom environment. Data on teacher education (highest degree obtained) and experience (in years) were collected from the application complet ed by teachers for grant participation. School SES was measured by obtaining t he median annual household income of the neighborhood in which the childcare c enter was located. The last site variable, classroom environment, was evaluated through the Early Literacy Observation Checklist (ELOC), which meas ured teacher-child interaction (e.g.,
57 use of open-ended questions) as well as t he literacy-related environment (e.g., availability of books in the classr oom, presence of print/signs). Outcome/Dependent Variable Literacy development in preschool children is the outcome variable examined in the current study. Literacy development was assessed through growth in expre ssive language and phonemic awareness. These skills were assessed through the adm inistration of the preschool version of the Individual Growth and Development Indicators (IGD I). The three subtests of the IGDI (Picture Naming, Alliterati on and Rhyming) served as the measure of rate and levels of literacy develop ment in the student participants. Measures Data were collected from an archival database for the teacher and student participants of the ELO grant for Cohort Three. See Table 6 for a summary of data sources for each variable. Table 6. Measures Used to Assess Research Variables Measure Research Variable Teacher Application for Inclusion in Grant Teacher Education (P) Teacher Experience (P) School SES (P) Student Demographic Data Sheet Race (P) Gender (P) Age (P) Home SES (P) Student Attendance Sheet Attendance (P) Individual Growth and Development Indicators Â– Preschool Version Early Literacy (O) Deveraux Early Childhood Assessment Total Protective Factors (P) Behavior (P) Early Literacy Observation Che cklist Classroom Environment (P) Note: P=Predictor Variable; O= Outcome Variable.
58 Demographic Sheet Information for each child was gathered via a demographic sheet completed by a childca re provider (see Appendix B). Data gathered on each child included: (1) date of birth, (2) gender, (3) race, (4) home zip code, and (5) primary language. A s eparate sheet (Appendi x C) was used to obtain attendance information for each child. Individual Growth and De velopment Indicators Â– Preschool Version. Early literacy skills were assessed utilizing the Individual Growth and Development Indicators (IDGI). The IGDI was devel oped by McConnell and McEvoy (2002) as part of the Early Child hood Research Institute on Measuring Growth and Development. These l anguage and literacy assessm ent tools are used with children from birth through eight year s of age and serve as General Outcome Measures (GOMs) of development. In a broader sense, GOMs depict individual childrenÂ’s growth and development over time thus tapping into both their current performance as well as their rate of development. Fuchs and Deno (1991) described GOMs as Â“reliable, valid and e fficient procedures for obtaining child performance data to evaluate intervention programsÂ” (p. 489). The IGDI used in the current study served as a GOM of literacy development, and can be used repeatedly over short periods of time to evaluate a studentÂ’s response to an intervention and to identify children at-risk. More specifically, the preschool IGDI is designed to measure early literacy skills in children from three to five years of age and was utilized with the student participants in all three cohorts. The three subtests of the IGDI used for this study were: Picture Naming, Alliteration, and Rhyming.
59 Picture Naming is a measure of expr essive language that is assessed by presenting the child with pictures commonly found in their environment (e.g., food, animals, toys, vehicles). This subtest is initiated by a four-item demonstration provided by the examiner to ensure the childÂ’s understanding of the task. Next, the child is given the oppor tunity to practice with the same four items, and feedback is provided to the ch ild. Following the sample items, the timed and scored portion of the IGDI is adm inistered. The child is instructed to name the pictures as fast as he or she c an. The examiner st arts the stopwatch as the first picture is presented and at t he one minute time limit the subtest is concluded. If at any time during the mi nute the child hesitates for three seconds the examiner prompts him or her by saying, Â“What do we call this?Â” The child is then given an additional two seconds at whic h time the next picture is presented, regardless of the childÂ’s response or lack thereof. The number of cards identified correctly is counted, thus becoming the childÂ’s Picture Naming score. Phonemic awareness is assessed by the Alliteration and Rhyming subtests of the IGDI. It is a vital el ement in reading success (Snow et al., 1998) that typically develops during the presc hool years (Lonigan, Burgess, Anthony, & Barker, 1998; Whitehurst & Lonigan, 1998). The Alliteration subtest measures phonemic awareness skills by assessing the st udentÂ’s ability to identify pictures that start with the same sound. The child is presented with cards containing one picture on the top (target pi cture) and three pictures ac ross the bottom. The child is instructed to, Â“point to the picture t hat starts with the sa me sound as the top picture.Â” The examiner dem onstrates the concept to the child in addition to
60 allowing him or her to prac tice before moving on to the actual administration. The demonstration is conducted with two standard cards and then four random cards are used for the childÂ’s sample item s. Once the demonstration and sample items are complete, the test administrati on begins. The subtest is two minutes in length and prompting is provided for child ren who reach the three second delay point (i.e., Â“Which one starts with the same sound as ____.Â”). The number of cards named correctly is considered the childÂ’s Alliteration score. Rhyming is another measure of phonem ic awareness. The child is presented cards that contain four pictures, with the ta rget on top and three on the bottom. The child is asked to Â“point to the picture that sounds the same as, or rhymes, with the top picture.Â” The child is provided with a demonstration as well as sample items for practice. If no res ponse is provided after three seconds, the examiner prompts t he child by asking, Â“Which one sounds the same as ____?Â” The number of cards identified correctly at the end of the tw o-minute time period is the childÂ’s Rhyming score. Priest, Davis, McConnell, and Shi nn (1999) examined the psychometric properties of the IGDI, s upporting its effectiveness in measuring early literacy skills in preschool children. Concurr ent validity coefficients between the IGDI Picture Naming and two norm-referenced meas ures of preschoo l language skills (e.g., Peabody Picture Vocabulary Test-T hird Edition, PPVT-3 and Preschool Language Scale-Third Ed ition) ranged from r =.47 to r =.69 Correlations between expressive language scores and chr onological age (assessing the toolÂ’s sensitivity to growth over time) we re assessed with samples of typically
61 developing children ( r =.63), children enrolled in Head Start ( r =.32) and children receiving services in a preschool special education classroom ( r =.48). One month alternate form reliabi lity coefficients ranged from r =.44 to .78. Missall & McConnell (2004) reported stabl e test-retest reliability of the Rhyming subtest ( r = .83 to .89, p < .01) when measured over three weeks. Validity also was examined for the Rhymin g subtest. Results revealed a positive correlation with other standardized meas ures of phonological awareness and early literacy development including the PPVT-3 ( r = .56 to .62, p < .05), Concepts About Print (CAP; Clay, 1985; r = .54 to .64, p < .01) and Test of Phonological Awareness (TO PA; Torgeson & Bryant, 1994; r = .44 to .62). Significant correlations were found between chronological age and IGDI scores (r =.46, p < .01) supporting the sensitivity of the instrum ent in measuring growing phonological skills (Missall & McConnell, 2004). Stable test-retest reliability also wa s found with the Alliteration subtest ( r = .46 to .80, p < .01; Missall & McC onnell, 2004) when m easured over three weeks. Tests of validity revealed that Alliteration correlated with the PPVT-3 ( r = .40 to .57, p < .01) TOPA (r = .75 to .79, p < .01), and CAP ( r = .34 to .55, p < .05). A positive correla tion with chronological age ( r = .61) also was found. Deveraux Early Childhood Assessment (DECA; LeBuffe & Naglieri, 1999). The Deveraux Early Childhood Assessment (DECA) is a strengths-based, normreferenced behavior rating scale completed by the preschool teacher. It was designed to assess the level of within-child pr otective factors (i.e., initiative, self control, and attachment) evidenced by t he preschool child and to measure the
62 level of emotional/behavior problems dem onstrated by the child in the early childhood environment. The DECA is a 27-item rating scale and a 10-item behavior concerns screener for children ra nging from ages two through five. LeBuffe & Naglieri (1999) reported that the internal reli ability for teacher informants ranged from .80 to .94 for the Behavioral Concerns and Total Protective Factors Scales respectively. Test-Retest reliability coefficients were obtained at a 24 to 72 hour interval and r anged from .68 to .94. Interrater reliability coefficients ranged from .59 to .77. Criterion related validity was conducted by comparing the scores of tw o groups of preschoolers; those with known emotional/behavioral problems and ty pical children within the community. Results revealed statistically signif icant differences between the groups indicating that the DECA discriminat es between groups of children with and without emotional/behavioral problems (LeBuffe & Naglieri, 1999). The DECA is the first published rating scale of within-child protective factors. Therefore, standar d measures of content and construct validity are not available. The content of the DECA wa s derived from the resilience literature (i.e., behavioral descriptions found in the lit erature to identify resilient children) and from focus groups conduc ted with parents and teachers (LeBuffe & Naglieri, 1999). LeBuffe and Naglieri (1999) conduc ted a principal factor analysis with varimax rotation on the standardization data set to obtain the Protective Factor Scales (Initiative, Self-control and Atta chment). Results of the factor study resulted in a three-factor solution with fa ctor loadings ranging fr om .46 to .74. The 10-item Behavior Concerns Scale wa s created by select ing two items with
63 the strongest factor loading fr om each of five scales on t he DECA (i.e., Attention problems, emotional cont rol problems, aggression, withdrawal/depression, and increased concern problems). Construct validity was assessed th rough an alternative technique that determines whether the instrument yields data that are consistent with the predictions generated from t he underlying theory of t he instrument. Lebuffe & Naglieri (1999) achieved this by obtai ning measures of risk and stress in the same group of preschool children for w hom data were collected using the DECA. A two-way Analysis of Variance (ANOVA) conducted by Lebuffe & Naglieri (1999) supported the use of t he DECA in measuring protective factors related to resilience in young children (p<.001). T hat is, children withÂ” high risk/lowÂ” protective factors scored higher on the Behavior Concerns Scale than those children withÂ” low risk/high protectiveÂ” factors. Criterion-related validity was measur ed by comparing a group of preschool children identified as having emotional/behav ioral problems to a group of Â“typicalÂ” preschool children (LeBuffe & Naglieri, 1999). Results of the Independent t-tests revealed statistically significant (p<.01) differences between the two groups on all scales of the DECA (LeBu ffe & Naglieri, 1999). Early Literacy Observation Checklist (ELOC; Justice, 2002) Â– The Early Literacy Observation Checklist (ELOC) a ssessed the literacy-related environment (e.g., availability of books in the classroom, presence of posters, signs and labels) as well as teacher-student intera ction variables (e.g., Does the adult ask the children to help read the title?, Does the adult praise the childrenÂ’s
64 participation?) related to literacy (J ustice, 2002). The literacy related environment refers to t hose settings and experiences that foster language and literacy growth through activities, includi ng talking, playing, reading and writing (National Head Start Association -NHSA) For children receiving childcare services outside the home, the childcare provider represents an important resource for facilitating this growth. Recognizing this, the HUR curriculum strives to equip teachers with the research based techniques and strategies necessary to foster literacy development in children. Therefore, the pur pose of the ELOC in the current study was to assess treatment integrity by examining the extent to which the childcare providers incorpor ated the knowledge and skills gained from the HUR curriculum and coaching sessions. The ELOC is comprised of four se ctions, (1) Storybook Reading, (2) Classroom Library, (3) Wr iting Center, and (4) Over all Print Environment. Literacy-related environment was assessed by teachersÂ’ responses to a variety of forced-choice (yes/no, and multichoice) and open-ended questions. For the purposes of this study, modifications we re made to the ELOC. Specifically, the literacy coaches and instruct ors of the LAE2000 course requested that the ELO evaluation team incorporat e the content of the Heads Up! curriculum to more accurately reflect the skills and informa tion taught to the teacher participants. The modifications proposed by the liter acy coaches and course instructors included the addition of two items in the Ov erall Print Environment section, (Â“Are printed materials displayed prominently in the early learning environment?Â” & Â“Are posters and signs displayed at eye level?Â”) and the extension of the rating
65 choices for two existing item s. The original version of the ELOC contained an open-ended question in the Storybook Read ing section regarding the frequency in which story time was held. The modi fication to this item required a forced choice response of never, one time per week, 2-3 times per week, once per day or more than once per day. The sec ond item suggested for modification (Â“Is there a specific space for childrenÂ’s i ndependent and group writing activities?Â”) is located in the Writing Center section and originally required a yes/no response. This was changed to a three-point scale, which provided the following response options: specific writing c enter, center set up only during choice time and no place for writing. The ELOC was completed as a preand post-measure to assess treatment integrity over time. Each item was assigned a weight on a 0 to 1 scale in .25 increments depending on the response format. Scores obtained included the four aforementioned se ctions in addition to an Overall Literacy Environment Score, which is the sum of all sections. Higher scores reflect a more literacy-rich childcare environment. Inter-rater agr eement was obtained by calculating the results from observations completed by dyadic pairs consisting of Program Evaluators and school psychology graduate students. Specifically, each dyad completed the ELOC while observing a litera cy activity in a preschool classroom. The number of agreements between each ob server was divided by the total number of items on the measure to dete rmine inter-rater agreement for the dyad. Inter-rater agreement of .85 or above was required prior to t he utilization of the instrument in the ELO grant. To date, there have been no attempts to obtain
66 psychometric properties of the ELOC, as th e developerÂ’s original intent of the measure was to provide a functional snapshot of the environment. Procedures for Original Study ELO grant activities were conducted between August 2003 and December 2004. Three cohorts were included. T able 7 depicts the activities and timeline relevant to each cohort. An application to conduct the ELO grant was submitted in August 2003 and subsequently approved. Recruitment of early childhood education teachers began in November 2003, followed by a re view of approximately 150 applicants and final selection of teacher participant s. The Coalition led the selection process, choosing one teacher from eac h childcare site represented in the applications. Teachers were systemat ically chosen based on their limited experience in an effort to provide them with incr eased resources and promote skill building. The interview process for Literacy Coaches (LCÂ’s) began in December 2003. Three baccalaureate-degreed female app licants were hired, all of whom had more than five years of experienc e within an early childhood education setting. Each LC was assigned seven to eight teachers to coach using the Early Literacy Learning Model (ELLM) for mentor ing teachers who engaged in literacy instruction. The ELLM model, developed by the Florida Institut e of Education at the University of North Florida, is a research-based comprehensive curriculum intended to improve language and early lit eracy skills of preschool children (Wood & Fountain, 2007). This model was chosen for use in the grant due to the
67 emphasis on instruction and coaching. Training was provided to the LCÂ’s by a consultant from Coordinated Chil d Care of Pinellas County. The application for University of South Florida (USF) IRB approval was submitted in January of 2004 and obtained in February 2004. Due to the archival nature of the current study, an IRB application was submitted to obtain permission to review the ELO database. The project evaluation component of the grant was headed by an Associate Professor at USF who hired three doctoral candidates (ELO head evaluators) from the School Psychology Program, including the author, to complete data collection activi ties. Two of the doctoral candidates were practicing School Psychologists at the Education Specia list (Ed.S). level and were employed by Pinel las County Schools. An additional seven graduate students we re recruited to assist in the evaluation efforts due to the large number of preschool students who participated and to ensure timely data collection. Training on the assessment materials (IGDI and ELOC) was provided for the seven graduate students and conducted by the three head evaluators. These training sessions included a presentation on the background of each measure, administration, scoring and interpretation procedures. Each graduat e student was given materials to use while assisting with ELO data collecti on and was provided the opportunity to practice the assessments during the trai ning. Additionally, the graduate students were required to administer the measures to three children outside the training
68 Table 7. Summary of ELO Grant Procedures and Timeline for Cohorts One, Two and Three Initial ELO Grant Activities (August 2003 Â– January 2004) Application submitted Recruitment and Selection of ear ly childhood education teachers Literacy Coaches (LCÂ’s) interviewed and hired Application for USF IRB approval submitted ELO Head Evaluation Team training in IGDI measures Recruitment and traini ng of graduate students Cohort One (January Â– May 2004) First semester of the Language Devel opment in Children course offered Treatment conditions (Concurr ent or Delayed) identified ELLM training for LCÂ’s Eligible Control group sites contacted by Head Evaluators Parental permission obtained Assignment of 3-4 childcare sites to each evaluator Weekly coaching for c oncurrent group provided Evaluation activities (ELOC and IGDI) were conducted two points in time Teacher Participants (n = 41) Source: Applicants chosen by the Coalition Concurrent (n = 12) Delayed (n = 10) Control (n = 19) Cohort Two (May Â– July 2004) Second, and final, semester of the Language Development in Children course offered (Summer 2004) Evaluation activities (ELOC and IGDI) were conducted two points in time Teacher Participants (n = 26) Source: Control group in Cohort One and reserve list of teachers not eligible for participation in Cohort One (family /home centers) Concurrent (n = 17) Delayed (n = 9) Control (n/a) Cohort Three (September Â– December 2004) DECA added as an evaluation measure Two progress monitoring points were added to assessment schedule Evaluation activities (ELOC, IGDI and DECA) were conducted four points in time Teacher Participants (n = 8) Source: Delayed group of Cohort Two Concurrent (n = 4) Delayed (n = 4) Control (n/a)
69 sessions while being observed for accu racy by their dyad partner. Dyads provided feedback to each other and continued to administer practice tests until 100% accuracy was obtained. Data were collected from teacher participants and the preschool children in their classroom for whom consent wa s granted, as well as through direct observation of the classr oom. These evaluation acti vities occurred in three consecutive stages spanning from January 2004 through December 2004 and consisted of three cohorts. Although data collection procedures were similar across the three groups, there were notabl e differences. The following sections are separated into Cohorts One, Two and Three to best describe the evaluation activities within each group. Cohort One The first Language Development in Children training course began during the IRB approval process in January 2004. It was through this venue that teachers were provided c onsent forms to document willingness to participate in the ELO grant evaluati on activities (see Appendix D for blank consent form). Once consent forms were signed and returned the teachers were designated to one of two treat ment conditions (concurrent or delayed coaching). These conditions were developed to evaluate the effectiveness of coaching on early literacy development (i.e., treatment integrity). This selection process was controlled to the extent that the age of students in the classrooms were equally represented in each treatment condition. For example, once a teacher was randomly selected to receive concurrent coaching, the age of the students was noted and all teachers with similarly aged st udents were placed in a pile. One
70 teacher was then randomly chosen from t hat group to receive delayed coaching. The control group was formed by solicit ing teachers who met criteria for participation but were not selected for participation in Cohort One. Nineteen teachers from seven centers agreed to parti cipate in the control group. Initial contact to the centers was made via phone by one of the three head evaluators and was followed up with an in-person visi t to the site director to discuss procedures and deliver consent forms. Permission from parents allowing their preschool children to participate was obtained via parental assent (See A ppendix E). An information letter (see Appendix F) accompanied the consent form both of which were sent home by the teacher participants. A second wave of forms was sent to parents if a response was not received within two weeks of the original distribution. Once consent forms were collected, file folder s were created for each childcare site with information pertinent to the data collection activities. Basic information included the name of the childcare site, name of the director, contact telephone number, site address, and map wi th directions from USF to the childcare site. Blank data collection forms also were included as were the ages and identification numbers of t he children for whom consent was obtained. The ID numbers were six digits in length, r epresenting the treatment condition (7 = concurrent, 8 = delayed, 9 = control) teacher number assigned by the lead program evaluators, and child number. For example, the 10th child in the classroom of teacher #29 under the delayed coaching treatment condition would be assigned the ID number of 829010. T he names of the children were never in
71 the file folders or directly linked with the data collected. To further ensure confidentiality the Director of each site hel d the master list of student participants, which contained their assigned number (e.g., t he last digit(s) in the ID number). Data were locked in a filing cabinet at USF upon completion of data collection. The three head evaluators assigned three to four childcare sites to each of the program evaluators, including them selves. This process considered the number of children in t he classroom, amount of time the graduate students dedicated to data collecti on, and geographical distance between each site to ensure equal caseloads. For example, the sites were grouped based on number of children in each classroom and t hen further grouped into geographical location. A three-week data collecti on window was set during which time the program evaluators were responsible for sc heduling visits and colle cting data. Observations of the preschool en vironment were conducted using the Early Literacy Observation Checklist (ELOC) and complet ed by the program evaluators and coaches. Prior to the onset of data collection for the study, interrater agreement trials were conducted. An agreement level of .85 or higher was required between dyad partner s. Actual agreement le vels ranged from .85 to .93. The coaches observed the classr ooms of the teacher participants in the concurrent coaching group while t he program evaluators observed the classrooms of both the del ayed coaching and control groups. The teacher participants in the concurrent coaching group were given feedback by their coach on the results of the ELOC as part of the weekly coaching session. Feedback on the classroom observations of th e delayed and coaching groups was not
72 provided. The ELOC took approximat ely 30 minutes to complete and was conducted in the last few weeks of Febr uary 2004. Teacher participants in all groups were briefly interviewed after t he observation to answer the few questions on the ELOC that could not be comp leted by examini ng the classroom environment (e.g., Are children permitt ed to borrow these books for home use? Are there specific times set aside during the day for reading activities?). Aside from obtaining valuable data regarding the cla ssroom environment, the observation provided the children with the opportunity to become comfortable with the examiner in their classroom prior to the individual asse ssment activities. The preschool participants for wh om consent was obtained were individually assessed for early literacy skills through the Individual Growth and Development Indicators (IGD I). The IGDI was administered in a separate area of the classroom to reduce distraction and pr event other children from prematurely viewing the materials. T he evaluators spent approximat ely 5 to 10 minutes with each child, giving verbal praise and a sticker at the end of the assessment session as a reward for participating. Administration of t he IGDI occurred within a three-week window beginning in the mi ddle of February and continuing through the beginning of March 2004. Each evaluator followed the standardized instructions for administration and reco rded scores from each child on the data form (See Appendix G) located inside the si te-specific folder. Data were then entered into a Microsoft Excel spreadsheet by one of the three head evaluators and rechecked for accuracy.
73 A second wave of data collection occu rred from the end of April to the middle of May 2004 at which time the classroom was observed utilizing the ELOC and the IGDI was re-administered to the preschool student participants. Additional data were collected on Cohor t One in May focusing on the preschool student participants who were identified as entering ki ndergarten in the 2004-05 academic year. Cohort Two The training course was offe red for a second and final time during the summer of 2004. Twenty-si x teachers participated (17 in the concurrent coaching group and 9 in t he delayed coaching group) in the ELO evaluation activities. The 17 teacher par ticipants in the concurrent coaching group were obtained from two sources: the spring cont rol group and the Â‘reserve listÂ’ of teachers who origin ally applied to participate in the grant but were not chosen for the first cohort. Evaluation ac tivities and measures used to collect data were identical to Cohort One. S pecifically, the IGDI and ELOC measured literacy growth and the presc hool environment respectively at two points in time. Activities and evaluations related to Cohort Two took place between May and July 2004. Cohort Three Two significant procedure differences are noted in Cohort Three as compared to the procedures and activities of Cohorts One and Two: use of the DECA and number of student assessments. The DECA was added to the evaluation component of the grant because the goals of the evaluation team shifted and focused on further examini ng school-related behavioral competencies (e.g., attention, self-control) in preschool children. These behavioral
74 competencies were originally explored in the previous two cohorts through a teacher-completed Ages and Stages Questionnai re (ASQ, Bricker, D., & Squires, J.), which was organized and managed by an ELO grant committee member employed by Directions for Mental Health in St. Petersburg, Florida. Homebased mental health services were offe red to families of children who scored within a predetermined range. The ASQ is a tool desig ned to screen infants and young children for developmental delays. Dir ections for Mental Health used the ASQ as part of their standard screening protoc ol of all children; therefore it was not used as part of the ELO evaluation ac tivities. Although data were gathered on Cohorts One and Two as part of this standard screening, they were not included in the analyses conduc ted by the ELO evaluation team. The evaluation team replaced the ASQ with the DECA as a behav ior screener to obtain information on both behavior concerns and with in-child protective factors of the preschool students. This replacement al so enabled the evaluation team to apply the same standardized procedures (ensur ing confidentiality, collecting data according to evaluation timeline, doubl e-checking scoring and data entry) to the behavior screener that were appl ied to all other grant related measures. The committee member at Directions for M ental Heath was provided with the name and contact information of those childr en who scored in the at-risk category on the Behavioral Concerns Scale of the DECA to continue with the home-based referral services. The DECA was given to the teacher participants along with the demographic sheet once consent forms were received from parents. Results on the DECA were compared to data obtained on the IGDI to explore the differences
75 in literacy development among children with behavior concerns who had high scores measuring within-child protective fa ctors in comparison to children with challenging behaviors who had low scores measuring within-child protective factors. The second difference in the eval uation procedure was the number of times the student participants were asse ssed. Specifically, two progress monitoring points were added to the pre and post points, resulting in four total assessments for each student. Due to the reduced number of teacher participants, the three head evaluators completed all observations and assessments for the third cohor t. Data were collected monthly during a five-day window, starting late September 2004 and ending mid December 2004. Procedures for Selection, Review and Analys is of Archival Data in Current Study Archival Data The archival data consists of demographic information at the teacher (education, years of ex perience), child (gender, age, race, attendance, and home SES), and childcare site (class size,) levels, as well as data from observations of the preschool classrooms. These data were obtained from the following sources: the application form completed by the teacher for grant participation, a demographic info rmation sheet, attendance sheet and the ELOC. In an effort to measure liter acy growth and behavior levels of the preschool children, archival data from the IGDI (pre/post and two progress monitoring points) and DECA was revi ewed. Child participant scores on the Behavior Concern Scale of the DECA was organized from lowest score to highest score. This method assisted in ex amining the outcome measure (literacy
76 development in preschool children) with re spect to the teacherÂ’s perception of the degree to which academic behavioral competencies are present in the classroom. Research Design The data were collected and entered into the comprehensive database, SPSS for Windows. Demographic characte ristics (i.e., gender, age and race of students, home SES), attendance, years of education and experience of teachers, and site SES of the sample were calculated and basic descriptive statistics, such as the mean and standard deviation, were gathered to provide a description of the sample characteristi cs. Descriptive statistics for literacy development, within-child protective factor s and behavior also were calculated. See Table 8 for a summary of the data s ource and respective range of data that were used for analysis in the current study. Table 8. Description of Measurement Data for Cohort Three. Data Measured By Data Range Age of student Age in months at time of data collection 36 to 72 months Gender Gender of participant 1 = Male, 2 = Female Race Race of participant 1 = White, 2 = African American, 3 = Hispanic, 4 = Asian, 5 = Other Student Attendance Number of days participant attended school during the course of the study 0 to 64 days Home SES Median household annual income of the neighborhood in which the child resides 1 = $10,000-19,000; 2 = 20,000-29,000; 3 = 30,000-39,000 etc. School SES Median annual income of the neighborhood in which the childcare site is located 1 = $10,000-19,000; 2 = 20,000-29,000; 3 = 30,000-39,000 etc.
77 Table 8 (continued). Description of Measurement Data for Cohort Three. Data Measured By Data Range Class Size Highest number of students enrolled in a teacherÂ’s classroom 12 to 20 Teacher education Highest degree obt ained 1 = High School Diploma or GED, 2 = Some College, 3 = AA, 4 = 4 Year Degree Teacher experience Years of experience in early childhood setting 1-20 years Classroom environment Total score on the Early Childhood Checklist (ELOC) 0 to 41 Behavior Total score on the Behavior Concerns Scale of the DECA 30 to 70 Total protective factors Sum of the three subtest scores on the DECA 30 to 70 Early Literacy IGDI Â– Picture Naming IGDI Â– Rhyming IGDI Â– Alliteration 0 to 60 The current study sought to exami ne a complex integration of data consisting of various levels, all nested within one another. These levels include individual child factors (i.e., race, age, gender, and attendance), the preschool environment (i.e., class size, site SES, teacher experience, and classroom environment), and the behavioral competencie s of the child (i.e., typical or challenging), all of which hav e the potential to affect lit eracy development. Within this nested organization, preschool student s sharing the same classroom teacher increase group homogeneity and tend to be more similar to each other than preschool students selected from the popul ation at random. As Osborne (2000) explained, this similarity o ccurs as a natural consequence at two levels. First,
78 students within a particular classroom typi cally come from the same geographic location and, thus, are not randomly assi gned from the larger population (school district as a whole or from the nati onal population). Se cond, students assigned to a particular classroom also share the experience of being exposed to the same environment, which consists of the same teacher, instruction and physical surroundings (Osborne, 2000). Therefore, the lower level unit of analysis (literacy growth of preschool students) is influenc ed by the higher-level variables (teacher and classroom environment characteristics) While traditional approaches (e.g., multiple regression, Analysis of Co variance) to multilevel data analysis disaggregate higher-level variables and/ or aggregate lower level variables, the results are not without significant flaws (O sborne, 2000). Specific flaws of these approaches include lower levels of robus tness, violation of independence of observations, and potential under/overestim ation of observed relationships due to elimination of within-group information (Osborne, 2000). Since factors (e.g., teacher education, classroom environmen t, SES, gender) at each level influence each other (Hofmann, 1997) and have the potential of affecting outcomes (e.g., early literacy in preschool students), it becomes necessary to use a multilevel approach that enables the data to be s eparated based on indi vidual and group effects. Hierarchical Linear Modeling (HL M) offers the option to examine both the effects of level 1 and level 2 variables on the outcome, as well as cross-level interactions. Through this statistical pr ocess, relationships between predictors and outcomes can best be estimated (Osborne, 2000). This is accomplished by
79 holding each level constant to examine nest ed layers for their role in the outcome variable. For a visual description of the HLM design see Appendix H. As with any statistical procedure, assu mptions are required. Specific to HLM, Bryk and Raudenbush ( 1997) initially discussed the issue of normality, suggesting that both individual outcomes and growth parameters assume normal distributions. This assumption c an be validated through examination of histograms (for outcomes) and outliers (for growth parameters) Should outliers be present, analyses will be conducted with and without the observations to determine their contribution to the result s. Covariance structure is the second assumption considered (Bryk and R audenbush, 1997). HLM does not require identical data collection design for each subj ect, rather, the flexib ility of the model accepts varying numbers of data point s and spacing between observations. Therefore, HLM uses a covariance structur e that estimates erro r variance. That is, it considers random effects (Bryk and Raudenbush, 1997). Last, assumptions regarding the metric used to assess the outcome variable require that each observation be measured on a common metric to allow for change in growth across time as opposed to changes in the measurement scale. The first research question was devel oped to explore the contributing factors of positive and negative classr oom behavior on the rate of literacy development. That is, how is litera cy achievement in preschool children impacted by classroom behavior? The teacher ratings of classroom behavior were obtained through the Behavior Concer ns scale of the DECA and were used as the predictor variable. Scores on t he IGDI represented t he outcome variable
80 (literacy development). The individual, or base, level of the HLM analysis reflected the literacy growth of the pre school children over four data points in time. This initially addressed the rate and levels of literacy development for the preschool children based on monitored perfo rmance on the IGDI. Specifically, the literacy scores obtained at Time O ne represented the intercept of the regression equation, while the slope docum ented the growth as measured during subsequent time points. The equati on for Level one reads as follows: Yti = oi + 1i ati + eti where Yti is the outcome measure (literacy) and a is the age, both representing time t for the ith child, oi and 1i are intercepts and slopes estimated for the ith child and e is the amount of error. T he covariance structure for the errors is assumed to be as follows: 2 = 1 2 0 0 0 0 2 2 0 0 0 0 3 2 0 0 0 0 4 2 In this structure, independence and equal variance are implied. Level two of the HLM model reflects the individual factors related to the preschool child (i.e., race, gender, attendance and home SES) and begins to answer the second question of whether t here are individual level variables associated with the variation across t he individuals. In other words, each variable within Level two was examined to ascertain their contribution in literacy development, addressing such questions as, Â“do girls have higher literacy score
81 than boys?Â” Within this level, the intercept and slopes from the Level one analysis are utilized as dependent variabl es generating the following equations: 0i = 00 + 01(behavior) + 02 (gender) + 03 (race) + 04 (attendance) + 05 (home SES) + r0i 1i = 10 + 11(behavior) + 12 (gender) + 13 (race) + 14 (attendance) + 15 (home SES) + r1i An unstructured covariance matrix of cova riance structure is assumed at level two, representing the following model: roi Too To1 r1i T11 The third HLM level focused on the preschool classroom variables, hypothesized as contributing to childrenÂ’sÂ’ literacy outcomes. These variables, considered predictors, include school SE S, class size, teacher experience, teacher education, and classroom environment. Their inclusion in the regression model is as follows: 00 = G000 + G001 (school SES) + G002 (class size) + G003 (teacher experience) + G004 (teacher education) + G005 (classroom environment) + u00 10 = G100 + G101 (school SES) + G102 (class size) + G103 (teacher experience) + G104 (teacher education) + G105 (classroom environment) + u10 It is important to note the possible estimation issue surrounding the Level 3 analysis. The HLM model is based on large sample theory, so most recommend large numbers of units at the highest level. However, the current study is limited to 8 groups (preschool classrooms), which lead to noncovergence
82 or an inadmissible solution (e.g., a negative variance estimate). As a result, the aforementioned classroom variables were presented as a fixed effect and the level-3 errors were dro pped from the model. Additional analyses were conducted to address the questions related to the within-child protective fa ctors. The DECA was completed at one point in time (during the initial evaluation activities of Cohort Three) and served as the data source for the third, and final, research question. These data were analyzed via a two level structure with the child fact ors representing the first level and the preschool site representing the second le vel. The current study sought to examine the differences in childrenÂ’s lit eracy development based on within-child protective factors for children who were rated as having challenging behaviors on the DECA. Specifically, this analysis ex amined whether there are differences in literacy development for children with challenging behavi ors who have high scores measuring within-child protective fa ctors versus children with challenging behaviors who have low scores measuring withinchild protective fa ctors. A two level HLM model was donducted using a m odification of the structural equation as discussed above for questions one and two. The modification centered on the inclusion of the DECA score as a pr edictor, which measures within-child protective factors.
83 CHAPTER FOUR RESULTS The purpose of this study was to ex amine the relationship between early literacy development and behavioral difficulties in preschool children. The role of within-child protective fact ors in literacy development also was explored. The current chapter will present results from eac h of the three research questions. First, descriptive information will be provid ed on each of the variables examined. Second, results of the hierarchical linear modeling (HLM) analysis will be presented to describe the relati onships of interest. Descriptive Statistics Demographic characteristics for ch ild (i.e., gender, age, race) and teacher (i.e., education, experience) participants we re examined (Table 9). The sample contained approximately the same number of boys (53%) and girls (47%). Seventy-one percent of t he participants were White. The age of the child participants ranged from 38 to 62 months with 79% of the children between 48 and 59 months of age. Level of educati on and years of experience varied across the eight teacher participants. More spec ifically, three teachers (37.5%) reported completing Â‘Some College,Â’ whereas fifty percent of the t eachers completed a two-year (n=2) or four-year (n=2) degr ee. Only one teacher listed Â‘High SchoolÂ’ as the highest level of education obtained. A notable degree of variability was present in the amount of experience the teacher participants had in the
84 classroom setting. Three teachers (37. 5%) had up to one year of experience, while 2 teachers (25%) reported tw enty years in the field. Table 9. Descriptive Statistics related to Child and Teacher Demographi c Characteristics Demographic Characteristics N % Child Characteristics Gender Male 44 53.01 Female 39 46.99 Age (in months) 36-41 2 2.41 42-47 9 10.84 48-53 41 49.40 54-59 25 30.12 60-65 6 7.23 66-72 0 0.0 Race White 59 71.08 African-American 13 15.66 Hispanic 6 7.23 Asian 2 2.41 Other 3 3.61 Teacher Characteristics Education High School degree 1 12.5 Some college 3 37.5 Two year degree 2 25.0 Four year degree 2 25.0 Experience (in years) Up to one year 3 37.5 2 Â– 6 1 12.5 7 Â– 11 1 12.5 12-16 1 12.5 17-21 2 25.0 Additional demographic information is reported in Table 10. More specifically, the mean annua l income for the househo ld was $32,015.44 (SD = $6,678.24) in comparison to $28,940.65 (S D = $6,050.70) reported for the site
85 based on the GIS zip code mapping syst em. The average number of days absent was reported as 5. 46 (SD = 0.66) out of 68 possible school days. Table 10. Descriptive Statistics for Demographic Variables Demographic Characteristics Mean SD Attendance (n=83) 5.46 0.66 Site SES (n=83) 28940.65 6050.70 Home SES (n=66) 32015.44 6716.66 Next, the data were explored to determine normality. Normality was determined by obtaining skewness and ku rtosis of the dependent measures (Table 11). Skewness (a measure of sym metry) and kurtosis (degree of peaks or flatness) refer to t he extent to which the sample distribution departs from the normal curve (Hatcher & Stepanski, 1994) In general, a normal distribution yields skewness and kurtosis values of zero, whereas an obtained value greater or less than one indicates a non-normal samp le distribution. For the current study, notable deviations in skewness and kurtosis occurred for the Alliteration subtest across all four points in time. The values were positive for both, indicating a right-skewed (meaning that ther e are relatively few high scores) and leptokurtic (an acute peak with the majori ty of scores falli ng around the mean) distribution. A higher kurtosis suggests that the variance is due to infrequent extreme deviations. Therefore, a vi sual inspection of the raw data was conducted, revealing one or more outliers. Consequently, the data were run both with and without the outliers to assess the sensitivity of the results due to these observations. Outcomes of the analyses we re not influenced by these outliers.
86 Table 11. Skewness and Kurtosis Values for Dependent Measures N Skewness Kurtosis Time One Picture Naming 76-0.24 -0.15 Alliteration 761.68 2.16 Rhyming 761.12 0.13 Time Two Picture Naming 76-0.18 0.07 Alliteration 761.36 1.30 Rhyming 760.87 -0.31 Time Three Picture Naming 73-0.78 1.01 Alliteration 731.62 2.23 Rhyming 730.84 -0.71 Time Four Picture Naming 68-0.07 -0.17 Alliteration 681.64 2.72 Rhyming 680.70 -0.59 Means and standard deviations for the dependent variables were calculated across the four time points (Table 12). Scores on all subtests measuring literacy (i.e., Picture Naming, Alliteration, Rhyming) appeared to increase over time. Descriptive statis tics for the DECA and ELOC also were calculated (Table 12). Scores for the To tal DECA ranged from 31-66, with an average score of 49.66 ( SD =9.21) while scores of the Behavior component ranged from 37 to 72 ( M =49.65, SD =9.97). Sixteen percent of the study sample obtained a behavior score above sixty, which is the threshold for Â“elevatedÂ” on the rating scale. Observations of t he classroom environment using the ELOC yielded scores ranging from 16.75 to 40 ( M =23.33, SD =6.97) for the pre-test and a range of 29 to 57 ( M =45.86, SD =7.88) for the post test. There was a 4-point increase in mean scores from the pr eto the postmeasure.
87 Table 12. Means and Standard Deviations for Dependent Variables Mean Standard Deviation Time One (n=76) Picture Naming 20.25 0.60 Alliteration 2.63 0.49 Rhyming 4.54 0.67 Time Two (n=76) Picture Naming 20.70 0.82 Alliteration 3.50 0.53 Rhyming 5.11 0.67 Time Three (n=73) Picture Naming 21.93 0.79 Alliteration 3.42 0.60 Rhyming 5.51 0.77 Time Four (n=68) Picture Naming 23.24 0.84 Alliteration 4.88 0.76 Rhyming 6.69 0.84 Other Measures DECA (n = 77) Initiative 50.51 1.12 Se lf Control 53.70 1.21 Attention 46.90 0.85 Total 49.66 1.10 Behavior 49.65 1.19 ELOC Â– Pre 23.33 6.97 ELOC Â– Post 45.86 7.88 Linear graphs were constructed to view the relationships between behavior and the three IGDI subtests (Figur es 1 through 3). To illustrate these trends, the student sample was split in to two groups using the median DECA behavior score and was labeled as the low DECA behavior group and the high DECA behavior group. It is interesting to note that the ex pressive language and phonemic awareness skills as measured by Picture Naming, Alliteration and Rhyming were lower for those children who had high scores on the behavior
88 scale. Conversely, children with lower scores on the behavior scale obtained higher scores on the three literacy measures. 0 5 10 15 20 25 Time 1Time 2Time 3Time 4TimeMean Picture Naming Score Low High Figure 1 Behavior and Picture Naming Scores.
89 0 1 2 3 4 5 6 7 Time 1Time 2Time 3Time 4TimeMean Alliteration Score Low Behavior High Behavior Figure 2 Behavior and Alliteration Scores.
90 0 1 2 3 4 5 6 7 8 9 Time 1Time 2Time 3Time 4TimeMean Rhyming Score Low Behavior High Behavior Figure 3 Behavior and Rhyming Scores. Linear graphs depicting the relationship between race and behavior on literacy development, as measured by the three IGDI subtests, also were constructed (Figures 4 through 6) due to t he tendency of the variables to covary. To explain further, research suggests t hat culturally and linguistically diverse students exhibit more externalizing behavio rs (Sbarra and Pianta, 2001; Epstein, March, Conners, and Jackson, 1998; Ma cMillan et al., 1996) and have lower performance on reading achievement (Diamond and Onwuegbuzie, 2001) than non-diverse students. Overall, the graphs indicate that White students have higher scores, regardless of behavior status It is important to note, however,
91 that the number of stu dents in each group varies with forty-one White students and 19 non-White students in the low behavior groups, along with nine and two students respectively in the high behavior groups. While discrepancies are noted, it is critical to explore these relationships while controlling for each variable to ensure the appropriate conc lusions are made regarding their relationship to early literacy. A mo re complex analysis was conducted and will be discussed later in the chapter. 0 5 10 15 20 25 30 Time OneTime TwoTime ThreeTime FourTimeMean Picture Naming Score White/Low Behavior White/High Behavior Non-White/Low Behavior Non-White/High Behaivor Figure 4 Relationship between Behavior and Race on Picture Naming Scores.
92 0 1 2 3 4 5 6 7 8 Time OneTime TwoTime ThreeTime FourTimeMean Alliteration Score White/Low Behavior White/High Behavior Non-White/Low Behavior Non-White/High Behaivor Figure 5 Relationship between Behavior and Race on Alliteration Scores.
93 0 1 2 3 4 5 6 7 8 9 10 Time OneTime TwoTime ThreeTime FourTimeMean Rhyming Score White/Low Behavior White/High Behavior Non-White/Low Behavior Non-White/High Behaivor Figure 6 Relationship between Behavior and Race on Rhyming Scores. Correlations between the continuous pr edictor variables and the outcome variable were conducted (Tables 13 through 16). Results are reported based on each data time point. Time One Picture Naming was negatively and significantly correlated with attendance (r = -.28, p<.05). This suggests that Picture Naming scores decreased as number of days absent increased. No other correlations to Picture Naming were revealed at time one. A lliteration was positively and significantly correlated to home SES (r=.38, p<.01) site SES (r=.43, p<.001), teacher experience (r=.25, p<.05), cl ass size (r=.28, p<.05), and within-child protective factors (r=.46, p<.01). A significant negative correlation (r=-.40, p< .001) was found between Alliteration and Behavior. The third subtest, Rhyming, was
94 positively and significantly correlated with home SES (r=.37, p<.01) site SES (r=.38, p<.001), teacher experience (r-.25, p<.05), class size (r=.26, p<.05), and within-child protective fact ors (r=.39, p<.01). Resu lts revealed a significant negative correlation between Rhyming and t he predictor variables of Behavior (r=-.35, p<.01) and the ELOC pos ttest (r=-.27, p<.05). Time Two Picture Naming was positively and significantly correlated with home SES (r=.37, p<.01), site SES (r=.32, p<.01), teacher experience (r=.40, P<.001), class size (r=.39, p<.001), and within-child protective factors (r=.31, p<.05) at time two. There was a negative corre lation (r=-.35, p<.01) to attendance. Alliteration was positively and significantly correlated to home SES (r=.48, p<.001), site SES (r=.45, p<.001) class size (r=.29, p<.05), and withinchild protective factors (r=.37, p<.01) There was a negative correlation (r=-37, p<.01) to Behavior and the ELOC pretes t (-.26, p<.05). Rhyming was positively and significantly correlated to home SES (r=.44, p<.001), site SES (r=.39, p<.001), teacher experience (r=.32, p<.01), class size (r=.33, p<.01), within-child protective factors (r=.41 p<.001), and the ELOC postte st (r=.28, p<.05). In addition, Rhyming was negatively correla ted (r=-.39, p<.001) with behavior. Time Three The results for time thr ee do not reveal significant correlations between Picture Naming and the other predictor variables. However, Alliteration was found to be pos itively correlated with home SES (r=.34, p<.01), site SES (r=.45, p< .001), class size (r=.23, p<.05), and within-child protective factors (r=.39, p<.01). A negative correlation (r=-.42, p<.001) was found between Alliteration and Behavior as well as the ELOC pretest (r=-.33,
95 p<.01). Positive correlations were revealed between Rhyming and home SES (r=.32, p<.01), site SES (r=.48, p<.001) class size (r=.44, p<.001), and withinchild protective factors (r=.42, p<. 001). The correlation between Rhyming and Behavior was negative (r=-.35, p<.01), along with the ELOC pretest (r=-.31, p<.01). Time Four As with time two, the finding revealed a significantly positive correlation between Picture Naming and home SES (r=.34, p<.01), site SES (r=.28, p<.05), teacher experience (r=.40, p<.001), class size (r=.26, p<.05). A negative relationship with attendance was found to be significant (r=-.35, p<.001). Alliteration was positively and signi ficantly correlated to site SES (r=.38, p<.001), teacher experience (r=.26, p<.05) class size (r=.37, p<.01) and withinchild protective factors (r=.57, p<. 01). Alliteration was negatively and significantly correlated to the Behavior score (r=-.32, p<.01). Finally, Rhyming was positively and significantly correla ted with home SES (r=.3 8, p<.01), site SES (r=.40, p<.001), teacher experience (r=.37, p<.01), class size (r=.46, p<.001). and within-child protecti ve factors (r=.35, p<.01). Hierarchical Linear Modeling Hierarchical Linear Modeling (HLM) was utilized to examine nested relationships within various levels using SPSS. The nested data structure consisted of three levels that were explor ed to ascertain their contribution to early literacy development. Level one addre ssed the differences in literacy development of preschool children over ti me as measured by scores on the IGDI.
96 Table 13. Correlations between Predictors and Lite racy Outcomes at Time One. Variable 1 2 3 4 5 6 7 8 9 10 11 1 Picture naming 2 Alliteration 3 Rhyming 4 DECA Behavior 5 DECA Protective 6 Home SES 7 Attendance 8 Site SES 9 Teacher Experience 10 Class size 11 ELOCPretest 12 ELOC-Posttest .43 .35 -.12 .14 .20 -.28 .20 .22 .19 -.02 .10 ** ** .62 -.40 .46 .38 -.15 .49 .20 .28 .27 .19 ** ** ** ** ** * -.35 .39 .37 -.19 .38 .25 .26 -.06 .25 ** ** ** ** * -.82 -.48 .28 -.46 .12 .05 .24 -.52 ** ** ** ** .47 -.35 .41 .32 .02 -.00 .63 ** ** ** ** ** -.27 .58 .53 .34 -.09 .31 ** ** ** ** -.24 -.32 -.02 -.08 -.33 ** ** .36 .54 -.36 .46 ** ** ** ** .59 .54 .49 ** ** .05 .07.45 ** Note : p < .05; ** p < .01
97 Table 14. Correlations between Predictors and Literacy Outcomes at Time Two. Variable 1 2 3 4 5 6 7 8 9 10 11 1 Picture naming 2 Alliteration 3 Rhyming 4 DECA Behavior 5 DECA Protective 6 Home SES 7 Attendance 8 Site SES 9 Teacher Experience 10 Class size 11 ELCOPretest 12 ELOC-Posttest .40 .49 -.18 .31 .37 -.35 .32 .40 .39 .11 .21 ** ** ** ** ** ** ** ** .65 -.32 .37 .48 -.20 .45 .14 .29 -.26 .10 -.39 .41 .44 -.20 .39 .32 .33 -.01 .28 ** ** ** ** ** ** -.82 -.48 .28 -.46 -.12 .05 .24 -.52 ** ** ** ** .47 -.35 .41 .32 .02 -.00 .63 ** ** ** ** ** -.27 .58.53 .34 -.09 .31 ** ** ** -.24 -.32 -.02 -.08 -.33 ** ** .36 .54 -.36 .46 ** ** ** ** .59 .54 .49 ** ** ** .05 .07.45 ** Note : p < .05; ** p < .01
98 Table 15. Correlations between Predictors and Literacy Outcomes at Time Three. Variable 1 2 3 4 5 6 7 8 9 10 11 1 Picture naming 2 Alliteration 3 Rhyming 4 DECA Behavior 5 DECA Protective 6 Home SES 7 Attendance 8 Site SES 9 Teacher Experience 10 Class size 11 ELCOPretest 12 ELOC-Posttest .36 .41 -.06 .21 .18 -.12 .19 .04 .15 -.15 -.05 ** ** .62 -.42 .39 .34 -.23 .45 .01 .23 -.33 .16 ** ** ** ** ** ** -.35 .42 .32 -.08 .48 .19 .44 -.31 .06 ** ** ** ** ** ** -.82 -.48 .28 -.46 -.12 .05 .24 -.52 ** ** ** ** .47 -.35 .41 .32 .02 -.00 .63 ** ** ** ** ** -.27 .58 .53 .34 -.09 .31 ** ** ** -.24 -.32 -.02 -.08 -.33 ** ** .36 .54 -.36 .46 ** ** ** ** .59 .54 .49 ** ** ** .05 .07 .45 ** Note : p < .05; ** p < .01
99 Table 16. Correlations between Predictors and Literacy Outcomes at Time Four. Variable 1 2 3 4 5 6 7 8 9 10 11 1 Picture naming 2 Alliteration 3 Rhyming 4 DECA Behavior 5 DECA Protective 6 Home SES 7 Attendance 8 Site SES 9 Teacher Experience 10 Class size 11 ELCOPretest 12 ELOC-Posttest .30 .33 -.11 .21 .34 -.18 .28 .40 .26 .11 .22 ** * ** .57 -.32 .41 .27 -.19 .38 .26 .37 -.12 .20 ** ** ** ** -.26 .35 .38 -.09 .40 .37 .46 -.06 .15 ** ** ** ** ** -.82 -.48 .28 -.46 -.12 .05 .24 -.52 ** ** ** ** .47 -.35 .41 .32 .02 -.00 .63 ** ** ** ** ** .27 .58 .53 .34 -.09 .31 ** ** ** -.24 -.32 .02 -.08 -.33 ** ** .36 .54 -.36 .46 ** ** ** ** .59 .54 .49 ** ** ** .05 .07.45 ** Note : p < .05; ** p < .01
100 Level two examined child vari ables that were identifi ed as potentially affecting literacy development. These variables inclu ded the child participantsÂ’ (1) age, (2) race, (3) gender, (4) home SES, (5) behavior and (6) attendance in school. Level three explored how the childcare site vari ables such as (1) class size, (2) years of teaching experience, (3) highest level of education earned by the preschool teachers, (4) site SES, and (5) cla ssroom environment, influenced literacy development. Intraclass Correlations Variance estimates fo r the unconditional two level models were examined first (Table 17). Specifically, intraclass correlation coefficients (ICCÂ’s) were obtained to meas ure the proportion of the variance in outcome between and within pers ons. ICC values range fr om 0 to 1, indicating complete within person variability or complete between person variability respectively. For the current study, ICCÂ’s ranged from .53 to .69 for the two level models, suggesting that the majority of the variability is attributed to between person variables as opposed to within perso n variables for all three measures of literacy. Table 17. Intraclass Correlation Coefficients Dependent Measures ICC Two level unconditional model Picture Naming .53 Alliteration .60 Rhyming .69 It was anticipated that individual diffe rences in the level one model were impacted by time. That is, child par ticipants were expected to demonstrate
101 growth in literacy scores across the four data points. Thus, the first analysis addressed research question number one: How does pos itive and negative classroom behavior contribute to the rate of literacy developm ent on preschool children? The Picture Naming growth m odel findings indicate that for each unit increase in time, Picture Naming growth increased, on average, by .87 points. The Alliteration and Rhyming growth model s yielded findings that indicate a growth increase of .66 on average for Allit eration and .69 points for Rhyming. Table 18 depicts the within-ch ild differences at level one for each subtest. Table 18. Linear Model of Literacy Growth Outcome Variables Average Intercept Average Slope Picture Naming Mean Score 21.30 Time .87 Alliteration Mean Score 3.55 Time .66 Rhyming Mean Score 5.37 Time .69 In addition, a random sample of child participants was taken to illustrate the relationship between time and growth on scores from the Picture Naming, Alliteration and Rhyming subtests (Figures 7 through 9). While growth in literacy development over time was supported, the variance in the rate of growth remains in question. This will be addressed later in this chapter.
102 0 5 10 15 20 25 30 35 40 45 1234Data PointPicture Naming Scores Figure 7 Growth Over Time for a Random Selection of Student Participants on the Picture Naming Subtest.
103 0 5 10 15 20 25 30 1234Data PointAlliteration Score Figure 8 Growth Over Time fo r a Random Selection of Student Participants on the Alliteration subtest.
104 0 2 4 6 8 10 12 14 16 18 1234Date PointRhyming Score Figure 9 Growth Over Time for a Random Selection of Student Participants on the Rhyming subtest. Next, six level two variables invest igating child characteristics were entered into the HLM model. These included age, gender, race, attendance, home socioeconomic status (SES), and behav ior are summarized in Table 19 for each of the three outcome variables. T he analysis at this level addressed the second research question: What fact ors (i.e., gender, ra ce, SES, teacher experience, classroom environment, class si ze) contribute to the rate and levels of literacy development for children i dentified with typical or challenging behaviors? Overall, the majority of the student level variables failed to significantly predict literacy scores and sl opes for the three subtests. Findings did, however, reveal significant relations hips pertaining to Picture Naming scores
105 and race and attendance. Specifically, it was noted that White students had significantly higher Picture Naming score s than non-White students. In addition, the scores of White students increased more across time in comparison to scores of non-White students. The se cond finding revealed a significant and negative relationship between attendance and Picture Naming scores. That is, the more often students were absent, t he lower their Picture Naming scores were. In examining scores over time results suggest that preschool studentsÂ’ Picture Naming and Rhyming scores incr eased significantly across time, a finding that was not evident with Alliterati on scores. Also worth noting is the lack of significant relationship between t he behavior variable and literacy development for each subtest. Table 19. Linear Model of Literacy Growth in cluding Child Characteristics Outcome Variables Predictor Parameters for Fixed Effects Standard Error p-value Picture Naming Intercept 20.76 .62 .000 Age .03 1.44 .984 Gender a -1.85 1.20 .132 Race b -4.02* 1.56 .013 Attendance -.28** .10 .010 Home SES .00 .00 .289 DECA Behavior .00 .08 .978 Time .71* .29 .016 Age Time -.53 .66 .426 Gender Ti me -.24 .56 .671 Race Ti me -1.90* .74 .014 Attendance Time -.02 .05 .707 Home SES Time 4.80 5.55 .392 DECA-Behavior Time .03 .04 .392 Note : p < .05; ** p < .01, aGender (0=Male, 1=Female), bRace (0=White, 1=Non-White)
106 Table 19 (Cont). Linear Model of Literacy Growth in cluding Child Characteristics Outcome Variables Predictor Parameters for Fixed Effects Standard Error p-value Alliteration Intercept 3.24 .52 .000 Age -.17 1.22 .891 Gender .54 1.02 .599 Race -1.45 1.32 .279 Attendance -.13 .09 .147 Home SES .00 .00 .080 DECA Behavior -.11 .07 .097 Time .35 .24 .161 Age Time .73 .57 .206 Gender Ti me -.27 .48 .580 Race Time .01 .62 .986 Attendance Time -.03 .04 .491 Home SES Time -3.36 4.72 .480 DECA-Behavior Time -.04 .03 .185 Rhyming Intercept 5.11 .63 .000 Age 1.35 1.47 .364 Gender 2.07 1.24 .101 Race -2.52 1.59 .120 Attendance -.04 .10 .724 Home SES .00 .00 .071 DECA Â– Behavior -.10 .08 .209 Time ..83** .29 .007 Age Time .42 .68 .536 Gender Ti me -.69 .57 .235 Race Time -.66 .75 .386 Attendance Time -.01 .05 .786 Home SES Time 4.28 5.69 .940 DECA-Behavior Time -.01 .04 .859 Note : p < .05; ** p < .01, aGender (0=Male, 1=Female), bRace (0=White, 1=Non-White) The variances of scores from Pictur e Naming, Alliteration, and Rhyming within each of the four assessm ent periods were different (Table 20). That is, the mean scores for each subtest varied signifi cantly across students. Additionally, the slopes for Alliteration and Rhyming va ried significantly across students, but
107 not for Picture Naming. Finally, the change in mean Rhyming score covaried significantly with the change in Rhyming slo pe. Significant findings of covariance were not found for Picture Naming or Alliteration. Table 20. Linear Model of Literacy Growth includ ing Child Characteristics: Variance Estimates Outcome Variables Parameter Estimate for Random Effects Standard Error p-value Picture Naming Within Students Time One 4.68 4.63 .312 Time Two 24.23** 5.53 .000 Time Three 11.34** 3.53 .001 Time Four 12.88* 6.06 .033 Between Students Mean Score 15.61** 4.01 .000 Time Slope 1.99 1.23 .106 Mean Score x Time Slope 2.17 1.54 .160 Alliteration Within Students Time One 2.40 2.57 .351 Time Two 6.92** 1.92 .000 Time Three 5.01** 1.80 .005 Time Four 10.79** 4.21 .010 Between Students Mean Score 11.92** 2.93 .000 Time Slope 1.68* .76 .028 Mean Score x Time Slope 1.99 1.14 .082 Rhyming Within Students Time One 8.03** 3.00 .007 Time Two 7.58** 2.05 .000 Time Three 6.98** 2.26 .002 Time Four 11.37* 4.51 .012 Between Students Mean Score 17.50** 4.14 .000 Time Slope 2.02* .99 .041 Mean Score x Time Slope 3.61* 1.48 .015 Note : p < .05; ** p < .01
108 Next, the four variables that exam ined classroom characteristics were entered into the HLM, creating the third le vel of the model (Table 21). Initial attempts at running the third level model resulted in non-convergence. A series of modifications were made to simplify t he variance structure. Specifically, the initial proposed structure was as follows: Level One: Yti = oi + 1i ati + eti Level Two: 0i = 00 + 01(behavior) + 02 (gender) + 03 (race) + 04 (attendance) + 05 (home SES) + r0i 1i = 10 + 11(behavior) + 12 (gender) + 13 (race) + 14 (attendance) + 15 (home SES) + r1i Level 3: 00 = G000 + G001 (school SES) + G002 (class size) + G003 (teacher experience) + G004 (teacher education) + G005 (classroom environment) + u00 10 = G100 + G101 (school SES) + G102 (class size) + G103 (teacher experience) + G104 (teacher education) + G105 (classroom environment) + u10 It was simplified to: Level One: Yti = oi + 1i ati + eti Level Two:
109 0i = 00 + 01(behavior) + 02 (gender) + 03 (race) + 04 (attendance) + 05 (home SES) + r0i 1i = 10 + 11(behavior) + 12 (gender) + 13 (race) + 14 (attendance) + 15 (home SES) + r1i Level Three: 00 = G000 + G001 (school SES) + G002 (class size) + G003 (teacher experience) + G004 (teacher education) + G005 (classroom environment) 10 = G100 + G101 (school SES) + G102 (class size) + G103 (teacher experience) + G104 (teacher education) + G105 (classroom environment) Although the model was conceptualized as a 3-level model, the removal of the error term at level three reduced it to a 2-level structure. In addition, a common variance was assumed for each m easurement occasion, leading to a single variance estimate at level one. The model was further simplified by merging two variables into one. That is, the SPSS output re vealed that the variable representing the posttest of the ELOC was considered redundant, and therefore did not add any additional informa tion to the analysis. As such, the average scores between the preand post-te sts were used as a way to collapse ELOC1 and ELOC2 into one variable. The relationship between the classroom environment and literacy was the concept of interest as opposed to the change in environment, making the collapse in vari ables logical. Although convergence was achieved, findings were non-signific ant for almost all predictors. The exception was a marginal relationship (p=.056) between level of teacher education and Alliteration scores. Thus, the higher the level of education the
110 teacher attained, the higher the Allite ration score. Additionally, there was a significant negative relationship betw een the Picture Naming slope and class size. The findings indicate that scores increase as the number of students in a class decrease. Table 21. Linear Model of Literacy Growth includi ng Child and Classroom Characteristics Outcome Variables Predictor Parameters for Fixed Effects Standard Error p-value Picture Naming Intercept 33.35*** 8.55 .000 Site SES -5.70 .00 .807 Teacher Experience .12 .38 .764 Teacher Education -.53 2.64 .842 Class Size .48 .56 .402 Class Environm ent -.35 .22 .124 Site SES Time -.00 .00 .325 Teacher Experience Time .31 .19 .105 Teacher Educatio n Time -.49 1.37 .721 Class Size Time -.55* .27 .050 Class Environment Time .15 .11 .174 Alliteration Intercept 1.41 6.88 .838 Site SES .00 .00 .236 Teacher Experience .04 .31 .886 Teacher Education 4.13 2.11 .056 Class Size -.36 .45 .427 Class Environment .07 .18 .708 Site SES Time -1.09 .00 .919 Teacher Experienc e Time -.14 .17 .411 Teacher Educatio n Time 1.98 1.25 .121 Class Size Time -.01 .25 .963 Class Environment Time .06 .10 .532 Rhyming Intercept 6.32 8.11 .440 Site SES .00 .00 .416 Teacher Experience .04 .36 .923 Teacher Education 4.32 2.48 .088 Class Size -.04 .53 .941 Class Environm ent -.02 .21 .942 Note : p < .05, *** p < .001
111 Table 21 (continued). Linear Model of Literacy Growth includi ng Child and Classroom Characteristics Outcome Variables Predictor Parameters for Fixed Effects Standard Error p-value Site SES Time -.00 .00 .180 Teacher Experience Time .16 .19 .408 Teacher Educatio n Time .29 1.36 .832 Class Size Time -.03 .27 .921 Class Environment Time .07 .12 .530 Note : p < .05, *** p < .001 Variance estimates for the third level also were calculated (Table 22). The mean score for each subtest varied signifi cantly across students. Additionally, the slope for the Picture Naming, Allit eration and Rhyming subtests varied significantly across students. Table 22. Linear Model of Literacy Growth includi ng Child and Classroom Characteristics: Variance Estimates Outcome Variables Parameter Estimate for Random Effects Standard Error p-value Picture Naming Within Students 15.51*** 2.32 .000 Between Students Mean Score 13.41*** 2.32 .000 Time Slope .35** 3.93 .001 Alliteration Within Students 6.02*** .88 .000 Between Students Mean Score 9.77*** 2.51 .000 Time Slope 1.84* .74 .011 Rhyming Within Students 8.20*** 1.21 .000 Between Students Mean Score 13.68*** 3.50 .000 Time Slope 1.94* .93 .037 Note : p < .05; ** p < .01, *** p < .001 The next analysis addressed litera cy development in children with challenging behaviors with and without the presence of within-child protective
112 factors (Table 23). Research question num ber three (What differences are there between literacy development in children with challenging behaviors who have high scores measuring within-child protecti ve factors in comparison to children with challenging behavior s who have low scores measuring within-child protective factors?) was the focal point in this analysis. Picture Naming, Alliteration and Rhyming scores did not incr ease significantly over time. Several significant findings were noted pertaining to the predictor variables and will be reported based on the indivi dual IGDI subtests Picture Naming. Results revealed significantly higher Picture Naming scores for White students ( M =22.97, SD =6) as compared to their non-White peers ( M =17.63, SD =6.63). In additi on, a significant and negative relationship was found between Picture Naming score s and attendance. That is, as absences increased, the Picture Naming score decreased, suggesting that the more often students were absent, the lowe r their Picture Naming scores were. Further, behavior had a signifi cant relationship with the Picture Naming slope. That is, children who had high scores on the DECA behavior scale demonstrated an increase in Picture Naming score ov er time, whereas children who had low Table 23. Linear Model of Literacy Growth including Child Characteristics Outcome Variables Predictor Parameters for Fixed Effects Standard Error p-value Picture Naming Intercept 20.36 .87 .000 Age .01 1.58 .997 Gender -1.85 1.21 .135 Race -3.35* 1.64 .047 Attendance -.25* .11 .023 Note : p < .05; ** p < .01
113 Table 23 (continued). Linear Model of Literacy Growth including Child Characteristics Outcome Variables Predictor Parameters for Fixed Effects Standard Error p-value Home SES .00 .00 .323 DECA Â– Behavior .12 .12 .330 DECA Â– Protective .18 .13 .195 Behavior x Prot ective -.00 .01 .810 Time .51 .39 .197 Age Time -.52 .69 .456 Gender Time -.26 .53 .626 Race Time -1.32 .74 .081 Attendance Time .01 .05 .916 Home SES Time 3.73 5.29 .485 DECA-Behavior Time .14* .05 .013 DECA Â– Protective* Time .16* .06 .011 Behavior x Protecti ve Time -.00 .00 .975 Alliteration Intercept 1.56 .65 .021 Age -1.76 1.19 .147 Gender .84 .91 .365 Race -.93 1.23 .450 Attendance -.11 .08 .162 Home SES .00* 9.09 .021 DECA Â– Behavior -.04 .09 .640 DECA Â– Protective .15 .10 .139 Behavior x Prot ective -.02** .01 .002 Time .04 .35 .902 Age Time .41 .63 .512 Gender Time -.29 .48 .546 Race Time .01 .66 .992 Attendance Time -.04 .04 .398 Home SES Time -3.08 4.81 .526 DECA-Behavior Time -.04 .05 .391 DECA Â– Protecti ve Time .00 .05 .941 Behavior x Protecti ve Time -.00 .00 .251 Rhyming Intercept 3.73 .83 .000 Age .18 1.53 .905 Gender 2.33* 1.17 .053 Race -1.75 1.57 .271 Attendance .00 .10 .992 Home SES .00* .00 .036 DECA Â– Behavior .02 .12 .856 Note : p < .05; ** p < .01
114 Table 23 (continued). Linear Model of Literacy Growth including Child Characteristics Outcome Variables Predictor Parameters for Fixed Effects Standard Error p-value DECA Â– Protective .23 .13 .086 Behavior x Prot ective -.02 .01 .072 Time .82 .43 .061 Age Time .46 .76 .554 Gender Time -.71 .59 .237 Race Time -.63 .81 .439 Attendance Time -.01 .05 .837 Home SES Time 3.91 5.88 .947 DECA-Behavior Time .01 .06 .923 DECA Â– Protecti ve Time .02 .07 .776 Behavior x Protective Time .00 .01 .899 Note : p < .05; ** p < .01 scores on the behavior scale had Picture Na ming scores that were similar across time. No significant relationships were found between behavior and Alliteration or Rhyming. Lastly, the DECA within -child protective factors also had a significant relationship with the Picture Naming Slope. Students who were rated as having high within-child protective factors had Picture Naming scores that increased over time; however, those ch ildren who were rated as having low within-child protective fact ors had Picture Naming scores that did not increase as much over time. The sample of student s was split into two groups using the median DECA protective factors score, yielding a low and high within-child protective factors group.
115 0 5 10 15 20 25 Time 1Time 2Time 3Time 4TimeMean Picture Naming Score Low High Figure 10 Relationship between Within-Child Protective Factors and the Picture Naming Slope. Alliteration. Results pertaining to the Alliteration subtest revealed a significant and positive relationship with Home SES. That is, the higher the Home SES of the student, the hi gher their Alliteration score ( p < .02). In addition, behavior moderated the relationship between within-child protective factors and Alliteration scores. As shown is Figure 11, when children were rated as having behavior issues, within-child protective factors did not have much of a relationship with the mean Alliteration score However, when children were rated as not having behavior issues, within-ch ild protective factors had a positive relationship with the mean score of the subtes t. It is important to note that the mean Alliteration scores were higher fo r those children with high ratings for
116 within-child protective fa ctors regardless of behavior ratings as compared to those children who were ranked with low within-child protecti ve factors. 0 5 10 15 20 25 30 LowHighBehaviorMean Alliteration Score Low Protective Factors High Protective Factors Figure 11 The Moderating Effect of Within -Child Protective Factors on the Relationship between Behavior and Alliteration. Rhyming. Gender and Home SES we re found as having a significant relationship with Rhyming scores over time. Specifically, girls ( M =7.02. SD =6.90) had higher scores than boys ( M =4.03, SD =5.37, p<.05 ). Further, a significant and positive relationship was noted with Home SES, suggesting that the higher the Home SES of the studen t, the higher their Rhyming score ( p < .04). The variances of scores from Pict ure Naming, Alliter ation, and Rhyming within each of the four a ssessment periods are present ed (Table 24). Findings indicate that the varian ce of scores were different within each assessment
117 period. That is, the mean score varied si gnificantly across students. The slopes for the Alliteration and Rhyming subtests varied significantly across students, a trend not noted for Picture Naming. Las tly, the change in mean Rhyming score covaried significantly with the change in Rh yming slope. Significant covariance parameters were not found for Pict ure Naming or Alliteration. Table 24. Linear Model of Literacy Growth includ ing Child Characteristics: Variance Estimates Outcome Variables Parameter Estimate for Random Effects Standard Error p-value Picture Naming Within Students Time One 3.72 4.62 .421 Time Two 24.56*** 5.58 .000 Time Three 10.68*** 3.29 .001 Time Four 14.51* 6.02 .016 Between Students Mean Score 15.88*** 4.13 .000 Time Slope 1.70 1.17 .148 Mean Score x Time Slope 1.60 1.51 .290 Alliteration Within Students Time One 3.04 2.60 .243 Time Two 6.43*** 1.83 .000 Time Three 5.40** 1.88 .004 Time Four 10.60* 4.34 .015 Between Students Mean Score 8.93*** 2.39 .000 Time Slope 1.64* .79 .039 Mean Score x Time Slope 1.70 1.04 .102 Rhyming Within Students Time One 8.02** 2.97 .007 Time Two 7.67*** 2.04 .000 Time Three 6.98** 2.28 .002 Time Four 11.15* 4.56 .015 Between Students Mean Score 15.34*** 3.77 .000 Note : p < .05; ** p < .01, *** p < .001
118 Table 24 (continued). Linear Model of Literacy Growth includ ing Child Characteristics: Variance Estimates Outcome Variables Parameter Estimate for Random Effects Standard Error p-value Time Slope 2.27* 1.05 .031 Mean Score x Time Slope 3.77* 1.48 .011 Note : p < .05; ** p < .01, *** p < .001 Next, the five variables that examined classroom characteristics were entered into the HLM, creating the final leve l of the model (Table 25). As with the previous research question, initial a ttempts at running the third level model resulted in non-convergence. The sa me procedure was followed to obtain convergence. Findings were non-significant for all predictors with the exception of two. Class size and environment had a significant relationship with Picture Naming scores. More specifically, as the number of student s within a classroom increase, the scores on the Picture Naming subtest decrease. Additionally, as the scores on the ELOC (measure of cl assroom environment) increase, scores on the Picture Naming subtest increase. Table 25. Linear Model of Literacy Growth includi ng Child and Classroom Characteristics Outcome Variables Predictor Parameters for Fixed Effects Standard Error p-value Picture Naming Intercept 46.96 8.89 .000 Site SES 3.27 .00 .881 Teacher Exper ience -.21 .37 .577 Teacher Education -2.58 2.55 .318 Class Size 1.28* .57 .031 Class Environment .74** .24 .003 Site SES Time -.00 .00 .324 Teacher Experience Time .22 .20 .259 Teacher Educatio n Time -1.58 1.47 .291 Note : p < .05; ** p < .01, *** p < .001
119 Table 25 (continued). Linear Model of Literacy Growth includi ng Child and Classroom Characteristics Outcome Variables Predictor Parameters for Fixed Effects Standard Error p-value Picture Naming Class Size Time -.22 .31 .480 Class Environment Time .02 .13 .853 Alliteration Intercept 5.79 7.19 .425 Site SES .00 .00 .126 Teacher Exper ience -.13 .30 .667 Teacher Education 2.73 2.05 .189 Class Size -.02 .46 .960 Classroom Envir onment -.09 .19 .638 Site SES Time -4.79 .00 .964 Teacher Experienc e Time -.15 .18 .392 Teacher Educatio n Time 2.14 1.34 .116 Class Size Time -.06 .29 .844 Class Environment Time .08 .12 .482 Rhyming Intercept 14.01 8.82 .120 Site SES .00 .00 .268 Teacher Exper ience -.17 .37 .636 Teacher Education 2.89 2.51 .257 Class Size .44 .57 .441 Classroom Envir onment -.25 .23 .298 Site SES Time -.00 .00 .172 Teacher Experience Time .18 .20 .367 Teacher Educatio n Time .36 1.48 .811 Class Size Time -.04 .32 .909 Class Environment Time .07 .13 .576 Note : p < .05; ** p < .01, *** p < .001 Variance parameters for the third le vel of the model also were obtained (Table 26). Findings reveal that vari ances in scores were different between students for each subtest at each assessment period. That is, the mean score varied significantly across students. The slopes for the Alliteration and Rhyming subtests varied significantly across students, a trend not noted for Picture Naming.
120 Table 26. Linear Model of Literacy Growth including Child and Classroom Characteristics: Variance Estimates Outcome Variables Parameter Estimate for Random Effects Standard Error p-value Picture Naming Within Students 15.23*** 2.27 .000 Between Students Mean Score 10.88*** 3.52 .001 Time Slope .43 1.05 .686 Alliteration Within Students 6.05*** .89 .000 Between Students Mean Score 8.21*** 2.23 .000 Time Slope 1.91* .78 .014 Rhyming Within Students 8.19*** 1.21 .000 Between Students Mean Score 12.72*** 3.38 .000 Time Slope 2.13* .99 .031 Note : p < .05; ** p < .01, *** p < .001 Summary In conclusion, few predictors emerged in this study as having a significant relationship with literacy development in pre school children. In the first analysis, race and attendance were the significant pr edictors noted in relation to Picture Naming scores. In addition, a significant amount of variance was noted between students for the mean scores of all three areas of literacy development. Notable variance between students regarding slope was evident for Alliteration and Rhyming only. The third level of the anal ysis addressed site characteristics in relation to literacy development in preschoo l children. Findings were insignificant for all predictors, except Teacher Educ ation for the Alliteration subtest. The average score of each subtest varied signi ficantly between st udents at the third level. Change in slope across students varied for Alliteration and Rhyming.
121 For the second HLM analysis, the leveltwo child characteristics of age and attendance were significant predictor s for Picture Naming. Significant predictors for Alliteration were Ho me SES and the behavior/within-child protective factors interaction. The child characteristic of Home SES also was a significant predictor related to Rhymi ng along with gender. A significant amount of variance was revealed between students for Picture Naming, Alliteration, and Rhyming. Significant variance su rrounding the slopes across students was evident for Alliteration and Rhyming. Cla ssroom variables were explored at the third level of the analysis. Class size and environment were noted as having a positive and significant relationshi p with expressive language scores as measured by the Picture Naming subtes t. Variances in scores suggest a significant difference in means for all subtests across students, while slopes for the Alliteration and Rhyming subtests were noted as varying significantly.
122 CHAPTER FIVE DISCUSSION The purpose of this study was: (1 ) to examine the relationship between early literacy development in preschool ch ildren as it relates to challenging behavior, and (2) to explore the role of wit hin-child protective factors in early literacy for children rated with and without challenging behaviors. The current chapter will provide a synopsis of the results and will discuss the findings in response to the three research questions and in the context of existing research. Implications of this study, limitations and suggestions for future research will be addressed. Responses to Research Questions Research Question #1: How does pos itive and negative classroom behavior contribute to the rate of literacy development in preschool children? Minimal support was documented for this hypothesis. That is, a relationship between behavior and literacy developmen t was found for the Picture Naming subtest only. This research question wa s designed to explore the differences in early literacy development among presc hool children who had elevated scores on a teacher completed behavior rating scale as compared to their peers who did not have elevated scores. It was hypothes ized that children who had high ratings of behavior would have lower scores measuring early literacy skills than those children who had ratings indicating typi cal behaviors. A review of two HLM
123 analyses revealed that scores on the DE CA Behavior Scale had a significant relationship with the Picture Naming sl ope only. Picture Naming scores of students who had high DECA Behavior rati ngs increased over time but the Picture Naming scores of students w ho had lower DECA Behavior ratings remained relatively unchanged over time. This finding conflicts with previous research linking behavior to achievement (Al Otaiba & Fuchs, 2002; Nelson, Benner, & Gonzalez, 2003). In this st udy behavior was assessed at Time One only. Therefore it is not known if t he behavior of the children who initially had higher ratings on the DECA improved from time one to time four. It could be hypothesized that being enrolled in pre school provided these children with a structured environment that aided in curb ing their negative behaviors. If this were the case, then an increase in skill development over time might be expected. Conversely, it could be pos ited that behavior scores increased as a result of an improvement in the childre nÂ’s expressive language, a skill that is measured through Picture Naming. No other relationship was found bet ween literacy and behavior. Although significant relationships were not found for Alliteration and Rhyming using the HLM analysis, arithmetic differences we re found between the hi gh score and low score behavior groups based on descriptive statistics of their average scores on all three subtests across the four points in time. The children with the higher scores on the behavior scale consistently had lower literacy scores, thus revealing a trend that was predicted. This generates the question of whether or not the relationship would have been signif icant if the sample size was larger.
124 Further, only 11 of the 71 students for which behavior scores were obtained had scores above sixty (a score above 60 is considered a clinical sign of behavioral issues). Overall, these findings e licit additional questi ons regarding the contribution of positive and negative cla ssroom behaviors on the rate of literacy development. Research Question #2: What factors cont ribute to the rate and levels of literacy development for children identified with typical or challenging behaviors? The two factors explored in this questi on include child (e.g., gender, race, age) and classroom (e.g., teac her experience, class size) characteristics. Child characteristics comprised level two of t he HLM analyses. Results indicated that race and attendance have a significant re lationship with expressive language skills. Specifically, White students obtained higher scores and had a greater slope on the Picture Naming subtest of the IGDI as com pared to Non-White students. The support for race as a si gnificant predictor of early literacy development is consistent with research a ffirming that culturally and linguistically diverse students attain significantly lower performance levels on measures of reading achievement (Diamond and Onwu egbuzie, 2001; Meece and KurtzCostes, 2001). Attendance also was rela ted to the growth of expressive language in preschool children. As t he number of days absent increased, the score on the Picture Naming subtest dec reased. This finding aligns with previous research (Easton & Englehar d, 1982; Moonie et al., 2008; Gottfried, 2009) conducted with the school-aged popu lation. Although absenteeism and achievement in preschool has not been widely explored, the link appears to be
125 logical. To expand, academic engaged time is the amount of time a child is attending to the curriculum. Therefore, the level of learning is related to the amount of time the child spends actively engaged in the academic environment (Shapiro, E. S. & Heick, P., 2004). G illiam and Shahar (2006) explored the rates and predictors of preschool expulsion and suspension in Massachusetts. Results indicate that expulsion rates we re 13 times higher than the national K-12 rate. Expulsion is similar to absenteeism in that they are both examples of loss of academic engaged time. It is not surpri sing, therefore, that attendance and expressive language have a significant rela tionship during these early years. The sole child characteristic that was identified as a significant predictor of phonemic awareness, as measured by bot h Alliteration and Rhyming subtests, was Home SES. The relationship was positive, suggesting that the higher the median income for the neighborhood in wh ich the child resides, the higher the score on the two subtests. This is consis tent with research that explored the link between SES and academic achievement (Nichols, Rupley, Rickelman and Algozzine, 2004; Orr, 2003; St ipek, 2001). The age of the preschool population is a pivotal age in which this link bec omes more apparent (Sattler, 1990). Gender was identified as the second child characteristic that resulted in a significant predictor of phonemic awareness. This relationship was present for the Rhyming subtest only and indicated hi gher performance levels for girls as compared to boys. Previous rese arch (Diamond and Onwuegbuzie, 2001) supports this finding.
126 Despite the significant relationshi p for race, gender and attendance, with expressive language as measured thr ough the Picture Naming subtest, the majority of the predictor s (age, gender, race, attendance, and behavior) at level two yielded non-significant findings for t he Alliteration and Rh yming subtests. A potential explanation for this lack of significa nt relationship relates to the variation in skill requirements necessary for the ch ild to successfully complete the tasks of the three subtests. Similarl y, the difficulty level, vari es within the three subtests (Alliteration and Rhyming are more comp lex than Picture Naming), making equal comparisons impossible. Expressive l anguage begins at birth, with the newborn using sounds to indicate pain or pleasur e. This skill develops over time to include gestures, babbling, single words, and sentences, all with the intent of conveying wants/needs or to express m eaning to others. While expressive language skills continue to be refined as children get older, these skills are already present at entry into pre school. In contrast, phonemic awareness typically emerges during the course of the preschool years (Lonigan, Burgess, Anthony, & Barker, 1998; Whitehurst & Lonigan, 1998) and theref ore may not be present in some children in the beginning of their preschool experience. While the analysis did not reveal a significant finding, the mean Alliteration and Rhyming scores increased throughout the four assessment windows. Since the skill level required for each subtest differs, it elicits thought about the rate of skill development in phonemic awareness throughout the academic year, and whether progress-monitoring extending to the end of the year would tap into a significant finding. In other words, if phonemic awareness is a skill that is
127 acquired during the preschool years, then a significant relationship is more likely to be found if skill assessment is conduc ted throughout the year as opposed to the beginning months of the school year. Classroom characteristics were expl ored at the third level of the HLM models, where child factors were controll ed. Class size appears to play a role in the growth of expressive language skills in preschool children. Specifically, scores on the Picture Naming subtest increased as the number of children enrolled within a classroom decreased. Th is is supported by previous research examining the relationship between class size and achi evement (Blatchford et al., 2003; Nye, Hedges, and Konstantopoulos, 1999). Further results highlighted a marginal relationship (p=.056) between level of education attained by the teacher and Alliteration scores. This re lationship is supported by previous research (Darling-Hammond, 1999; t he National Center for Educational Research, 2000) that demonstrated the re lationship between levels of teacher education and student academic achievement. While these studies indicate a positive and significant relationship between the two variables, teacher certification as well as a major in the fi eld were more powerful predictors of reading achievement, even when student SES and language status were accounted for (Darling-Hammond, 1999). It is therefore, not surprising that the results from the current study revealed only a marginal relationship. Qualitative data regarding type of certific ation and major were not available for this archival research; however, the use of such data in future studies may be valuable.
128 There was no empirical support linking the remaining third level variables (e.g., site SES, teacher education, teacher experience, class size, class environment) to literacy development. The small sample size at level three (N=8 classrooms), may have contributed to the insigni ficant findings at this level. More specifically, it is hypothesized that sample size resulted in non-convergence during the initial attempt at running the analysis. With eight classrooms and the initial six predictor variabl es, the model was too comple x. Although two of the variables (Early Literacy Observation Checklist Â– ELOC1 and ELOC2) were collapsed into one allowing for convergenc e, the number of predictors at the second level essentially utilized the majo rity of the account ed variance, leaving little variance for the five predictors at le vel three. This resulted in the need to remove the level three error terms. It is believed that a larger sample size (i.e., more classrooms) would have prevented these issues from occurring. Question #3: What differences are there between literacy development in children with challenging behavior who hav e high scores measuring within-child protective factors in comparison to children with challengi ng behaviors who have low scores measuring within-child protective factors? An HLM analysis was constructed to address the link between behavior and within-child protective factors. The findings suggest that within-child protective factors had a pos itive and significant relationship with Alliteration scores for those children who did not hav e challenging behaviors. Interestingly, the Alliteration scores of children with c hallenging behaviors were not related to level of within-child protective factors. Overall, withinchild high protective factors
129 were associated with high scores measur ing phonemic awareness in comparison to low within-child protective factors, regardless of the presence of behavioral issues. The dialogue around this resear ch question generates thought about the potentially strong influence of behavior on ac hievement. That is, are challenging behaviors more powerful than within-child protective factors? Based on the current study, within-child protective fa ctors are influential to those children who do not have challenging behaviors. Combin ing the finding of the current study with previous research supporting t he link between behavior and achievement (Nelson, Benner, & Gonzalez, 2003; Al Otaiba & Fuchs, 2002; Torgesen, 2000) provides a strong foundation for prevent ion and early intervention at the preschool level. Summary Results from this study provided val uable information regarding the factors contributing to literacy development in pre school children. Previous research has focused on examining such factors only with school-aged children. In general, support was found for some variables (race, attendance, gender, home SES, class size, teacher education, classr oom environment, behavior, within-child protective factors) thus providing support that child and classroom factors are related to literacy development prior to the elementary school years. Implications for the Profession of School Psychology The findings from this study can benef it practitioners and researchers who collaborate with early childhood educators. First, it wa s documented that several factors (e.g., child, cla ssroom) influencing achievement in elementary and
130 secondary education also contribute towa rds the development of early literacy skills in children three to five years of age. Second, behavioral issues and the presence of within-child protective factors play a role in literacy development in the preschool setting. Therefore, support ex ists for providing early intervention in the preschool settings, with a focus on academic skill development as well as prosocial skills. The need for early intervention in this area is further supported by Carter et al (2010) w ho reported that approximatel y one in five children met the criteria for behavioral issues during the transition to formal schooling. Sociodemographic and psychosocial fa ctors such as persistent poverty beginning in early childhood, limited parental education, and low family expressiveness were explored and found to be significantly associated with mental health issues in the preschool pop ulation. Therefore, screening and early intervention by practitioners in the field during the preschool years is warranted to increase the chances of academic and soci al-emotional success in the transition to formal schooling. Practitioners can help support early childhood educators in creating classroom environments that are litera cy-rich and promote prosocial behaviors. Such support should include screening pres chool children for early identification of problematic behaviors and/or deficits in literacy grow th, focusing on academic engaged time and increasing language expo sure. Additionally, including the family in the efforts to increase the skill level of the ch ildren should not be ignored. According to a longitudinal study conducted by Hart and Risley (1995), childrenÂ’s vocabulary size at age three were high correlated to language scores
131 in subsequent years. No tably, the size of the childÂ’s vocabulary varied significantly between low and high income families, thus providing support for parent training as another facet in a m odel of design for practitioners. The use of the Preschool IGDI as an assessment tool to gauge early literacy skills and monitor progress has been supported in the current study. The use of this assessment tool provides practitioners with data to monitor the progress of skill development in early literac y, which can, in turn, assist in the identification of children who require additional re sources/instruction in expressive language and phonemic awareness. Responses to intervention also can be monitored with this assessment t ool, providing practitioners and early childhood educators with dat a to work towards the goal of kindergarten readiness. In summary, it has been noted through this and previous research that children acquire early literacy skills during the preschool y ears. Therefore, practitioners and educators are at a pivotal poi nt to impact the trajectory of these children and provide them with t he academic and behavioral competencies needed to succeed in school. Additionally, the use of Hierarchica l Linear Modeling as the statistical analysis to explore growth, as well as hi gh rate of change in growth, proved to be valuable. School psychologists serving a scientist-practitioner role can benefit from utilizing this method as a means to ascertain the relationship of nested variables. Given that the school setting is nested by nature (children within a classroom, classrooms within a school, etc) HLM enables school psychologists to
132 better explore rates of learning, which wi ll contribute towards the development of interventions and subsequent monitoring of t he childÂ’s response to intervention.. Finally, the results of the study generate discussion pertaining to policy development as it relates to the qua lity of teachers and classrooms at the preschool level. According to Barnett (2004), the educational qualifications of preschool teachers are related to ear ly learning and development; however, there are no consistent qualifications for teachers prior to the kindergarten level. Barnett (2004) reported that fewer than half of the pr eschool teachers held a bachelorÂ’s degree, with many teachers r eporting high school as their highest level of education. The results of this study show that as teacher education and the richness of the liter acy environment increase so do scores measuring phonemic awareness and expressive langua ge respectively. It therefore strengthens the notion of requiring presch ool teachers to have a college degree with specialized training in early child hood education. Periodic training and professional development for teachers in the preschool setting also should be considered as policy to ensure current cert ification as well as dissemination of Best Practice for teaching in the preschool classroom. Limitations The current study contributed both t heoretically and practically to the existing research surrounding liter acy development and behavior. Notwithstanding, there are seve ral limitations to this stud y. First, teachers were selected based on a convenience sample. This prevented a random selection of study participants. Random selection allo ws for an equal chance of participation,
133 thus resulting in a distribution comparable to that of the population from which the sample is drawn. A typical shortcomi ng in research, convenience sampling often leads to the question of whether the char acteristics of the teachers would differ under an alternative sele ction process. Second, the use of single measure such as a behavior rating scale to identify behavior and within-child protecti ve factors hinders t he accuracy of the interpretation of those data. Ideally, data are best derived from multi-source (records review, interview, observations testing), multi-informant (teacher, parent, child) conditions collected across mult iple settings and points in time. In the current study, data on t hese variables were collected at one point in time, namely the beginning of the ac ademic year. This generates questions including: (1) did the teacher have enough time to fo rmulate an accurate picture of the childÂ’s behavior prior to completing the scale and (2) was the behavior maintained at the original level throughout the school ye ar? That is, did any of the children who had high scores on the behavior scale improve over time or did any of the children who had low scores wo rsen over time? As cited by Gilliam and Shahar (2006), approximately 8% of a ll preschool children exhibit behavioral problems that are diagnosable, which are associated with future behavior issues, poor peer relationships, and decreased achiev ement in kindergarten. Given this statistic, it is questionable as to w hether the current st udy under-identified behavioral issues in the student sample, further warranting additional research. The data collected to measure protective factors in the preschool sample was limited to within-child factors (attachme nt, initiative, self-c ontrol) and did not
134 account for external factors including the home or community. The quality of the home environment is a powerful predictor of the outcome for children (Benard, 1991) and includes factors such as ca ring, support, and parental warmth. Further, researchers have posited that care giving is the most powerful predictor of resiliency in children that lasts through childhood and adolescence (Demos, 1989; Werner & Smith, 1982; Rutter, 1979) The absence of data examining external protective factors can, therefore, be consider ed a shortcoming to this study. Third, the socioeconomic status fo r the individual child was based on household zip code due to lack of family income data. Although the use of zip codes to determine socioeconomic status is supported for use when specific information is not available (Krieger Williams & Moss, 1997; Krieger, 1992), family-specific data would result in a greater confidence in understanding the relationship between SES and early literacy. The duration of the data collection phase is a fourth limitation. Although four data points were included in t he study, the duration of data collection consisted of three months. Long-term pr ogress monitoring extending to the end of the academic school year would prov ide valuable information to address the research questions interesting this study. Finally, the number of participating schools in the study was small, affecting the analyses at the third level of both models. Although steps (setting the third level variables as a fixed effe ct, dropping the level-2 errors from the model and collapsing two similar variables in to one) were taken to address this
135 issue, an increased sample size is re commended to ensure accurate parameter estimates. Based on these limitations, the results should be interpreted with caution. Although significant findings were found linking child and classroom factors to early literacy development in preschool, additional research is warranted and encouraged. Future Research Despite the insightful results gleaned from the current study, additional questions have been generated, paving the way for future research in this area. First, and critical, is the following: would the extension of progress monitoring to the end of the school year yield additi onal significant relationships? To elaborate, the methodology of the current study excluded t he last five months of the school year, preventing a comprehensiv e assessment of skill development in literacy for the preschool children. As di scussed in a previous section, the skills assessed through the Alliteration and Rhym ing subtests typically emerge over the course of the preschool year. Therefore, it make s sense to monitor progress for the length of the school year as opposed to limiting data collection to the first four months, when phonemic awareness is just beginning to emerge for many students. Second, is behavior more influential than within-child protective factors? Results from the current study indicate that within-child protective factors are advantageous for literacy development in those children who demonstrate typical behaviors. However, withinchild protective factor s did not appear to have a
136 significant relationship with the childr en who had challenging behaviors based on the behavior rating scale. To explore this finding in depth, it is suggested that future research include a larger sample of children with challenging behaviors. In addition, it is suggested that future res earch utilize a more accurate method for obtaining data on behavior and protective fa ctors. Methodologi cal changes also are suggested for future exploration in th is area. Such changes include overall sample size and variability. More specifically, obtai ning a larger and randomly selected sample may have provided access to schools and teachers with greater variability in both child and teacher characte ristics. As described in chapter three, the goal of the study was to a ccess teachers who had limited experience in the classroom and with early lit eracy training in an effort to promote skill building and to provide resources. Therefore, the sample in the current study was restricted to teachers who were identified as needing skills and resources to aid in the literacy development of their preschool student s from schools located in low SES areas of the county. The current area of res earch would benefit from the expansion of the sample to include high SES schools as well as teachers with higher qualifications (i.e., year s of experience, years of education, certification) and skills in an effort to explor e the differences in statistical results.
137 REFERENCES Achilles, C. M. (2005). Class size and learni ng. In L. W. Hughes (Ed.). Current issues in school leadership. Topics in educational leadership. Mahwah, NJ: Lawrence Erlbaum Associates. Al Otaiba, S., & Fuchs, D. (2002). Characteristics of children who are unresponsive to early literacy interventi on: A review of the literature. Remedial and Special Education, 23 (5), 300-316. Alexander, K. L., Entwisle, D. R., and Olson, L. S. ( 2001). Schools, achievement, and inequality: A seasonal perspec tive. Educational Evaluation and Policy Analysis, 23 171-91. American Psychiatric Association. (2000) Diagnostic and stat istical manual of mental disorders Â– Text Revision(4th ed.) Washington, DC: American Psychiatric Association. Appleyard, K., Egeland, B. Van Dulmen, M. Sroufe, L.A. (2005). When more is not better: the role of cumulative risk in child behavior outcomes. Journal of Child Psychology and Psychiatry, 46 (3), 235-45. Armstrong, K. H., Dedrick, R. F., & Greenbaum, P. E. (2003) Factors associated with community adjustment of yo ung adults with serious emotional disturbance: A longitudi nal analysis. Journal of Emotional and Behavioral Disorders, 11 (2), 66-76. Arnold, D. (1997). Co-occurrence of externalizing behavior problems and
138 emergent academic difficulties in y oung high-risk boys: A preliminary evaluation of patterns and mechan isms. Journal of Applied Developmental Psychology, 18 317-330. Arnold, D. H., Ortiz, C., Curry J. C., Stowe, R. M., Goldst ein, N. E., Fisher, P. H., Zeljo, A., Yershova, K. (1999) Promoting academic success and preventing disruptive behavior disor ders through community partnership. Journal of Community Psychology, 27 (5), 589-598. Atzaba-Poria, N. Pike, A., and Deater-De ckard, K. (2004). Do risk factors for problem behaviour act in a cumulative manner? An examination of ethnic minority and majority children through an ecological perspective. Journal of Child Psychology and Psychiatry, 45 (4), 707-718. Axcher, C. and Fruchter, N. (2001). Eacher quality and student performance in New York CityÂ’s low performing schools. Journal of Education for Students Placed at Risk, 6 (3), 199-214. Baker, J. A., Derrer, R.D ., Davis, S.M., Dinklage-Travis H.E., Linder, D. S., and Nicholson, M.D. (2001). The flip side of the coin: Understanding the school's contribution to dropout and completion. School Psychology Quarterly 16 (4), 406-426. Barnett, S. W. (2004). Bette r teachers, better preschools: Student achievement linked to teacher qualifications Preschool Policy Matters, 2 New Brunswick, NJ: NIEER. Barrera, M., Biglan, A., Taylor T. K., Gunn, B. K., Smolko wski, K., Black, C., Ary,
139 D. V., and Fowler, R. C. (2002). Early elementary school intervention to reduce conduct problems: A r andomized trial with Hispanic and nonHispanic children. Prevention Science, 3 (2), 83-94. Baydar, N., Brooks-Gunn, J. and Fustenberg, F. (1993). Early warning signs of functional illiteracy: Predictors in childhood and adolescence. Child Development, 64 815-829. Benard, B. (1991). Fostering resiliency in ki ds: Protective factors in the family, school and community. Portland, OR: Northwest Regional Educational Laboratory. Berk, L. E. (2000). Child Development (5th Edition). Boston, MA: Allyn and Bacon Blatchford, P. Bssett P., Goldst ein, H., and Martin, C. (2003). Are class size differences related to pupilsÂ’ educational progress and classroom processes? Findings from the institute of education class size study of children aged 5-7 years. British Educati onal Research Journal.. 29 (4), 709-730. Bronfenbrenner, U. (1979). The ecology of human development. Cambridge, MA: Harvard University Press. Br ooks-Gunn, J. and Markman, L. B. (2005). The contribution of parenting to ethnic and racial gaps in school readiness. Future of Children, 15 (1), 139-168. Buckner, J. C., Bassuk, E. L., and Weinreb, L. F. ( 2001). Journal of School Psychology. Predictors of academic achievement among homeless and low-income housed children, 39 (1), 45-69. Burkam, D. T., Ready, D. D., Lee, V. E., and LoGerfo, L. F. (2004) Social-class
140 differences in summer learning betw een kindergarten and first grade: Model specification and estimation. Sociology of Education 77 (1), 1-31. Bryk, A. S. and Raudenbush, S. W. (1997). Application of hierarchical linear models to assessing change. Psychological Bulletin, 101 (1), 147-158. Caprara, G. V., Barbaraneli, C., Pastorelli, C., Bandura, A., and Zimbardo, P. G. (2000). Prosocial foundations of ch ildrenÂ’s academic achievement. Psychological Science, 11 (4), 302-306. Carter A. S., Wagmiller, R. J, Gray, S. A., McCarethy, K.J., Horwitz, S. M., Briggs-Gowan, M. J. (2010). Prev alence of DSM-IV disorder in a representative, healthy birth cohor t and school entry: Sociodemographic risks and social adaptation. Journal of the American Academy of Child and Adolescent Psychiatry, 49 686-698. Child Trends and Center for Child Heal th Research (2004). Early childhood development in social context: A c hartbook. The Commonwealth Fund: New York. Christenson, S. L., Rounds, T., and Frank lin, M. J. (1992). Home-school collaboration: Effects, issues, and opportunities. In S. L. Christenson & J.C. Conoley. (Eds.), Home-School Co llaboration: Enhancing ChildrenÂ’s Academic and Social Com petence, The National Association of School Psychologists: Maryland. Conroy, M. A. and Brown, W. H. (2004). Early Identificati on, prevention, and
141 early intervention with young children at risk for emotional or behavioral disorders: Issues, trends, and a call for action. Behavioral Disorders, 29 (3), 224-236. Cooper H. M. (1989). Does reducing student to instru ctor rations affect achievement? Educational Psychologist, 24 (1), 79-98. Cusumano, D. L., Armstrong, K., Cohen, R., & Todd, M. (2006). Indirect impact: How early childhood educ ator training and coaching impacted the acquisition of literacy skills in pre school children. Journal of Early Childhood Teacher Education, 27 363-377. Darling-Hammond, L. (2000). Teacher quality and student achievement: A review of state policy evidence. Education policy analysis archives, 8 (1). Diamond, P. J. and Onwuegbuzie A. J. (2001). Factors associated with reading achievement and attitudes among elementary school-aged students. Research in the Schools, 8 (1), 1-11. Dodge, K. A., Pettit, G.G., and Bates, J. E. (1994). Socializati on mediators of the relation between socioeconomic status and child conduct problems. Child Development, 65 Easton, J. Q. and Engelhard, G. (1982). A longitudinal re cord of Elementary School absence and its relationship to reading achievement. The Journal of Educational Research, 75 (5), 269-274. Entwisle, D. R., Alexander, K. L. and Olsen, L. (1997). Children, Schools, and Inequality. Boulder, CO: Westview Press. Finn, J.D., Gerber, S. B. and Boyd-Zaharias, J. (2005). Small classes in the
142 early grades, academic achievement and graduating from high school. Journal of Educational Psychology, 97 (2), 214-223. Flanagan, K. S., Bierman, K. L. and Kam, C. M. ( 2003). Identifying at-risk children at school entry: The usef ulness of multibehavioral problem profiles. Journal of Clinical Ch ild and Adolescent Psychology, 32 (3), 396-407. Frick, P. J., Kamphaus, R. W., Lahey, B. B., Loeber, R., Chri st, M. G., Hart, E. L., and Tannenbaum, L. E. ( 1991). Academic underachievement and the disruptive behavior disorders. Jo urnal of Consulting and Clinical Psychology, 59 (2), 289-294. Garmezy, N. (1994). Reflections and commentary on risk, resilience and development. In R.J. Haggerty, L. Sherrod, N. Gamezy, & M. Rutter (Eds). Stress, risk, and resilience in children and adolescent s: processes, mechanisms, and interventions (pp. 1-19). New York: Cambridge University Press. Gauthier, Y. (2003). Infant mental health as we enter the third millennium: Can we prevent aggression? Infant Mental Health Journal, 24 (3), 296-308. Gilliam, W. S. & Shahar, G. (2006). Preschool and child care expulsion and suspension: Rates and predictor s in one state. Infants and Young Children, 19 (3), 228-245. Green, M. T., Clopton, J. R., Pope, A. W. (1996) Understanding gender differences in referral of children to mental health services. Journal of Emotional and Behavioral Disorders, 4 (3), 182-190.
143 Gregory, E., Williams, A., Baker, D., & Stre et, B. (2004). Intr oducing literacy to four year olds: Creating classroom cu ltures in three schools. Journal of Early Childhood Literacy, 4, (1), 85-107. Gottfried, M. A., (2009). Evaluating the relationship between student attendance and achievement in urban elementary and middle schools: An instrumental variables approach. Educational Evaluation and Policy Analysis 31 (4), 392-415. Hart B., & Risley, R. T. (1995). Meaningful differ ences in the everyday experience of young American children. Baltimore, Paul H. Brookes. Hindshaw, S. (1992). Ac ademic underachievement, a ttention deficits, and aggression: Comorbidity and implicati ons for intervention. Journal of Consulting and Clinical Psychology, 60 (6), 893-903. Hindshaw, S. (1992). Externalizi ng behavior problems and academic underachievement in childhood and adoles cence: Causal relationships and underlying mechanisms. Ps ychological Bulletin, 111 127-155. Hofmann, D. (1997). An over view of the logic and rationale of hierarchical linear models. Journal of Management, 23 (6), 723-744. Howse, R. B., Calkins, S. D., Anastopoulos A. D., Keane, S. P., Shelton, T L. (2003). Regulatory contributors to childrenÂ’s kindergarten achievement. Early Education and Development, 14 (1), 101-119. Huaqing Qi, C. (2003). B ehavior problems of preschool children from lowincome families: review of the lit erature. Topics in Early Childhood Special Education, Winter
144 Jimerson, S. R. (2001). Meta-analysis of grade retention research: Implications for practice in the 21st century. School Psychology Review, 30 420-437. Johnson, L. J., Gallagher, R. J., Cook, M., and Wong, P. ( 1995). Critical skills for kindergarten: Perceptions from kinder garten teachers. Journal of Early Intervention, 19 (4), 315-349. Kalff, A. C., Kroes, M., Vles, J.S.H., Hendri ksen, J.G.M., Feron, F.J.M., Steyaert, J., van Zeben, T.M.C.B., Jolles, J., and van Os, J. (2001). Neighbourhood level and individual leve l SES effects on child problem behaviour: A multilevel analysis. Journal of Epidemiology and Community Health, 55, (4), 246-250. Kamps, D. M. (2003). Curriculum influences on growth in early reading fluency for students with academic and behavioral risks: A descriptive study. Journal of Emotional and Behavioral Disorders, Winter Landry, S. (2002). Early childhood educat ion: From efficacy research to improved practice. A summit on early childhood cognitive development (April 30). Found on http://www.ed.gob.teacher/how/earl/ cognitivedevsummit02/page_pg3.html Lane, K. L., Wehby, J. H., Menzies, H. M., Gregg, R. M., Doukas, G. L., and Munton, S. M. (2002) Early Literacy Instruction for first grade students at Â–risk for antisocial b ehavior. Education and Tr eatment of Children, 25 (4), 438-458. Lonigan, D.J., Burgess, S. R., Anthouny, J. L., & Ba rker, T. A. (1998).
145 Development of phonological sensitivity in 2 to 5 year old children. Journal of Educational Psychology, 90 294-311. MacMillan, D. L., Gresham, F. M., Lopez M. F., Bocian, K. M. (1996). Comparison of students nominated fo r prereferral interventions by ethnicity and gender. Journal of Special Education, 30 (2), 133-151. Magnuson, K. A., and Waldfogel, J. (2005). Early childhood care and education: Effects on ethnic and racial gaps in school readiness. Future of Children, 15 (1), 169-196. McLeod, J. D., and Nonnemaker, J. M. (2000). Poverty and child emotional and behavioral problems: Racial/Ethnic diffe rences in processes and effects. Journal of Health and Social Behavior, 41 (2), 137-161. Meece, J. L., & Kurtz-Cost es, B. (2001). Introduction: The schooling of ethnic minority children and youth. Educational Psychologist. 36 (1), 1-7. Moonie, Sheniz, Sterling, D. A., Figgs, L. W., and Ca stro, M. (2008). The relationship between school absence, academic performance, and asthma status. Journal of School Health, 78 (3), 140-148. National Center for Education Statistics (1998a). The condition of education 1998 (NCES 98-013). Washington, DC: U.s. De partment of Education, Office of Educational Resear ch and Improvement. National Center for Education Statisti cs (1998b). Statistical analysis report: AmericaÂ’s teachers: profile of a profession, 1993-94 (NCES 97-460). Washington, DC: U.S. Department of Education. National Center for Education Statistics ( 1998c). Statistical analysis report: Urban
146 schools: The challenge of loca tion and poverty (NCES 96-184). Washington, DC: U.S. Department of Education. National Head Start Association. T he Early Years of HeadsUp! Reading: A Report from the National Head Start Association. Alexandria, VA: National Head Start Association. Retrieved 1/11/04 from www.huronline.org National Institute of Child Health and Human Developm ent (NICHD) Early Child Care Research Network (2003). Does am ount of time spent in child care predict socioemotional adjustment duri ng the transition to kindergarten? Child Development, 74 (4), 976-1005. National Institute of Child Health and Human Developm ent (NICHD) Early Child Care Research Network (2004). Does cl ass size in first grade relate to childrenÂ’s academic and social per formance or observed classroom processes, 40 (5), 651-664. National Research Council (2001). Understanding dr opouts: Statistics, strategies, and high stakes testing. Washington, DC: National Academy Press. Nelson, J. R., Babyak, A, Gonzalez, J., & Benner, G. J. (2003). An investigation of the types of problem behaviors exhibited by k-12 students with emotional or behavioral disorders in public school settings. Behavioral Disorders, 28 (4), 348-359. Nelson, J. R., Benner, G. J. Gonzalez, J. (2003). Learner characteristics that
147 influence the treatment effectiveness of early literacy interventions: A meta-analytic review. Learning Disabi lities Research and Practice, 18 (4), 255-267. Nye, B., Hedges, L V., & Konstantopoulos, S. (1999). The long term effects of small classes: A five-year follow-up of the Tennessee class size experiment. Educational Ev aluation and Policy Analysis, 21 (2), 121-142. Orr, A. (2003). Black-white differences in achievement: The importance of wealth. Sociology of Education, 76, (4), 281-304. Osborne, J. Wl (2000). Advantages of hi erarchical linear modeling. Practical Assessment, Research and Evaluation 7 (1). Retrieved August 14, 2005 from http://PAREonline. net/getvn.asp?v=7&n=1. Patterson, G. R., Reid, J. B., and Dishion, T. J. (1992). Anti social Boys. Eugene, OR: Castalia. Pikulski, J. J., and Tobin, A. W. (1989) Factors associated with long-term reading achievement of early readers. In S. McCormick, J. Zutell, P. Scharer, and P. O.Keefe (Eds). Cogn itive and social perspective for literacy research and instruction. Ch icago: National Reading Conference. Priest, J., Davis, K., McC onnell, S., McEvoy, M., & Shin J. (1999, December). Individual Growth and Development Indicators of preschoolersÂ’ Â“expressing meaningÂ” skills: Follow t hat trajectory! Presentation at the annual meeting of the Division for Earl y Childhood, Council for Exceptional Children, Washington, D.C. Ramey, C. T. & Ramey, S. L. (2004) Early learning and school readiness:
148 Can early intervention make a differ ence? Merrill-Palmer Quarterly, 50 (4), 471491. Ramey, C. T., & Ramey, S. L. (2002). Early childhood e ducation: From efficacy research to improved practice. A summit on early childhood cognitive development (April 30). Found on http://www.ed.gob.teacher/how/earl/ cognitivedevsummit02/page_pg2.html Ready, D. D., Lee, V. E., & Welner, K. G. (2004). Educ ational equity and school structure: school size, overcrow ding, and schools-within schools. Teachers College Record 106 (10), 1989-2014. Rimm-Kaufman, S.E., Planta, R. C., and Cox, M. J. (2000). TeachersÂ’ judgments of problems in transition to kindergar ten. Early Childhood Research Quarterly, 15 147-166. Rivera-Batiz, F. L., & Marti, L. (1995). A school system at risk: A study of the consequences of overcrowding in New York City Public Schools. New York: New York City CitizensÂ’ Commission on Planning for Enrollment Growth. Rutter, M. (1979). Protective factors in childrenÂ’s responses to stress and disadvantage. In M.W., Kent and J.E. Rolf (Eds.) Primary prevention of psychopathology, 3, Hanover, NH: Un iversity of New England Press. Sameroff, A.J., Bartko, W.T., Baldwin, A., Baldwin, C., and Seifer, R. (1998). Family and social influences on the development of child competence. In M. Lewis and C. Feiring (Eds.),F amilies, Risk, and Competence. Mahwah, N: Erlbaum.
149 Sattler, J. M. (1988). Assessment of Children (3rd Edition). San Diego. Scarborough. H. S. (1989). Prediction of reading dysfunction from familial and individual differences. Journal of Educational Psychology, 81 101-108. Schultz, D, and Shaw, D. S. (2003). Bo ysÂ’ maladaptive social information processing, family emotional clim ate, and pathways to early conduct problems. Social Development, 12 (3). 440-460. Stipek, D. J. (2001). Pathways to constr uctive lives: The importance of early school success. In Bohart, A. C. and St ipek, D. J. (Eds), Constructive and Destructive Behavior: Implications for Family, School, and Society. American Psychological Association: Washington, DC. Stipek, D., and Miles, S. ( 2008). Effects of aggression on achievement: Does conflict with the teacher make it worse? Child Development, 79, 17211735. Torgesen, J. (2000). Individual differences in response to early interventions in reading: The lingering pr oblem of treatment re sisters. Learning Disabilities Research and practice, 15 (1), 55-64. Torgesen, J., Wagner, R. K., Rashotte, C. A., Lindamood, P., Rose, E., Conway, T., & Garvan, C. (1999). Preventi ng reading failure in young children with phonological processing disabilities: Group and individual responses to instruction. Journal of Educational psychology, 91 579-593 United States Commission on Civil Rights (1967) Racial isolation in the public schools. Washington, DC: US Government Printing Office. Unitied States Department of Education (2008). Promoting educational
150 excellence for all American families: Ea rly Reading First. Retrieved 5/5/08 from http://www.ed.gov/program s/earlyreading/index.html. Werner, E. E. (1992). T he children of Kauai: resiliency and recovery in adolescence and adulthood. Journal of Adolescent Health, 13 (4), 262-68. Werner, E., & Smith, R. (1982) Vulnerable but invincible : A longitudinal study of resilient children and youth. New York: Adams, Bannister and Cox. Whitehurst, G.J., & Lonigan, C. J. (1998) Child development and emergent literacy. Child Development, 69 848-872. Zigler, E., and Styfco, S. J. (2001). Can early childhood intervention prevent delinquency? A real possibility. In B ohart, A. C. and Stipek, D. J. (Eds), Constructive and Destructive Behavior: Implications for family, school, and society. American Psychological Association: Washington: DC.
152 Appendix A: Application and Agreement Form for ELO Teacher Participation HEADS UP! READING PLUS LITERACY PROJECT SCHOLARSHIP APPLICATION Applicant Name: Day Phone: Evening Phone: Highest Level of Education: (Check one) H.S Diploma G.E.D. Some College 2 Yr. College Degree 4 Yr. College Degree Advanced Degree Site Employer Name: Work Address: City: State: Zip: Center Director (if applicable): Type of Work Site: (Check one) Family Child Care Child Care Center Private Pre-K Privat e Kindergarten Pre-K ESE Head Start Public Kindergarten Home Visitor Program Number of years you have worked in Early Childhood: Age of Children you are currently working with: (Check all that apply) 0-1 1-2 2-3 3-4 4-5 5-6 Number of Children currently in your care: Number of Children in your care whose first language is not English: Please list any previous training in Early Childhood Literacy: 1) 2) 3) Preferred Campus if selected: (Check one) Seminole St. Pete/Gibbs No Preference I understand that: 1) If eligible, I will receive more information about the requirements of participation for me and my Director (if applic able); 2) If employed at a Child Care Center, my Director must support my participation in this project. 3) If selected, there is no charge that I must attend all 15 classes and these classes are for college credit X X Applicant Signature Director Signature (if applicable)
153 Appendix A (Continued): Applicat ion and Agreement Form for ELO Teacher Participation Training Participant Contract I agree to participate in the Pinellas Early Literacy Community Project Training and Coaching Program, and will fulfill the following obligations: 1. Obtain the support and commitment from my Center Director to participate in the program. 2. I will attend the Orientation Session and all 14 satellite training session. (Will be allowed to miss one session to allow for illness or family obligations.) Should I miss a session, I will view the videotape of the session. 3. I will implement the literacy idea, activities and strategies learned in eh training/coaching program in my classroom. After each session, I will develop a brief action plan detailing how I will implement the strategy discussed, and return to the next training session with the plan. 4. I agree to share the specific printed literacy activities provided at each training session with my Director and at least one other teacher. I will assist my fellow teacher in developing an action plan, and bring to the next training session. 5. I will distribute books and materials to the families of children in my classrooms. 6. I will hold at least one Â“literacy eventÂ” for families of children in my classroom. 7. I agree to work with the Literacy Coaches in my classroom, and participate in six coaching visits. 8. I agree to participate in the evaluation, by completing surveys, encouraging parents to complete their surveys and assisting the Evaluator in connecting with families for literacy surveys. 9. I agree to participate in the Literacy Learning Community Showcase, and to bring a display of activities, photographs and other visual materials of how they implemented literacy activities in their classrooms. ________________________ ____________ ________________ Signature of Applicant Date ________________________ ____________ ________________ Signature of Director Date
154 Appendix B. Demographic Information Sheet Center: Address: Center Director: ____________________________________ Phone Number: _______________________________________________ ChildÂ’s Name: DOB Age Gender Race Home Zip Code Primary Language 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.
155 Appendix C: Attendance Information Sheet For each child, please record the number of days he/she was absent within the specified month. This form can be retu rned by using the se lf-addressed stamped envelope provided. Thank you again for your time and dedication to this project! ChildÂ’s Name Sept. Oct. Nov. Dec. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. Thank You!
156 Appendix D: Teacher Consent Form ________________________ __________________ ______________________Adult Informed Consent for Child Care Providers Social and Behavioral Sciences University of South Florida Information for People Who Take Part in Research Studies The following information is being presented to help you decide whether or not you want to take part in a minimal risk research study. Please read this carefully. If you do not understand anything, ask the person in charge of the study. Title of Study: Evaluation of Pinellas Early LIteracy Learning Community Project: Early Learning Opportunities (LCP: ELO)] Principal Investigator: Kathleen Hague-Armstrong. You are being asked to participate in the evaluation of LCP: ELO because you have applied to participate in the Â“Language Development In Young ChildrenÂ” course at St. Petersburg College. General Information about this evaluation: This evaluation intends to document the implementation and impact of the LCP: ELO. The LCP: ELO is a unique comprehensive approach towards improving literacy, reading readiness, and social-emotional functioning of children ages 0-5. The project will be conducted in Pinellas County, Florida, and will provide opportunities for ca regivers and teachers from publicly funded and private children's programs to increase their level of professional education, earn college credits, gain early literacy teaching skills, tools and materials for their classrooms, and promote healthy social-emotional development in the children they serve. In addition, parent educators with expertise in early childhood mental heatlh will provide support to families to enhance the young child's social and behavioral development. The evaluation goals include : (1) determine if LCP activities and objectives are implemented in a timely fashion; (2) determine if the home visiting component enhances family confidence and competence; (3) determine if the home visiting component enhances child social and emotional functioning; (4) determine if the classroom-teaching component increases knowledge and skills in child care providers; (5) determine if the mentoring and coaching of child care providers improve their confidence and competence in implementing early literacy strategies; (6) determine if children participating in LCP activities show improv ement in the of language and literacy skills; (7) determine if children transitioning to kindergarten demonstrate kindergarten readiness skills; (8) determine if it is feasible to implement this collaborative model within the community; (9) and determine the cost of implementing this model. Where the study will be done: Pinellas County early childhood centers, St. Petersburg College, Directions for Mental Health, Inc., and Florida Mental Health Institute at the University of South Florida.
157 Appendix D (continued): Teacher Consent Form Plan of Study : The evaluation will be conducted within the natural context of your classroom and childcare center. If you consent to participate, you may be asked to participate in individual interviews and/or an audiotaped one-hour focus group, and to complete rating scales and simple data collection forms. We will want to collect your information throughout the semester you are taking the Â“Language Development in Young ChildrenÂ” course in addition to the semester before (for those on the waiting list) and one-two semesters after the completion of the course. An evaluator will meet with you three times per semester for visits up to one hour and one half. These visits may be conducted during your regular meeting times with Â“Language Development in Young ChildrenÂ” or during your working hours. Payment for Participation : There will be no additional payment for participation in the evaluation. Benefits of Being a Part of this Research Study : By taking part in this evaluation, you will provide valuable information about the implementation and outcomes of the LCP: ELO project. This information will be used to modify and improve the current project. Risks of Being a Part of this Research Study : There are no known risks to participating in this evaluation. Confidentiality of Your Records : Your privacy and research records will be kept confidential to the extent of the law. Authorized research personnel, employees of the Department of Health and Human Services, and the USF Institutional Review Board may inspect the records from this research project. The results of this evaluation may be published. However, the data obtained will be combined with data from other childcare providers in the publication. The published results will not include your name or any other information that would personally identify you in any way. A pseudonym will be used in place of your name on all documents related to the evaluation and all data will be stored in locked files. Data stored within data bases will be entered with the pseudonym and will be only accessible to the research team through the use of a password. How many other people will take part ? About 50 Â– 150 children care providers, 1500 children, and 50 families. Volunteering to Be Part of this Research Study : Your decision to participate in this evaluation is completely voluntary. You are free to participate or to withdraw at any time. There will be no penalty or loss of benefits you are entitled to receive if you stop taking part in the evaluation. Questions and Contacts If you have any questions about this evaluation, please contact Kathleen Armstrong, Ph.D. at (813) 974-8530.
158 Appendix D (Continued): Teacher Consent Form If you have questions about your rights as a person who is taking part in an evaluation, you may contact the Division of Research Compliance of the University of South Florida at (813) 974-5638. Consent to Take Part in This Research Study By signing this form I agree that: I have fully read or have had read and explained to me this informed consent form describing this research project. I have had the opportunity to question one of the persons in charge of this research and have received satisfactory answers. I understand that I am being asked to participate in research. I understand the risks and benefits, and I freely give my consent to participate in the research project outlined in this form, under the conditions indicated in it. I have been given a signed copy of this informed consent form, which is mine to keep. _________________________ ________________________ _______________ Signature of Participant Printed Name of Participant Date Investigator Statement I have carefully explained to the subject the nature of the above evaluation. I hereby certify that to the best of my knowledge the subject signing this consent form understands the nature, demands, risks, and benefits involved in participating in this evaluation. _________________________ _________________________ __________ Signature of Investigator Printed Name of Investigator Date or authorized research investigator designated by the Principal Investigator Investigator Statement: I certify that participants have been provided with an informed consent form that has been approved by the University of South Flor idaÂ’s Institutional Review Board and that explains the nature, demands, risks, and benef its involved in participating in this evaluation. I further certify that a phone number has been provided in the event of additional questions. _________________________ ____________________ _______________ Signature of Investigator Printed Name of Investigator Date
159 Appendix E. Parental Assent Form Child Informed Assent Social and Behavioral Sciences University of South Florida Information for People Who Take Part in Research Studies The following information is being presented to help you decide whether or not you want your child to take part in a minimal risk research study. Please read this carefully. If you do not understand anything, please contact the person in charge of the study. Title of Study: Pinellas Early LIteracy Learning Community Project: Early Learning Opportunities (LCP: ELO)] Principal Investigator: Kathleen Hague Armstrong. Your child is being asked to participate because he/she is in a classroom whose teacher is attending the Â“Language Development In Young ChildrenÂ” course at St. Petersburg College. General Information about the Research Study : This is an evaluation of the Pinellas Early Literacy Learning Community Project, which assesses the implementation of the Â“Language Development In Young ChildrenÂ” course activities and outcomes related to literacy development in children. The LCP: ELO is a unique comprehensive approach to improving literacy, reading readiness, and social-emotional functioning of children ages 0-5. The project will be conducted in Pinellas County, Florida, and will provide opportunities for caregivers and teachers from publicly funded and private children's programs to increase their level of professional education, earn college credits, gain early literacy teaching skills, tools and materials for their classrooms, and promote healthy social-emotional development in the children they serve. Parent educators with expertise in early childhood mental health are also available to support families and provide home-based training to enhance t he young child's social and behavioral development. The evaluation goals include: (1) determine if LCP activities and objectives are implemented in a timely fashion; (2) determine if the home visiting component enhances family confidence and competenece; (3) determine if the home visiting component enhances child social and emotional functioning; (4) determine if the classroom-teaching component increases knowledge and skills in child care providers; (5) determine if the mentoring and coaching of child care providers improve their confidence and competence; (6) determine if children partici pating in LCP activities show improvement in the of language and literacy skills; (7 ) determine if children transitioning to kindergarten demonstrate readiness; (8) determine if it is feasible to implement this collaborative model within the community; (9) and determine the cost of implementing this model. Where the study will be done: This is a collaboration of Pinellas County early childhood centers, St. Petersburg College, Directions for Mental Health, Inc., and Florida Mental Health Institute at the University of South Florida.
160 Appendix E (Continued): Parental Assent Form Plan of Study: The study will be conducted within the natural context of the classroom and childcare center. If you give your child permission to participate, your child may be selected to complete several assessment s that measure language and literacy skills, such as the Individual Growth and Developmental Indicators (IGDI; Carta, Greenwood, Walker, Kaminski, Good, McConnell & McEvoy), which involves naming items on flashcards. If your child is transitioning to kindergarten, he/she will be administered the ESI-R, which is a brief assessment that m easures kindergarten readiness skills, such as drawing a line and naming objects, that is uti lized on all children entering kindergarten in Pinellas County. Additionally, with your consent, your childÂ’s teacher will complete the Devereux Early Childhood Assessment (LeBuffe & Naglieri, 1998), the Ages and Stages Communication Questionnaire (ASQ) and the Screening for Early Literacy Learning (SELL). These rating scales are designed to assess social/emotional functioning and communication skills in preschool children. If selected, your child also may be observed within his/her classroom setting using a preschool observation checklist that looks at academic and social behaviors. Finally, upon your assent, your child may be photographed and videotaped to document his or her progress in the classroom. You can give permission for your child to receive the assessments and not the photographing or vice versa. Payment for Participation : There will be no payment for participation. Benefits of Being a Part of this Research Study: By taking part in this study, you will provide valuable information about the implementation and outcomes of the LCP: ELO project. This information will be used to modify and improve the current project to increase the early literacy skills of the children in the program. Risks of Being a Part of this Research Study: There are no known risks to participating in this study. Confidentiality of Your Records: Your privacy and evaluation records will be kept confidential to the extent of the law. Authorized research personnel, employees of the Department of Health and Human Services, and the USF Institutional Review Board may inspect the records from this evaluation project. The results of this study may be published. However, the data obtained will be combined with data from other childcare centers in the publication. The published results will not include your childÂ’s name or any other information that would personally identify your child in any way. A pseudonym will be used in place of your childÂ’s name on all documents related to the study and all data will be stored in locked files. Data stored within data bases will be entered with the pseudonym and will be only accessible to the research team through the use of a password. How many other people will take part? About 50 Â– 150 child care providers and about 1500 children and families.
161 Appendix E (Continued): Parental Assent Form Volunteering to Be Part of this Research Study : Your decision to allow your child to participate in this research study is completely voluntary. You are free to allow your child to participate in this research study or to withdraw at any time. There will be no penalty or loss of benefits you or your child are entitled to receive if you stop taking part in the study. Questions and Contacts If you have any questions about this research study, please contact Kathleen Armstrong, Ph.D. at (813) 974-8530.If you have questions about your rights as a person who is taking part in a research study, you may contact the Division of Research Compliance of the University of South Florida at (813) 974-5638. Investigator Statement I have carefully described this study to the parent regarding the nature of the above research study. I hereby certify that to the best of my knowledge that this form explains the nature, demands, risks, and benefits involved in participating in this study. _________________________ _________________________ __________ Signature of Investigator Printed Name of Investigator Date Or authorized research investigator designated by the Principal Investigator
162 Appendix E (Continued): Parental Assent Form Consent to have child take part in this research study (please review options 1 and 2 below) By signing this form I agree that: I have fully read or have had read and explained to me this informed consent form describing this research project. I have had the opportunity to question one of the persons in charge of this research and have received satisfactory answers. I understand the risks and benefits, and I freely give my assent for him/her to participate in the research project outlined in this form, under the conditions indicated in it. I have been given a signed copy of this informed consent form, which is mine to keep. OPTION #1: Permission for assessment and photographing/video-taping 1. I give permission for (___________________________) to participate in this ChildÂ’s name Study by receiving both the assessm ents mentioned in this form and to be photographed and video-taped. _____________________________ ___________________________ Signature of Caregiver of Participant Printed Name of Caregiver Date If you do not wish to have your child participate in one or both components, please sign one of the three options below and return this form to your childÂ’s school or childcare center. OPTION #2: Permission for one component only or No Permission 1. I give my child (____________________) permission to participate in the ChildÂ’s Name assessments but DO NOT give my child permission to be photographed or videotaped. ___________________ ___________________ _________ Signature of Parent Printed Name of Parent Date 2. I give my child (____________________) permission to be photographed/ ChildÂ’s Name video-taped but DO NOT give permission to participate in the assessments. _____________ __________________ __ ______ Signature of Parent Printed Name of Parent Date
163 Appendix E (Continued): Parental Assent Form 3. I DO NOT wish to have my child (____________________) participate in any. ChildÂ’s Name part of this study ________________ ___________________ ________ Signature of Parent Printed Name of Parent Date
164 Appendix F: Parental Information Letter Learning Community Project 8823 115th Avenue, North, Lar go, Florida 33773 Phone (727) 547-4566 Fax (727) 547-4599 Dear Parent/Guardian, I have been selected to participate in a Learning Community Project designed to increase literacy and school readiness for young children in Pinellas County. Along with 3 college credits and free tuition, I will get resource books and materials for my classroom. A literacy coach will make regular visits to help me use what I am learning. As a part of this project all children in the classroom will be screened using different tools, such as a measure of your childÂ’s literacy skills and his/her social and emotional development. Parent educat ors will be available to work with families of children showing signs of needing further screening, and if your child scores meets the criteria or the teacher has concerns, a referral will be made to the appropriate agency. Thank you for supporting me to become better educated so I can provide high quality care for your child. Sincerely, VERY IMPORTANT: FILL IN ALL INFORMATION BELOW! Teacher Name___________ __________________ ______ Center__________________ _______________________ ChildÂ’s Full Name____________ ____________________
165 Appendix G: IGDI Recording Form Picture Naming 1 min Alliteration 2 min Rhyming 2 min Data Pt. Date Student Name Score (# correct) Score (# correct) Score (# correct) 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4
166 Appendix H. HLM Path Diagram Outcome Variable Level One: Rate and Level of Literacy Growth in Preschool Children (IGDI Administration Time 1 Â– Time 4) Level Two: Child Variables (Race, Gender, Home SES, Attendance) Level Three: Childcare Site Variables (Site SES, class size, teacher experience, teacher education, and classroom environment) Early Literacy
ABOUT THE AUTHOR Melissa Farino Todd received her Bac helor of Arts Degree in Psychology in 1997 from the University of South Flor ida (USF). Her education continued at USF as she received her Masters of Arts and Education Specialist Degrees from the School Psychology program, in 1999 and 2003 respectively. She continued to work towards her doctorate while empl oyed full time as a school psychologist in the Pinellas County School District, w here she primarily se rved the Emotionally Handicapped/Severely Emotionally Disturb ed (EH/SED) population. Additionally, Melissa was employed as a Program Eval uator at the Louis de la Parte Florida Mental Health Institute where she ev aluated the outcomes of an early literacy and social development grant. Melissa also held a position at Tampa General Hospital in the Early Interventi on Program, conducting developmental assessments of children birth through age thr ee as well as providing social skills training to school aged children. Thr oughout her employment, Melissa was both nationally certified and state licensed. Since 2005, Melissa has coauthored th ree publications and presented at numerous state and national conferences pertaining to early literacy and behavior. In 2008, Melissa and her co lleagues were awarded an Honorable Mention for the Taylor and Francis Annual Award for Distinguished Journal of Early Childhood Teacher Education Article of the Year. Most recent, she has taken a sabbatical from her professional role to engage in the most important and rewarding position, raising her two daught ers, Kaitlyn Rose and Olivia Rose Todd.