|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
Prevalence of Client Violence against Social Work S tudents and Its Effects on Fear of Future Violence, Occupational Commitment, a nd Career Withdrawal Intentions by Pamela Myatt Criss A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy School of Social Work College of Behavioral and Community Sciences University of South Florida Major Professor: Lisa Rapp-Paglicci, Ph.D. Susan Allen, Ph.D. Marian Dumaine, Ph.D. Alison Salloum, Ph.D. Date of Approval: February 17, 2009 Keywords: social worker, workplace violence, career turnover intentions, career commitment, safety training Copyright 2009, Pamela M. Criss
Dedication This work is first dedicated to my husband, Steve C riss. You are my partner and best friend. You have been patient beyond words, of ten tiptoeing quietly past my workspace, when I knew that you would love to stop and have a conversation. I admire your passion for safety and justice. I look forward to the future with you. I really am done now! This is further dedicated in honor of my mother, Yv onne Myatt, and in memory of my dad, William (Bill) Myatt. You planted in me a d esire to serve people and the ability to see the best in people. I am privileged to be yo ur daughter. Thanks for believing in me. To my precious daughters, Jill, Jaclyn, and Jennif er: You have grown into such beautiful, strong women. Thanks for your support an d love. You patiently sacrificed time with me during my masters program and now again dur ing another degree program. To you, too, I say, I really am done now. I look forwa rd to more time with you! Also, to my stepson, Scott, thanks for your legal expertise, yo ur love, and respect. I am proud of you! To my grandchildren: Cameron, Taylor, Dante, and El lie. You gave me many reasons to smile during the last four years. I pra y that this study will be an inspiration to you one day. Go as far as you can with your educati on. Learn all that you can about the career of your dreams. To all social work students from Southeastern Unive rsity in 2001-2009: You continue to inspire me. Thanks particularly to Trac y Miller, who unknowingly inspired this project.
Acknowledgements To my God, who has blessed me with this rich opport unity, all honor is yours. Thank you, Dr. Rapp-Paglicci, for being a dedicated guide through this doctoral process. I appreciate your interest in the subject, but more than that, I am so appreciative of you being available to field all the questions. Thanks to my other committee members: Dr. Susan All en, Dr. Marian Dumaine, and Dr. Alison Salloum. I appreciate the time inves ted and your thoughtful comments. Thank you, Dr. Roger Boothroyd, for patiently helpi ng me with statistics questions and never once making me feel inadequate. My thanks also go to the administration of Southeas tern University, in particular Dr. Robert Herron and Dr. Lyle Bowlin, past and cur rent Vice Presidents of Academic Affairs, for approving the tuition reimbursements w hich paid for almost all of this degree. Thanks also to my mom for helping to fund a part of my research and degree. Marleen Milner, what would I have done without you? From the application process until this very day, you have helped to kee p me going. You have always had encouragement for my project. Thank you seems ina dequate, but it will have to suffice. Thank you to my technical guru, Patti Andrews, who patiently helped me with formatting, creating the questionnaire, and answeri ng every computer question. Thanks to my husband, children, sons-in-law, grandc hildren, and Ruth for helping to prepare almost 4000 pieces of mail perfectly. Yo u are all truly amazing!
i Table of Contents List of Tables .................................... ................................................... ............................ viiiList of Figures ................................... ................................................... ............................. xiiAbstract .......................................... ................................................... ............................... xiiiChapter I Introduction ............................ ................................................... .......................... 1Statement of Problem .............................. ................................................... ............. 3Purpose ........................................... ................................................... ...................... 4Definition of Terms................................ ................................................... .............. 4Client Violence ................................... ................................................... ..... 4Indirect Experience of Client Violence............. .......................................... 5Physical Assault .................................. ................................................... ..... 5Threat of Physical Harm ........................... .................................................. 5Property Damage ................................... ................................................... .. 5Occupational Commitment ........................... .............................................. 5Career Withdrawal Intentions ...................... ............................................... 5Hypotheses ........................................ ................................................... ................... 6Delimitations ..................................... ................................................... ................... 8Significance and Contribution of the Study ........ ................................................... 8Chapter II Review of the Literature ............... ................................................... ................ 10Social Workers Risk of Client Violence ........... ................................................... 10Theoretical Framework ............................. ................................................... ......... 12Person in Environment Perspective ................. ......................................... 12Work Stress Theory ................................ .................................................. 12Theory of Occupational Commitment ................. ..................................... 14Theory on Beliefs, Attitudes, Intentions, and Behav iors .......................... 14Prevalence and Incidence of Violence to Social Work ers .................................... 15Global Studies in Violence to Social Workers ...... ................................... 15U. S. Studies of Social Work Violence ............. ........................................ 18Client Violence Studies in Specific Regions of the U.S. .............. 19Client Violence in Specific Practice Settings in th e U.S. ............. 20Methods of Sampling in U.S. Studies. .............. ............................ 22Summary of U.S. Social Work Client Violence Studies ............. 30Violence in Social Work Education Student Perspec tive ....................... 31Violence in Social Work Education-Field Director Pe rspective .............. 36
ii Characteristics of Violence towards Social Worker ............................................ 38Types of Violence ................................. ................................................... 38Direct Exposure to Client Violence. ............... .............................. 38Indirect Exposure to Client Violence. ............. .............................. 39Personal/Individual Factors Influencing Client Viol ence ......................... 40Gender. ........................................... ............................................... 40Age. .............................................. ................................................. 4 1Experience......................................... ............................................ 42Ethnicity. ........................................ ............................................... 44Educational Level. ................................ ........................................ 45Organizational/Practicum Related Factors Influencin g Client Violence ..................................... ................................................... ....... 46Place of Client Violence. ......................... ..................................... 46Time of Client Violence............................ .................................... 47Client Violence Per Practice Setting. ............. ............................... 48Race /Ethnicity Match of Field Educator and Student .............................. 51Cross Cultural Supervision. ....................... ................................... 51Supervision Functions in Social Work. ............. ........................... 55Integration of Knowledge on Cross-Cultural Supervis ion and Supervision Functions. ................... ................................... 56Training on Safety and Client Violence............. ....................................... 56Training Student Perceptions ..................... ............................................. 57Training in Social Work Program Venues. ........... ........................ 58Field Agency Safety Policies. ..................... .................................. 59Training Content Areas. ........................... ..................................... 59TrainingField Director and Field Educator Perspec tive ......................... 60Student Training................................... ......................................... 60Field Educator Training. .......................... ..................................... 62Social Work Program Policies. ..................... ................................ 63Field Agency Safety Policies. ..................... .................................. 64Summary on Training. .............................. .................................... 64Effects of Violence on Social Workers ............. ................................................... 65Short Term Effects of Violence on Social Workers .. ............................... 65Long Term Effects of Violence on Social Workers ... ............................... 67Effects of Violence on Social Work Students........ ................................... 69Fear of Future Client Violence .................... ............................................. 70Compromised Occupational Commitment ............... ................................ 71Correlates of Occupational Commitment. ............ ........................ 72Occupational Commitment in Social Workers. ........ .................... 73Occupational Commitment in Students. .............. ......................... 75Measurement of Occupational Commitment. ........... .................... 77Career Withdrawal Intentions ...................... ............................................. 77Evolution of the Variable. ........................ ..................................... 77Career Withdrawal Intentions in Other Fields. ..... ........................ 78Career Withdrawal Intention in Social Work. ....... ....................... 80
iii Measurement. ...................................... .......................................... 81Summary ........................................... ................................................... ................. 82Chapter III Methodology .......................... ................................................... .................... 86Research Design and Methods ....................... ................................................... .... 86Study Sample ...................................... ................................................... ............... 86Data Collection ................................... ................................................... ............... 89Instrumentation ................................... ................................................... ............... 91Personal Demographic Variables .................... .......................................... 91Gender. ........................................... ............................................... 92Age. .............................................. ................................................. 9 2Years of Social Service Experience. ............... .............................. 92Ethnicity. ........................................ ............................................... 92BSW/ MSW. ......................................... ........................................ 92Place of Client Violence. ......................... ..................................... 93Time of Day of Client Violence..................... ............................... 93Practicum Practice Setting. ....................... .................................... 93Field Educator/Student Ethnicity/Race Match........ ...................... 93Training. ......................................... ............................................... 93Subscales: Experience with Client Violence ........ .................................... 94Direct Violence. .................................. .......................................... 94Indirect Exposure to Violence. .................... ................................. 94Subscale: Fear of Future Violence ................. ........................................... 95Subscale: Occupational Commitment ................. ...................................... 95Subscale: Career Withdrawal Intentions ............ ....................................... 97Instrument Pilot Test ............................. ................................................... 97Statistical Procedures ............................ ................................................... ............. 99Variables ......................................... ................................................... ....... 99Statistical Procedures ............................ .................................................. 100Chapter IV Results ............................... ................................................... ....................... 101Descriptive Statistics ............................ ................................................... ............ 101Study Sample of Social Work Students .............. .................................... 101Personal Demographics of Social Work Student Sample ....................... 101Gender. ........................................... ............................................. 102Age. .............................................. ............................................... 102Experience......................................... .......................................... 104Ethnicity/Race. ................................... ......................................... 105Current Social Work Degree Program. ............... ........................ 106Social Work Students Practicum Characteristics ... ............................... 106Home Visits. ...................................... ......................................... 106Work During Evening Hours. ........................ ............................. 107Practicum Practice Settings........................ ................................. 108Race/Ethnicity of the Field Educator. ............. ............................ 109Field Educator/ Student Race Match. ............... .......................... 111Safety Training.................................... ................................................... 112
iv Totals of Training Content Areas. ................. ............................. 113Total Training in the Social Work Program Per Conte nt Area. ........................................ ............................................... 115Individual Training Content in Social Work Practice Classes....................................... ............................................. 117Individual Training Content in Social Work Field Se minar. ...... 118Individual Training Content in Field Practicum Agen cy. ........... 118Other Training Sources. ........................... ................................... 119Additional Analyses Concerning Training. .......... ...................... 119Extent Prepared to Deal With Violent Clients. ..... ...................... 121Extent Prepared Relationships With Other Variables. ................ 122Prevalence of Direct Client Violence against Social Work Students ..... 124Prevalence Rate of Direct Incidents of Client Viole nce. ............ 124Prevalence Rate by Types of Violence. ............. ......................... 124Total Incidents of Direct Exposure to Client Violen ce. .............. 124Transformation of Direct Violence Variables. ...... ...................... 125Quantities of Direct Incidents by Type of Client Vi olence. ....... 126Prevalence of Indirect Client Violence towards Soci al Work Students ..................................... ................................................... ...... 127Prevalence Rate of Indirect Exposures to Client Vio lence. ........ 128Prevalence Rate by Types of Indirect Exposure to Vi olence. .... 128Total Incidents of Indirect Exposure to Client Viol ence. ........... 129Quantities of Indirect Client Violence by Type of C lient Violence. .................................... ............................................ 129Fear of Future Violence ........................... ............................................... 130Reliability and Sampling Adequacy. ................ .......................... 130Results of Exploratory Factor Analysis. ........... .......................... 131Fear of Violence Descriptive Information. ......... ........................ 131Occupational Commitment ........................... .......................................... 132Reliability and Sample Adequacy. .................. ............................ 132Exploratory Factor Analysis. ...................... ................................ 132Affective Commitment (AC). ........................ ............................. 133Normative Commitment (NC). ........................ ........................... 133Continuance Commitment (CC). ...................... .......................... 133Occupational Commitment Descriptive Information. .. ............... 134Career Withdrawal Intentions ...................... ........................................... 136Findings Related to Prevalence of Client Violence H ypotheses ........................ 137Hypothesis 1-1 .................................... ................................................... 137Hypothesis 1-2 .................................... ................................................... 139Hypothesis 1-3 .................................... ................................................... 141Hypothesis 1-4 .................................... ................................................... 143Hypothesis 1-5 .................................... ................................................... 147Hypothesis 1-6 .................................... ................................................... 149Hypothesis 1-7 .................................... ................................................... 151Hypothesis 1-8 .................................... ................................................... 155
v Hypothesis 1-9 .................................... ................................................... 157Hypothesis 1-10 ................................... ................................................... 160Rate of Violence Per Training in Social Work Classe s. ............. 160Rate of Violence Per Training in Field Seminar. ... ..................... 161Rate of Client Violence Per Training at the Field A gency. ........ 161Rate of Client Violence Per Training in Other Place s. ............... 162Hypothesis Testing................................. ..................................... 162Extent of Preparation. ............................ ..................................... 167Findings Related to Fear as Mediator Hypotheses ... .......................................... 167Client Violence Relationship with Occupational Comm itment and Career Withdrawal Intentions ................. ........................................... 168Client Violence and Affective Commitment. ......... .................... 169Client Violence and Continuance Commitment. ....... ................. 169Client Violence and Normative Commitment. ......... .................. 170Client Violence and Career Withdrawal Intentions. .................. 170Fear of Future Violence as Mediator ............... ....................................... 170Hypothesis 2-1 .................................... ................................................... 171Direct Client Violence and Fear of Future Violence. ................. 171Indirect Exposure to Client Violence and Fear of Fu ture Violence. .................................... ............................................ 172Hypothesis 2-2 .................................... ................................................... 173Fear of Future Violence and Affective Commitment. ............... 174Fear of Future Violence and Normative Commitment. ............. 174Fear of Future Violence and Continuance Commitment. ........... 175Hypothesis 2-3 .................................... ................................................... 175Predictor of Violence Analyses .................... ................................................... ... 176Predicting Client Violence Based on the Effect of I ndependent Variables .................................... ................................................... ..... 176Methods of Analysis. .............................. .................................... 177Data Screening. ................................... ........................................ 178Screening For Multicolinearity. ................... ............................... 178Assumptions of Multiple Regression Analyses. ...... ................... 179Multiple Regression Analyses Results............... ..................................... 180Predicting Sum of Client Violence. ................ ............................ 180Predicting of Physical Assault. ................... ................................ 181Predicting of Threatened Physical Harm. ........... ........................ 182Predicting Verbal Abuse. .......................... .................................. 182Predicting Threat of Lawsuit. ..................... ................................ 183Predicting Damage to Personal or Professional Prope rty. .......... 183Chapter V Summary and Discussion .................. ................................................... ......... 185Summary of Findings ............................... ................................................... ........ 185Response Rate ..................................... ................................................... 185Characteristics of Sample ......................... .............................................. 185Gender/Age. ....................................... ......................................... 185Experience......................................... .......................................... 186
vi Race/Ethnicity. ................................... ......................................... 186Degree Program. ................................... ...................................... 186Home Visits. ...................................... ......................................... 187Evening Hours. .................................... ....................................... 187Practice Settings. ................................ ......................................... 187Training. ......................................... ............................................. 188Occupational Commitment. .......................... .............................. 190Career Withdrawal Intentions. ..................... ............................... 192Summary of Prevalence of Direct Exposure to Client Violence ............ 192Summary of Prevalence of Indirect Exposure to Clien t Violence .......... 193Summary of Prevalence of Client Violence Hypotheses ........................ 194Summary of Additional Analysis..................... ....................................... 200Summary of Fear as Mediator Hypotheses ............ ................................. 203Age and Client Violence ........................... .............................................. 206Work Experience and Client Violence ............... .................................... 208Ethnicity and Client Violence ..................... ............................................ 210Degree Program and Client Violence ................ ..................................... 212Practice Setting and Client Violence .............. ........................................ 216Field Educator/Student Race Supervision Dyads and C lient Violence ..................................... ................................................... ..... 218Training and Client Violence ...................... ............................................ 218Client Violence and Occupational Commitment ....... ............................. 221Client Violence and Career Withdrawal Intentions .. .............................. 222Client Violence and Fear of Future Violence ....... .................................. 223Fear of Future Violence and Occupational Commitment ....................... 225Fear of Future Violence and Career Withdrawal Inten tions ................... 227Fear of Future Violence as Mediator ............... ....................................... 227Implications for Social Work Education............. ................................................ 22 8Micro Level Planning and Intervention ............. ..................................... 228Field Placement and Practice Considerations. ...... ...................... 228Field Faculty Consultation to Students. ........... ........................... 229Use of Field Seminar to Support Students. ......... ........................ 230Mezzo Level Planning and Intervention ............. .................................... 231Field Agency Selection. ........................... ................................... 231Agency Policies and Procedures. ................... ............................. 232Social Work Educations Curriculum on Violence. ... ................ 233Content of Safety Training Curriculum. ............ ......................... 234Methods of Training on Safety and Violence. ....... ..................... 237Training for Field Educators. ..................... ................................. 239Macro Level Planning and Interventions ............ .................................... 240Social Work Program Policies. ..................... .............................. 240CSWE Policies. .................................... ....................................... 241Federal Policy on Social Work Safety. ............. .......................... 242Coordination of Safety Training Throughout the Soci al Work Program. ................................ ....................................... 243
vii Risk Management Concerns. ......................... ............................. 245Study Limitations ................................. ................................................... ............ 247Suggestions for Future Research ................... ................................................... .. 248Conclusion ........................................ ................................................... ............... 252References ........................................ ................................................... ............................ 253Appendices ........................................ ................................................... ........................... 265Appendix A: Tables and Figures .................... ................................................... 266Appendix B: IRB Letter ............................ ................................................... ....... 334Appendix C: in Focus NASW Information ............. ........................................... 336Appendix D: Client Violence Questionnaire ......... ............................................. 338Appendix E: First Mailing Cover Letter ............ ................................................. 3 43Appendix F: Second Mailing Cover Letter ........... .............................................. 345Appendix G: Pre-Mailing Post Card ................. .................................................. 347Appendix H: Emailed Permissions to Use Scales/Quest ions ............................. 348 About the Author End Page
viii List of Tables Table 1 Summary of Hypotheses, Variable, Instrum ents, and Statistical Methods........................................ ................................................... ................ 267 Table 2 Personal Demographics of Respondents ... ................................................... ... 269 Table 3 Organizational/Professional Demographics .................................................. 271 Table 4 Total Training in the Social Work Progra m per Content Area ....................... 273 Table 5 Percentages of Students Who Had Safety T raining per Training Venue ....... 274 Table 6 Percentages of Students Who Did Not Have Safety Training per Training Venue .................... ................................................... ........................ 276 Table 7 Prevalence Rate of Direct Client Violenc e during Practicum ........................ 277 Table 8 Rate of Direct Violence by Type of Viole nce ............................................... 278 Table 9 Total Incidents of Direct Client Violenc e during Practicum .......................... 279 Table 10 Quantities of Incidents of Direct Client Violence .......................................... 280 Table 11 Prevalence of Indirect Exposure to Client Violence during Practicum .......... 282 Table 12 Prevalence of Indirect Exposure to Client Violence by Type of Client Violence .......................... ................................................... ............................. 283 Table 13 Total Incidents of Indirect Exposure to C lient Violence ................................ 284 Table 14 Incidents of Indirect Exposure to Client Violence by Type of Violence........ 285 Table 15 Descriptive Statistics of Fear of Future Violence .......................................... 287 Table 16 Results of Exploratory Factor Analysis of Occupational Commitment (Revised per Initial Factor Analysi s) ................................................ .............. 288
ix Table 17 Descriptive Statistics of Occupational Co mmitment (After Factor Analysis Adjustments) ............. ................................................... .................... 289 Table 18 Descriptive Statistics of Career Withdraw al Intentions .................................. 29 0 Table 19 Results of Chi Square Tests for Practicum Exposure Rate to Client Violence by Gender ................ ................................................... ..................... 291 Table 20 Results of t-tests for Incidents of Clien t Violence by Gender ........................ 293 Table 21 Results of Chi Square Tests for Practicum Exposure Rate to Client Violence per Age Category ......... ................................................... ................. 294 Table 22 Results of ANOVAs with Means and Standard Deviations of Client Violence among Age Categories ..... ................................................... ............ 296 Table 23 Results of Chi Square Tests for Practicum Exposure Rate to Client Violence per Experience Level ..... ................................................... ............... 297 Table 24 Results of Pearsons Correlations for Inc idents of Client Violence By Years of Experience ............ ................................................... ................... 298 Table 25 Results of ANOVAs with Means and Standard Deviations of Client Violence among Experience Categorie s ................................................. ........ 300 Table 26 Results of Chi Square Tests for Practicum Exposure Rate to Client Violence by Race/Ethnicity ........ ................................................... ................. 301 Table 27 Results of ANOVAs with Means and Standard Deviations of Client Violence among Race/Ethnic Groups ................................................... ......... 303 Table 28 Results of Chi Square Tests for Practicum Exposure Rate to Client Violence by Degree Program ........ ................................................... ............... 304 Table 29 Results of t-tests for Incidents of Clien t Violence by Degree Program ........................... ................................................... ............................. 306 Table 30 Practicum Exposure Rate to Client Violenc e per Place of Violence .............. 307 Table 31 Results of ANOVAs with Means and Standard Deviations of Client Violence Incidents of Students Maki ng Home Visits per Number of Home Visits Made During Practicum ................................................... ......... 308 Table 32 Practicum Exposure Rate to Client Violenc e per Time of Day ...................... 309
x Table 33 Results of Chi Square Tests for Practicum Exposure Rates per Percent of Evening Hours Worked ... ................................................... ........... 310 Table 34 Results of ANOVAs with Means and Standar d Deviations of Client Violence per Percentage of Evening Hours Worked ..................................... 311 Table 35 Results of Chi Square Tests for Practicum Exposure Rates to Client Violence per Practice Settin g ................................................. .............. 312 Table 36 Results of ANOVAs with Means and Standard Deviations of Client Violence among Practice Settings .. ................................................... ............. 314 Table 37 Results of Chi Square Tests for Practicum Exposure Rates per Race Match of Student/Field Educato r ................................................. .......... 315 Table 38 Results of ANOVAs with Means and Standard Deviations of Client Violence per Race/Ethnic Match of S tudent/Field Educator .......................... 31 7 Table 39 Results of Pearsons Correlations between Incidents of Client Violence and Total Training in Trai ning Venues ....................................... .... 318 Table 40 Results of Pearsons Correlations between Incidents of Client Violence and Total Training per Tra ining Content Area ................................ 319 Table 41 Results of Pearsons Correlations between Direct Incidents of Client Violence and Occupational Co mmitment/Career Withdrawal Intentions ............. ................................................... ..................... 321 Table 42 Results of Pearsons Correlations between Indirect Exposure to Client Violence and Occupational Co mmitment/Career Withdrawal Intentions ............. ................................................... ..................... 322 Table 43 Results of Pearsons Correlations between Direct Experience of Types of Violence and Fears of Same Type of Violence and General Fear of Violence .......... ................................................... ................... 323 Table 44 Results of Pearsons Correlations between Indirect Exposure to Types of Violence and Fears of Same Type of Violence and General Fear of Violence .......... ................................................... ................... 324 Table 45 Results of Pearsons Correlations between Fear of Violence and Occupational Commitment ........... ................................................... ............... 325 Table 46 Model Summary for Predictors of Occurrenc e of Overall Client Violence .......................... ................................................... ............................. 326
xi Table 47 Model Summary for Predictors for Frequenc y of Overall Client Violence .......................... ................................................... ............................. 327 Table 48 Model Summary for Predictors of Occurrenc e of Physical Assault............................ ................................................... .............................. 328 Table 49 Model Summary for Predictors of Occurrenc e of Threat of Physical Harm ..................... ................................................... ......................... 329 Table 50 Model Summary of Predictors of Occurrence of Verbal Abuse ..................... 330 Table 51 Model Summary of Predictors of Occurrenc e of Threat of Lawsuit ........................... ................................................... .............................. 331 Table 52 Model Summary for Predictors of Occurrenc e of Property Damage ............................ ................................................... ............................ 332 Table 53 Percentages of Field Directors Reporting Giving Training Content versus Student Reporting Re ceiving Training .................................. 333
xii List of Figures Figure 1 Research Model ........................ ................................................... .................. 266 Figure 2 Percent of Social Work Students Receivi ng Client Violence Training Content Areas by Venue ........... ................................................... .................. 275 Figure 3 Types of Direct Client Violence Incident s against Social Work Students ......................... ................................................... .............................. 281 Figure 4 Types of Indirect Exposure to Client Vio lence by Social Work Students ......................... ................................................... .............................. 286 Figure 5 Mean of Summed Client Violence Incidents per Gender ............................... 292 Figure 6 Mean of Summed Client Violence Incidents per Age Categories .................. 295 Figure 7 Mean of Summed Client Violence Incidents per Experience Categories ...... 299 Figure 8 Mean of Summed Client Violence Incidents per Student Race/Ethnicity ..... 302 Figure 9 Mean of Summed Client Violence Incidents per Degree Program ................ 305 Figure 10 Mean of Summed Client Violence Incidents per Practicum Setting .............. 313 Figure 11 Mean of Summed Client Violence Incidents per Student/Field Educator Race Match ....................... ................................................... .......................... 316
xiii Prevalence of Client Violence Against Social Work S tudents and Its Effects on Fear of Future Violence, Occupational Commitment, and Caree r Withdrawal Intentions Pamela Myatt Criss ABSTRACT Social work literature has documented that social w orkers may be the victims of client violence. However, to date, no studies have documented the nationwide prevalence of client violence towards social work students. Th is study examined direct and indirect incidents of physical assault, threats of physical harm, verbal abuse, threats of lawsuit, and property damage. The randomly selected national sample of social work students were selected from the National Association of Soci al Workers ( N = 595). Findings revealed that 41.7% of social work studen ts directly experienced client violence during their practicum. The highest rate o f the violence reported by students was verbal abuse (37.5%) while the lowest rate of repor ted violence was physical assault (3.5%). Being male was the most significant predict or of social work students exposure to client violence. Other factors related to increa sed violence were found, such as ethnicity and degree program. This study also exami ned whether students received safety training in 17 content areas and where they receive d the training. Fewer than 50% of students received training in most training content areas, regardless of where training was received. Furthermore, increased safety training in the field agency was significantly related to increased threats of physical harm and o verall client violence. When training
xiv from all venues was totaled, increased training was significantly related to increased verbal abuse, property damage, and overall direct c lient violence. This study found that when students experience clie nt violence directly or indirectly, they have increased fear of future viol ence in social work practice. Implications for social work programs, field agenci es and educators and social work students are discussed. Training content and strate gies are suggested.
1 Chapter I Introduction Consider the 2006 news story from Henderson, Kentuc ky: A nine month old boy, his mother and her boyfriend were still missing Wednesday, two days after a social worker was found dead in the mothers home. Boni Frederick, 67, had taken the bo y, who has been in foster care, to his mothers house for a visit on M onday. Police found her body after she failed to return to work, and she ap peared to have been beaten to death... It was bluntforce trauma and sharpinstrument laceration, Sgt. John Nevels of the Henderson Poli ce Department said Tuesday. There was definitely a struggle. Advocat es said the slaying emphasizes the danger of social work (Lenz, 2006). During their career, social workers have a high ris k of encountering client violence in the agencies where they serve. More tha n two dozen studies over the past twenty years have documented the occurrence of viol ence to social workers while they are in their workplaces. A recent national study of 5000 social workers was conducted by NASW Center for Workforce Studies (NASW, 2006b). It found that 47% of social workers had concerns about personal safety and near ly 30% reported that their employers do not adequately address safety concerns. National prevalence rates for social workers exposure to violence indicate that between 65%-86% of social workers have encountered client violence at some time during their career (B eaver, 1999; Ringstad, 1995). When social workers are the victims of client viole nce, they can experience the acute effects of trauma that include becoming timid withdrawn, frightened, nervous, and
2 angry (Guy & Brady, 1998; Horwitz, 2006; Norris, 19 90; Snow, 1994). With increased fear of future occurrences of violence and less con fidence in their abilities to serve clients, social workers may be reluctant to risk pu tting themselves in a similar situation. They may be prone to abandon their commitment to th e profession of social work and may have intentions of leaving the profession. Like veteran social workers, social work students m ay also be at risk of experiencing client aggression. When these emerging members of the profession experience client violence, they may be immobilized by the fear that this could occur again. They may question their choice of career and make a decision to change their career before they ever become independent social w ork practitioners. Unfortunately, much is still unknown about the prev alence of client violence toward social work students. In the 7 studies that have surveyed social work students, sample sizes have been small and regionalized. In t hese studies, prevalence rates for students encountering violence in their practicum h ave ranged from 21% 54% (Knight, 1999; Mama, 2001). Studies indicate that prevalence statistics for social workers exposure to client violence may be underestimated d ue to the fact that many incidents go unreported (Norris, 1990; MacDonald & Sirotich, 200 1). Thus the actual incidents of client violence toward social work students are lik ely to be more than the statistics reveal. Even if the lower prevalence rate were more accurat e, one in five social work students may be at risk of harm from client violence.
3 Statement of Problem Social work students may be harmed by client violen ce during their practicum experience. Because few studies have explored incid ence of client violence toward social work students, it is unclear how often it occurs. I f students are harmed, there may be far reaching personal, professional, and organizational effects. As with any other victims of violence, when students are hurt or purposefully fr ightened by clients, they may have acute reactions to this trauma. Fear of future viol ence in the field of social work may weaken their commitment to the profession. They may then make plans to leave the profession before they begin. Their potential as a social worker may never be reached. Organizationally, field agencies and social work sc hools may suffer financially if they are liable for injuries when they have not ade quately prepared the student for encountering the risk of violence. Agencys employe es may be affected negatively through their awareness of the potential that they too could be harmed. Finally, and perhaps most important to the mission of social wor k, clients may be affected as potential members of the profession are lost. Violence statistics may minimize the impact that cl ient violence has on students lives. Even those who work most closely with social work students may be lulled into forgetting that violence may irretrievably wound pe ople. It is the responsibility of social work educators and practitioners to learn as much a s possible about this phenomenon in order to help to prevent occurrences of violence to students.
4 Purpose This study is a national probability study that que stions social work students experience with client violence. It examines the pr evalence of client violence as it relates to personal factors such as gender, age, previous e xperience, ethnicity, and social work degree program. Additionally, it explores organizat ional factors such as place and time of the violent incident, training on safety and violen ce, social work practice setting, and ethnic mix of the field supervision dyad to see how these influence exposure to client violence. From these factors, analyses will be cond ucted to determine which factors best predict the occurrence of client violence. Finally, this study is the first to reach beyond prevalence and incidence studies of client violence toward social work students, as it considers possible effects of client violence on th e lives of social work students. This study explores relationships between social work st udents exposure to violence, fear of future violence and the potential decrease in caree r commitment and increase in career withdrawal intentions. Definition of Terms Client Violence Intentional personal or agency property damage, thr eats, verbal abuse, threat of lawsuit, attempted physical harm, or actual physica l harm against social workers or other service providers by individuals who are applicants recipients, or former recipients of those services (Newhill, 2003).
5 Indirect Experience of Client Violence Violent or aggressive client violence event that a social worker has heard about and/or witnessed during ones social work practicum Physical Assault An incident in which an individual actually lays ha nds or a weapon on another individual with the intent to harm (Newhill, 2003). Threat of Physical Harm Verbal or written threat to harm another individual or deliberate threatening physical gesture (including stalking) from one indi vidual to another (Newhill, 2003). Property Damage An incident in which an individual intentionally da mages another individuals personal property or property the individual was se en using at the time of the incident (Newhill, 2003). Occupational Commitment A persons belief in and acceptance of values of hi s or her chosen occupation or line of work and willingness to maintain membership in that occupation (Vandenberg & Scarpello, 1994). Career Withdrawal Intentions The extent to which an individual has thought about leaving his/her profession (Blau, 1985).
6 Hypotheses Based on the literature review of social work stude nts and their experience with client violence, five personal demographic variable s and five organizational demographic variables are incorporated in this study in order t o gain understanding about the prevalence of violence towards social work students The following hypotheses regarding personal demographics are examined: Hypothesis 1-1 Male social work students will exper ience more of every type of client violence than female social work students. Hypothesis 1-2 Younger social work students will ex perience more of every type of client violence than older social work students. Hypothesis 1-3 Less experienced social work student s will experience more of every type of client violence than more experienced social work students. Hypothesis 1-4 There will be no difference in expos ure to client violence by students of differing ethnicities/ racial backgroun ds. Hypothesis 1-5 BSW students will experience more of every type of client violence than MSW students. The relationship between the following organization al factors related to the practicum setting and prevalence of client violence are explo red through these hypotheses: Hypothesis 1-6 There will be a significant differen ce in numbers of client violence incidents occurring to social work student s according to the place of social work practice (home visit, offi ce, other).
7 Hypothesis 1-7 There will be a significant differen ce in numbers of client violence incidents occurring to social work student s according to the time of day. Hypothesis 1-8 There will be a significant differen ce in numbers of client violence incidents occurring to social work student s according to their practice setting. Hypothesis 1-9 There will be a significant differe nce in numbers of client violence incidents occur ring to social work students according to the race/ethnicity match of the students and their field educators. Hypothesis 1-10 Social work students who have had t raining in more content areas regard ing safety and violence will have fewer incidents o f client violence. The effects of client violence on social work stude nts are explored by examining the relationship between incidents of client violence, fear of future violence, occupational commitment, and career withdrawal intention. Fear o f future violence is proposed as a mediator between experience of client violence and the outcome variables, such that: Hypothesis 2-1 Experience of client violence has a positive and direct effect on fear of future client violence. Hypothesis 2-2 Fear of future client violence has a negative and direct effect on career commitment. Hypothesis 2-3 Fear of future client violence has a negative and direct effect on career turnover intention.
8 Delimitations The following are delimitations of this research st udy: 1. The sample for this study was drawn from the nat ional student membership in the National Association of Social Workers. Requirement for student membership in this organization is that students must be enrolled in social work programs that are in candidacy for accreditation by the Council of So cial Work Education (CSWE) or are fully accredited by CSWE. 2. The study included students who were currently p articipating in a field placement or current students who had completed their field p lacement in their social work program. Significance and Contribution of the Study To date, there have been few studies of client vi olence toward social work students. The studies that exist are predominantly non probab ility studies within specific social work programs, although one study included students from eight different programs (Knight, 1996). This study was the first national p robability study of social work students on the subject of client violence. As such, it was anticipated that the findings of the study would have greater generalizability to the entire p opulation of social work students in the U.S. It was also a matter of interest that in manua lly reviewing approximately six most recent years of issues of two journals on social wo rk education, this researcher could only find seven national random samples of social work s tudents on any subject. Most empirical studies drew their samples from one or tw o schools of social work. in order to
9 better understand the experiences of social work st udents at large, larger probability studies should prove helpful. Secondly, no previous social work studies have so ught to examine the association between client violence, career commitment, and car eer withdrawal intentions of social workers or social work students. In fact, few studi es have looked at the professional socialization of social work students, what types o f variables cause them to be committed to the profession of social work, and at what point this commitment initially takes place (Baretti, 2004). Finally, only one study has attempted to look at specific violence training that social work students have received and the possibility tha t training may reduce incidents of client violence (Elwood & Rey, 1996). It is hoped that the findings of this research will help schools of social work and field practicum sit es to develop training programs that will help to protect social work students from beco ming victims of violence, decrease their fear of future violence and sustain their int erest in the profession of social work.
10 Chapter II Review of the Literature This chapter contains a review of relevant literat ure pertaining to the key concepts of this study. It begins with a general introductio n to client violence towards social workers and theoretical underpinnings of the study. There follows a discussion of prevalence of client violence towards social worker s globally and in the United States. Methodologies of the U.S. studies are summarized. T he literature then focuses on client violence towards social work students. The next sec tion of the review focuses on literature concerning the personal and organization al factors that are the focus of the hypotheses in the study. Finally, literature concer ning the effects of client violence is discussed, with specific focus on fear of future vi olence, occupational commitment, and career withdrawal intentions. Social Workers Risk of Client Violence Social workers are at increased risk of becoming vi ctims of workplace violence. Bureau of Labor (BLS) data from 2000 showed that 48 % of all nonfatal injuries from occupational assaults and violent acts occurred in health care and social services (OSHA, 2004). BLS further reports that health care and soc ial services are at high risk for violent assault at work. The workplace injury rate for soci al service workers is seven and a half times higher than the workplace injury rate for the overall private sector.
11 In the Department of Justices National Crime Victi mization Survey, the average annual rate of non fatal violent crimes for mental health professionals was 68.2 per 1,000, compared with 12.6 per 1,000, the average rate for all occupations (OSHA, 2004). This increased risk for mental health professionals is n oted to result from systemic issues, such as: increased use of hospitals for the care of acut ely disturbed individuals; increasing numbers of chronically mentally ill people being re leased from hospitals without follow up care; long waits at agencies; and low staffing l evels. Low staffing levels may lead to situations of doing isolated work during treatment, and solo work in remote locations, with no back up or way to get assistance. Additiona lly, societal problems such as the prevalence of handguns, increasing members of gangs and increasing presence of drug and alcohol use may contribute to increased client violence. The Department of Justice survey also notes that there is a lack of training in recognizing and managing escalating and assaultive behavior. Some social work researchers have stated that risk of violence by clients is not inevitable, indicating that we need to be careful a bout accepting client violence as a reality for social workers (Ringstad, 2005). Howev er, other scholars have emphasized that without acknowledging that client violence tow ard social workers is a reality, we will not have the capacity to both prevent it and respon d effectively when it has occurred (Guy & Brady, 1998; Newhill, 2003). Brockmann (2002 ) stated that because social workers work with the most vulnerable populations, it may be more realistic to reduce, rather than eliminate violence. Social workers, par ticularly those newest to the field,
12 should be prepared to avoid violent encounters, yet be confident in seeking support if it is necessary to reduce the effects of client violence. Theoretical Framework Person in Environment Perspective Breakwell and Rowett (1989) noted that there had be en an absence of theory building perspective on the subject of client viole nce toward social workers. They then conducted their study using Steadmans situational approach to the study of violent incidents. The premise of this theory is that viol ence is influenced by a number of variables stemming from the interaction between the aggressor and the victim in certain situational context. (p. 242-243). The situational context of violence is critical in determining the beginning of the violence, the cour se of the violence and consequences. Some studies have outlined the need to understand b oth preventive and reactive strategies in order to effectively reduce the risk of violence in the workplace. In his article on an ecological view of psychological trauma, Harv ey (1996) discussed the use of a Person X Event X Environment Influences model. He s tated that individual vulnerability to victimization and individual recovery patterns a re determined by interactions between person factors, such as age and gender; event facto rs, such as the severity, duration, and frequency of trauma; and environmental factors, suc h as the victims support system. Work Stress Theory Theoretically, the phenomenon of client violence ca n be studied using the tenants of work stress theory. This theory comes from a psy chological stress framework which says that psychological stress is a particular rel ationship between the person and
13 environment that is appraised by the person as taxi ng or exceeding his or her resources and endangering his or her well-being (Lazarus & F olkman, 1984, p.19). It is therefore the relationship or the transaction between the sti mulus and response that indicates a stressor. What is stressful to one person may not b e stressful to another. It is the appraisal of the situation that ultimately determines the out come for an individual. The cognitive appraisal of a stimulus such as workplace violence determines the degree of stress and therefore the level of response. In the primary app raisal, the individual assesses whether they have been harmed and whether they may be at ri sk of the same harm in the future. The secondary appraisal helps to determine what wil l be done to manage the threat of harm. For example, Barling (1996) hypothesized that fear of violence would be a major consequence of workplace violence. Fear, in turn, h as been found to be associated with thoughts about quitting an organization (Rogers & K elloway, 1997). When a victim of workplace violence appraises his/her situation as t hreatening, fear of future victimization may act as a mediator to determine whether the pers on will cope by wanting to withdraw from the organization or the career to which they w ere previously committed. Lazarus and Folkman (1984) stated that fear is the manifestation of a specific stress appraisal. Changes in fear level indicate th at there are changes in the way the person is appraising his/her relationship with the environment. If the person has higher self efficacy or belief that they can master a situ ation, they feel a sense of control and appraise a situation as being less threatening. App lication of this theory to client violence would indicate that when a social work student has a stronger belief that a future client
14 violence situation can be avoided, there will be le ss fear of future violence, which would in turn have positive occupational effects for the student. Theory of Occupational Commitment Theoretical discussion of occupational commitment h as pointed to at least three views of an individuals career behaviors. The firs t is a differentialist view espoused by Van Maanen & Schein (1977). They emphasized individ ual or personality predictors of how an individual views their career. The organizat ional view, discussed by the same researchers, emphasized organizational or situation al factors as influences of a persons career motivation. Using these ideas, London (1983) intertwined the two, stating that both individual and organizational variables may in fluence ones career choice and continuance. Theory on Beliefs, Attitudes, Intentions, and Behav iors Fishbein and Ajzens (1975) theory on the relations hip between beliefs, attitudes, intentions and behaviors may be beneficial to help explain career withdrawal intentions. This theory emphasizes the role of behavioral inten tions in understanding the relationship between attitudes and behaviors. Beliefs are develo ped from information that we have about an object or incident. Attitudes are then dev eloped about the object or incident based on the beliefs that one has. Favorable attitu des are generally developed toward objects that we associate with good things and unfa vorable attitudes are typically associated towards objects we associate with bad th ings. Behavioral intentions then develop on the basis of a persons attitudes. Fishb ein and Ajzen define a behavioral intention as a persons subjective probability tha t he will perform some behavior
15 (p. 288). Behavioral intentions are typically follo wed by the behavior itself. Using Fishbein and Ajzens theoretical model, studies hav e looked at behavioral intentions to stay or to leave a place of employment. Mobley and colleagues (1979) proposed a complex model of employment turnover and stated tha t the best predictor of turnover is intention to quit. Though the current study looks a t career withdrawal intentions, as opposed to employment withdrawal intentions, this t heory may help to explain students behavioral intentions to stay or to leave the profe ssion of social work. Prevalence and Incidence of Violence to Social Work ers Global Studies in Violence to Social Workers In addition to documentation of prevalence of clie nt violence towards social workers in the U.S., studies have documented the pr evalence of violence in social work in Great Britain (Littlechild, 2005; Lyons, Lavalle & Grimwood, 1995; Norris, 1990; Rowett, 1986; Smith & Nursten, 1998), Australia (Gr een, Gregory, & Mason, 2003), Finland (Littlechild, 2005), Scotland (Leadbetter,1 993) and Canada (MacDonald & Sirotich, 2005; Schat & Kelloway, 2003; Snow, 1994) Rowett (1986) published the first major study in England, which found that when 132 p rogram managers were asked about incidents of violence towards their workers, the pr ojected ratio of physical assault was 1 in 259 workers. However, when Rowett asked 728 soci al workers directly, 1 in 4 had actually been assaulted, suggesting that program ma nagers may not always be aware when their workers are encountering violence or the y may minimize or deny the extent of client violence towards workers. This is pertinent to the current study because it indicates that program managers may not be aware of the exten t of client violence. More accurate
16 accounting of client violence may best come from th ose who have directly or indirectly encountered the violence. Another international study that has relevance to t he current study was completed by Norris (1990). The key contribution to the liter ature was the finding that much client violence is never reported. All of 38 field worker and residential childrens home social workers in Norris first study had been threatened, yet only half of them had told managers about the threats. His further study with 79 social service and probation agencies found that 40% did not monitor threats tow ards workers. This study found an average of 2.6 incidents of client violence annuall y. Another study in Scotland likewise found that reporting client violence was a problem (Leadbetter, 1993). This researcher was one of the few who has used multimodal methods to investigate client violence. One of the four methods used was a two week diary study with residential staff at a childrens home. It was found that 53 staff listed 131 inciden ts of verbal abuse, 14 incidents of moderate physical aggression, and 10 incidents of p hysical assault. Of major concern was the fact that none of the incidents was formally re ported. Lyons, LaValle, and Grimwood (1995) conducted the only study to date that looked at violence as a possible reason for turnove r in the profession of social work. They completed a study on career patterns of 791 social workers who had graduated from 21 randomly chosen Welsh and English academic institut ions in four predetermined years: 1991, 1987, 1983, and 1979. The intent of the study was to determine how many social workers left the profession after their training an d the reasons for leaving. A pilot survey indicated that verbal and physical abuse by clients was one potential reason that social
17 workers may leave the field. Respondents reported t hat 92% of the respondents had been verbally abused, 32% experienced physical violence and 68% were threatened with violence in their career. Twenty four percent of th ose who had left the field of social work had experienced client violence and another tw elve percent of those who were victims of client violence said that they had consi dered leaving their career. A multidisciplinary study of healthcare providers i n Ontario, Canada included a small percentage (6.9%) of social workers in an ove rall sample of 225 providers (Schat & Kelloway, 2003). This study is one of only a few st udies that have looked at attempting to reduce adverse consequences of workplace aggression and violence. By using a series of moderated multiple regressions, the researchers fou nd that organizational support moderated the effects of physical violence, vicario us exposure to violence, and psychological aggression on emotional well being, s omatic health, and job-related affect. It did not moderate fear of future violence. Though this study included only a small number of social workers, from the work stress theo retical perspective, it contributed much toward understanding possible negative effects of workplace violence on personal and professional outcomes. In one of four international qualitative studies, S mith and Nursten (1998) asked 24 social workers from a social service department in England to recall one distressing experience in their working life and to comment on what had been helpful in their attempts to process the experience. Fear of being a ssaulted was the experience that was most often mentioned by the participants, indicatin g that the psychological stress response of fear can have potent long term effects on social workers.
18 Because of the possible differences in expectations for social workers in different countries, it may be difficult to generalize the re sults of international studies to social workers in the U.S. However, these studies suggest that violence against social workers is a global phenomenon that demands to be addressed by the social work profession as a whole. U. S. Studies of Social Work Violence The earliest studies in the U.S. containing informa tion about social workers exposure to client violence were studies of multipl e disciplines (Bernstein, 1981; Whitman, Armao, & Dent, 1976). These studies indica ted that social workers were not harmed as often as psychiatrists (Bernstein, 1981) and nurses (Carmel & Hunter, 1989) but they were being attacked and threatened at subs tantial rates and sometimes injured by clients (Carmel & Hunter, 1993). A more current multidisciplinary study surveyed al l staff at the University Of Rochester Medical Center Of Psychiatry concerning i ncidents of endangerment, threats and assaults (Privitara, Wiesman, Cerulli, Tu, & Gr oman, 2005). With a total sample of 380, the clinical staff made up 69% of the responde nts. Among them were 32 persons with an MSW and 13 persons with a BSW. Forty six pe rcent of those with an MSW had been endangered, 41% had been threatened, and 9% ha d been assaulted. Of the individuals with a BSW, 46% had been endangered, 62 % had been threatened and 23% had been assaulted. Though nurses, physicians and a dvanced practical nurses had significantly more incidents of all types, the perc entages for social workers were substantially high.
19 Client violence studies in specific regions of the U.S. Several studies about client violence have been focused in particular geographic areas of the country, making it somewhat difficult to generalize the findings to so cial work workers across the U.S. Among these have been a study in Massachusetts (Atk inson, 1991), California and Pennsylvania (Newhill, 1996; Newhill & Wexler, 1997 ), Montana (Horejsic & Garthwait, 1994), a western state (Rey, 1996), an unidentified state (Schultz, 1987; Schultz, 1989), Santa Fe Springs, California (Castellanos, 1998), N ew England (Horwitz, 2006) and two in Los Angeles (Seeck, 1998; Vergara, 2006). One of the earliest studies to document the proble m of violence towards social workers in the U.S. was a study of 150 randomly sel ected social service workers in an unidentified state (Schultz, 1987). Interestingly, several questionnaires that discussed some blatant forms of violence such as attempted ho micide and forcible rape were not included because they occurred after office hours. The study showed that 65% of the respondents had encountered violence in the work se tting. This study also documented verbal threats to harm or kill the social worker in every social service setting that was studied. Rey (1996) completed a study in a western state us ing both a purposive sample of 150 agency directors and a random sample of 300 lic ensed social workers from an undisclosed source. A total of 175 social workers r esponded. More than 80% of the directors and social workers felt that safety issue s in social work were an increasing concern. A majority (89%) of the respondents had be en verbally abused by a client and 23% had been physically assaulted during their care ers.
20 In the largest regional study, Newhill (1996) found that 78% of the 1,129 social workers from California and Pennsylvania agreed tha t client violence toward social workers is a significant issue and 52% said that th ey sometimes worried about their own safety. Fifty-seven percent had experienced at leas t one type of client violence. In Wrenns (2005) study of 600 NASW members from th e state of Illinois she addressed the relationship between personal trauma exposure and secondary traumatic stress for social workers. One of the specific area s of inquiry included social workers exposure to occupational violence, particularly cli ent violence. Twenty two percent had been assaulted by a client in their career, with 6% experiencing assault within the past year. Forty percent of the respondents had been th reatened during their career and 10% had been threatened within the past year. The major ity (74%) of the participants had experienced verbal abuse during their career, with 40% experiencing verbal abuse within the past year. Twenty nine percent had been threate ned with a lawsuit during their career. Social workers with recent exposures to lawsuit thr eats and verbal abuse had higher levels of secondary traumatic stress. Overall, wor kers with exposure to direct workplace stress tended to score higher on secondary traumati c stress. Qualitative responses indicated that positive co-worker and administrativ e support helped to mitigate the effects of workplace stress and trauma. This study was one of only two social work studies that empirically examined both direct and indirect traum a effects on workers exposed to client violence. Client violence in specific practice settings in th e U.S. Some client violence studies in the U.S. have focused on social workers in particular practice settings. Because
21 these studies focus on particular groups of social workers, they cannot necessarily be generalized to the entire population of American so cial workers. The most studied practitioners have been child protection workers (C astellanos, 1998; Horejsic & Garthwait, 1994; Horwitz, 2006; Newhill & Wexler, 1 997; Song, 2005). Newhill and Wexler (1996) found that child and youth social wor kers were significantly more likely to consider violence toward social workers to be a significant issue for the profession in general and in their practice specifically and they were more likely to worry about their own safety while working with clients. Forty one pe rcent of this sample had experience at least one type of violence. Song (2005) reported a slightly lower percent (31%) of child and family workers who had experienced client viole nce in the past year. These percentages greatly contrast with the 97% of Horejs ic and Garthwaits (1994) 166 childrens services workers who had been screamed a t or cursed at in the preceding 12 months. Twenty six percent of those verbally abused said that such abuse happens at least once per week, with one third of the workers having been threatened with death. These findings resonate more with Canadian child welfare workers, 95% of whom had been verbally assaulted (Snow, 1994). A notable child and youth social worker study was a study that focused on prevalence of client violence toward child and fami ly social workers and its effects on burnout, organizational commitment, and turnover in tention (Song, 2005). This study has been the only study to take into account the mediat ing psychological stress response to workplace violence, rather than focusing strictly o n direct associations between exposure to client violence and negative effects. This study was also first social work study to test
22 a proposed structural model by using Structural Equ ation Modeling (SEM) techniques to address the concept of client violence and to exami ne an association between client violence and organizational commitment. The SEM fou nd that past victimization had a direct and positive effect on fear of future victim ization. Fear of future victimization then had a direct and negative effect on affective socia l workers commitment to their organization. Additionally, when the social workers scored higher on burnout scales, they were more likely to have intentions to leave their jobs. However, when they had stronger commitment due their identification with the organi zation (affective organizational commitment) they were less inclined to contemplate leaving the organization (turnover intentions). Exposure to client violence had a dire ct positive effect on turnover intention. School social workers are the only other social wo rkers from a specific practice setting that have been studied concerning exposure to client violence (Astor et al., 1998). Researchers found that one third of the 576 school social workers feared for their safety, with a significantly higher percentage (71%) of inn er city social workers having fear for safety. Almost half of the workers in inner city sc hools who reported being fearful also thought of leaving the profession of social work. T hirty five percent of the respondents had been physically assaulted within the past year. Methods of sampling in U.S. studies. Most of the social work client violence studies in the U.S. have drawn at least part of the samples randomly. Two exceptions to this were Atkinsons (1991) and Vergaras (2006) qu alitative studies. Atkinsons study involved semi-structured open-ended personal interv iews with eight clinical workers who had been assaulted by a client within the past year Though only three were social
23 workers, this study broadened knowledge about clien t violence by talking about worker response to violence. Fear and anxiety, particularl y about being in situations that were reminiscent of the assault, were prevalent in all o f the workers, yet these fears were contained in the workplace. Workers generally looke d for support from their colleagues and did not discuss their fears at home. Making sen se of the assaults helped to ameliorate the negative effects from the assault. The majority of workers felt that assault is always possible and they had increased awareness of the ne ed for precautions. Vergaras (2006) study included 15 social workers i n Los Angeles County who were purposefully sought from the researchers pers onal social network. The focus of the study was on violence experienced by workers who re gularly made home visits. The participants had been doing home visits for an aver age of 12.5 years. The majority (86.7%) of the sample were workers of a minority ra ce or ethnicity. None of the social workers were under 30 years of age. Vergara found that the majority of her sample had experienced client violence while completing home v isits. Most had reported the violence to supervisors, though they had not officially docu mented it. They were overall pleased with the support of their supervisors and organizat ions. Like Atkinsons (1991) sample, many of these workers were cautious about discussin g their fears about job safety with family and friends. This is one of the few studies that included data from a large percentage of minority social workers and the only client violence study to date to focus exclusively on the practice of home visiting. Two other U.S. studies have been non-probability st udies that cannot be generalized to the larger population of U.S. social workers. The first was Horwitzs
24 (2006) study involving a convenience sample of 273 social workers who attended training seminars that he was conducting. This stud ys focus was on the trauma effects of negative workplace events, including being directly or indirectly exposed to client violence. The second non-probability study was Castellanos (1998) study of 31 workers from five Department of Children and Family Service s offices. Its focus was on social workers perceptions about working in dangerous nei ghborhoods. The sample size of this study was very small; however, the study merits inc lude the unusually high percentage of ethnic and racial minorities and the percentage of workers with few years of experience. The majority (76%) of the workers had been at the a gency less than three years and 71% were minority social workers. Sixty one percent of the workers had been involved in a threatening situation during a home visit and almos t 84% knew another worker who had been involved in a threatening situation. In spite of the fact that many had been threatened, at least half (51.6%) said that they ha d no problem working in an innercity area. Of client violence studies that have drawn samples randomly, the majority have selected social workers from the national NASW memb ership (Astor, Behre, Wallace, & Fravill, 1998; Beaver, 1998; Guterman, Jayarante, & Bargel, 1996; Jayarante, VinokurKaplan, Nagda, & Chess, 1995; Jayarante, Croxton, & Mathison, 2004; Ringstad, 2005, Song, 2005) and state NASW memberships (Newhill, 19 97; Newhill & Wexler, 1997, Wrenn, 2005). These studies have yielded usable res ponses from between 269 social workers (Song, 2005) and 1129 workers (Newhill, 199 6), thus the studies are
25 strengthened by both probability sampling methods a nd larger sample sizes. Other variations of probability sampling have been rando m draws of 150 social service workers from an undisclosed source (Schultz, 1987) and of 300 licensed social workers from an undisclosed source (Rey, 1996). There have been five studies that have used probabi lity sampling of large numbers of social workers in all practice arenas throughout the United States (Beaver, 1999; Jayarante, Vinokur-Kaplan, Nagda, & Chess, 1995; Ja yarante, Croxton, & Mattison, 2004; Guterman, Jayarante, & Bargal, 1996; Ringstad 2005). Each of these studies had added strength in that they sought more information than the prevalence of violence information gathered in many of the other social wo rk client violence studies. Additionally four of the five studies asked questio ns about client violence in similar ways, thus adding the possibility of comparing and contrasting the findings. Only Ringstads (2005) national study used different que stions concerning exposure to client violence. Jayarante and his first set of colleagues (1995) st udied 633 social workers randomly drawn from the 1993 NASW membership. Stude nts, retired and unemployed workers were excluded from the study. The study co ntained questions about exposure to client violence as well as items about job stress, psychosocial strain, and social support. Respondents were asked whether they had directly en countered various types of violence and if a coworker in their agency had encountered t hose types of violence within the past year. Seventeen percent of the sample reported bein g physically threatened by clients and a small number (2.8%) reported being physically ass aulted. Forty one percent said that
26 they knew of a coworker who had been physically thr eatened. Almost half of the respondents had encountered verbal abuse and 53% kn ew of verbal abuse to coworkers. Workers who had been threatened reported significan tly higher levels of irritability, depression, anxiety, and burnout. These researchers found that workers who had experienced client violence were no more likely to make an effort to find work with another employer than workers who had not experienc ed client violence. Researchers used logistic regression to analyze the effect of a ge, gender and experience on experiencing client violence. They also did bivaria te comparison of individuals who experienced abuse and threats in the workplace with individuals who had not had such experiences. There was no analysis of the effect of ethnicity on experiences of violence. Some preliminary associations might have been valua ble, considering the fact that there were about 60 minority respondents. In a study similar to the 1995 study that Jayarante coauthored, he and colleagues Croxton, and Mathison (2004) received surveys from 507 NASW members. Respondents were asked if they had experienced different types of client violence directly or if they had known of a co-worker being the victim of the sa me list of violent acts. This study did a multistage random sampling process in which they first drew a sample of African American, Asian American, Hispanic, and White socia l workers who had a MSW and were in direct practice. The second stage of sampli ng drew a sample of NASW members who identified their primary practice setting as pr ivate practice, thus allowing the researchers to examine private practitioners separa tely. Researchers found that verbal abuse was pervasive in all settings that they studi ed. Additionally, 22.8% of the
27 respondents had been physically threatened by a cli ent. Institutional mental health and child protective service settings were noted to be the most dangerous settings for the social workers they questioned. There was a signifi cant difference in the numbers of private practitioners who had been threatened with violence by clients (9%) versus agency practitioners who had been threatened (22.8% ). Similarly, only 1.6% of private practitioners knew a colleague that had been physic ally assaulted, compared to 24.6% of agency practitioners. A particular strength of this study was that more data was collected from different race groups, thus allowing for bette r analysis of the effects of ethnicity and race on exposure to client violence. A doctoral dissertation explored the extent of clie nt violence against social workers and how client violence impacted the functi oning and well-being of the social workers, using a random sample of 1500 national NA SW members (Beaver, 1999). With a 62.8% return rate, 942 social workers responded. The mean age of the sample was 48.2. Only 8.4% of the sample was aged 24-34, with no soc ial worker under the age of 24. Ninety three percent had a MSW. Only 2% had a BSW. Seventy five percent had more than 10 years experience and only one participant h ad less than 2 years experience. Sixty five percent of the social workers had experienced some type of client violence in their careers, with verbal abuse being the most frequent type of violence. The annual prevalence rate for experiencing any type of violen ce was 23.4%. Respondents reported a total of 1,227 violent incidents in the past year. Seventeen percent of those incidents were physical assaults. Beaver found that social worker s who had experienced client violence were significantly more likely to be dissatisfied w ith their jobs. Additionally, the study
28 analyzed the workers level of burnout using Maslac hs three subscales of emotional exhaustion, depersonalization, and personal accompl ishment. Beaver found that there was a significant positive correlation between emotiona l exhaustion scores and violence scores, but the correlations for depersonalization and personal accomplishment were not significant. General health of the workers was also measured and there was no significant correlation between levels of client violence and l evels of general health. When all of the factors were entered into a multiple regression tes t, age was the most significant predictor of client violence. In 2005, Ringstad investigated the aggressive and a ssaultive interchanges between social workers and their clients, using responses f rom a national random sample of NASW members. She had a 34% response rate, receivin g questionnaires from 1,029 social workers. Ringstad is the only social work re searcher to use a standardized instrument, subscales from the Revised Conflict Tac tics Scale, to measure incidents of physical and psychological aggression. Respondents indicated which types of incidents they had experienced in their career and within the past year. Additionally they were asked about being both the victim and the perpetrat or in the violent incidents. The study found that 86% of the respondents had experienced s ome type of client violence during their career. Psychological aggression, including t hreats, verbal abuse and property destruction, was the most common type of client vio lence, with 1,029 respondents experiencing 8,113 incidents in the past 12 months. Thirty percent of the respondents had experienced physical assault in their career and al most 15% had experienced physical assault (695 actual incidents) in the past year. Ri ngstad emphasized that the
29 overrepresentation of experienced workers might hav e inflated the career rates of client violence, as it might be surmised that the longer a person is in practice, the more violence she might suffer. One cross-national study did a comparison of experi ence of client violence between social workers in the United States and soc ial workers in Israel (Guterman, Jayarante, & Bargal, 1996). The study included a ra ndom sample of 1200 NASW members in the U.S. and 1,497 members of the Israel Association of Social Workers. Response rates were comparable, with 62.4% of the U .S. workers responding and 60.1% of the Israeli workers responding. The survey was s imilar to the one that Jayarante had used in his two previous U.S. studies, though the I sraeli version of the survey was translated into Hebrew. The questions about workpla ce violence were imbedded in a larger questionnaire about workplace conditions. E ighty-eight percent of the U.S. workers held a MSW, whereas the majority (77%) of t he Israeli sample held a BSW. Another major difference in the respondents was tha t a majority (68%) of the U.S. workers worked in private settings, whereas only 8% of the Israeli workers worked in private settings. The researchers used both chi-squ are to examine differences in patterns of difference cross-nationally and logistic regress ion to examine predictive differences cross-nationally among six types of violence. The s tudy found that 48.8% of Americans and 47.4% of Israelis had experienced at least one form of victimization from clients within the past year. Both groups experienced compa rable levels of verbal abuse and physical abuse. Americans experienced more physical threats. Of interest to this study is the fact that in the U.S. less experienced workers had higher levels of physical threats,
30 physical assaults, and verbal abuse within the past year. However, in the Israeli sample, increased experience did not decrease exposure to c lient violence. Merits of this study are that it is the only cross-national that has been co mpleted on this subject and it uses a substantial random sample of social workers in the U.S. However, the survey excluded social workers who were not working directly with c lients, so it is possible that student members were not included. If responses from stude nts were not included, the researchers may have lost an opportunity to gain in formation from prospective social workers who are regularly interning in agencies. Summary of U.S. social work client violence studies There have been more than 20 client violence studies conducted in the United States within the past 30 years. Almost half of those studies sampled respondents in a limi ted geographic area. Six studies (including 4 from limited geographic areas) address ed only social workers from particular practice areas. Five national probabilit y studies with large sample sizes may be the best studies from which to generalize the clien t violence experiences of social workers throughout the U.S. A major limitation with many of the studies on soc ial workers in the U.S. is that the samples are often cited as having a combination of the following characteristics: female, White, MSW, a mean age in the mid 40s to e arly 50s, and many years of experience (Astor et al., 1998; Beaver, 1998; Horej sic & Garthwait, 1994; Newhill, 1996; Newhill & Wexler, 1997; Jayarante et al., 1995; Rin gstad, 2005; Song, 2005; Wrenn, 2005). In short, the studies that used a random dra w from the NASW membership may have results that are generalizable to the populati on of social workers in the U.S., but it is
31 unknown if these results are informative concerning the client violence experience of social work students. Violence in Social Work Education Student Perspec tive Studies of social work students have begun to show that field students are falling prey to violence while in the course of completing their practicum experience. In an early review of the literature, Star (1984) included info rmation from personal communication with Hawthorne about an informal survey conducted a mong first and second year graduate social work students attending the Univers ity of Southern California. Results of this survey indicated that client violence was one of the three most prevalent issues confronting students when they were beginning field placements. Regardless of agency setting, every student in this survey had worked wi th one overtly or potentially explosive client. This study was not published and it was tr ansmitted via conversation, making it vulnerable to inaccuracies in the data. However, it s merit is that is the first time that client violence toward social work students is specificall y discussed in the literature. Tully, Kropf, and Price (1993) completed a study of social work students at the University of Georgia. There were 49 BSW students a nd 72 MSW students. They reported that during the practicum, 26% of students had been verbally abused or threatened. Thirteen percent were threatened physic ally and 13% felt unsafe or threatened during their practicum experience. An additional 25 % had known or seen violence toward other personnel in their practicum site. Thi s study indicated that there was no significant difference in exposure to violence base d on practicum site. The students in this study had a 100% response rate to the question naire, likely because the surveys were
32 handed to the students by faculty. Field educators also received mailed surveys and 96 (74%) responded. Sixty two percent stated that they had been verbally abused by a client during their career and 24% had been physically ass aulted by a client. Mamas (2001) study of 37 senior BSW students in New Jersey revealed that 54% had been affected by some type of violence. Fif teen percent of those students reported experiencing physical assault by a client, while 49% had experienced verbal abuse in the form of verbal attack, cursing, or sex ual advances. Additionally, 40% of the students encountered threatening behavior. Mama use d the same survey that was earlier used by Tully et al (1993). Seventeen field educato rs also completed the survey. Eighty five percent had experienced verbal abuse by a clie nt in their career. Twenty nine percent had been physically assaulted in the form of being slapped, punched or robbed. Mamas sample is a very small convenience sample, which pr ohibits any generalization from the findings. The statistics reported were frequencies which may have been suitable for that phase of research on client violence of social work students, but they lack the stability of more rigorous analysis. Knight (1999) studied 110 BSW students from the Uni versity of Maryland. She did a test prior to the beginning of placement and another toward the end of the placement. The scores from the two tests could only be aggregated because individual students were not matched on their pretest and post test. At the beginning of placement, 37% of the students felt that there might be a risk to their personal safety while working with clients. At the end of placement 17% of the st udents reported that they actually felt that their personal safety was at risk with their c lients. This reduction in percentage was
33 in spite of the fact that during the course of the practicum, 33% of the students were verbally threatened and 19% were physically threate ned. Knight surmised that actual experiences with clients might have served to reduc e fears for the students safety. Students in the study who did not like the way that the school responded to their encounter with violence had a higher tendency to re think their career choice and to believe that social work is a dangerous profession. This study involved students from only one institution and the students were predomin antly from the region where the school is located. Analyses were limited to descrip tive data and one series of t tests. In 1998, Lyter and Abbott (as cited in Lyter and Ab bott, 2007) did an anonymous safety survey with 39 students who were in a social work practicum. Two thirds of the students were fearful of clients. A third (34%) had been threatened by a client. Three (9%) had been physically assaulted, though only one had reported this to the field director. The other students said that compassion a nd commitment kept them from reporting the incidents. Nothing is known about the methodology of this study other than the fact that it was anonymous. The small sample si ze prohibits generalizing the results. The largest client violence study of social work st udents to date was completed by Knight (1996). She surveyed 380 BSW students from 8 schools in Maryland, Virginia and the Washington D.C. area and had a response rat e of 258 (67.9%). in this study 20.9% reported at least one incident in which they had been verbally or physically assaulted. Of those, 83% felt that their encounter with danger made them more cautious in practice. Eighty-five percent of all of the stud ents in the study intended to pursue an MSW within three years. These students were signifi cantly more likely to view social
34 work as a dangerous profession. Additionally, they were more likely to be uncomfortable engaging in traditional social work practice activi ties such as making home visits, working evening hours, working directly with groups and families, and visiting resources in the community. They were also more likely to des ire to go into private practice. Knight interpreted these findings to mean that stud ents with the more advanced skills may be the least likely to engage in the most criti cal social work activities. Knight stated that this was an exploratory study in which certain statistical analyses, such as correlation matrixes, could not be completed due to the small n umber in the sample. Client violence toward 36 MSW students was the subj ect of a masters thesis at California State University (Schwarzmueller, 1998). This rather small response rate (25.7%) was obtained via a recruitment letter in st udent mailboxes. This study used the same survey that Knight (1999) had developed. Almos t half (48.6%) said that they worked with potentially violent clients at least so me of the time. The more familiar the students were with the community where they worked the less concerned they were for their safety. Analysis included Pearsons correlati ons and some t -tests. The researcher acknowledged that due to the homogenous sample it w as not possible to complete any tests using demographic variables of age and gender Additionally, she acknowledged giving the survey when students were busy trying to meet thesis deadlines. One student wrote in a comment that some of the questions on th e survey were confusing. Only one study has been done comparing awareness an d fear of violence of social work students versus students from another professi on. Elwood and Rey (1996) compared the violence experiences and training of s ocial work and medical students,
35 comparing a sample of 78 social work students and 4 6 medical students from the University of Nevada. Seventy five percent of the m edical students had heard reports of violence from peers or had observed violence in med ical settings and 8.6% of them had been physically assaulted by patients. Less social work students (43.6%) were aware of other social workers who were victims of client vio lence, but a similar number (6.4%) had been directly personally assaulted. Many (41%) of the social work students were fearful of client violence by their clients. These researchers acknowledged a possible sample bias due to the differential manner in which the data was gathered. The medical students had surveys mailed to their homes, where-a s the social work students received the surveys in field classes. For the social work s tudents, the surveys were completed in the presence of the researcher; thus, they had the benefit of being able to clarify questions or concerns about items on the survey. Another limi tation of the study was that the sample sizes were small and the study was conducted in one university. Merits of the study include its comparison groups. The only international study of social work student s occurred in Australia (Maidment, 2003). There were 39 participants in th eir third and fourth year of their BSW program. Data collection consisted of a mailed ques tionnaire sent at the end of the field placement. Statistics reported were predominantly f requencies, with one chi square table. Maidment found that 31% of the students had been ve rbally abused by clients and 2.5% had been physically abused. Again, in this study, t here is small convenience sample and findings cannot be generalized to other social work programs, particularly social work programs in the United States that do not have comp arable requirements for field.
36 Violence in Social Work Education-Field Director Pe rspective There have been four studies of field directors pe rceptions of client violence toward their students. Ellison (1996) sent a questi onnaire to 200 randomly selected CSWE accredited schools and received response from 147 schools, including 96 BSW programs and 51 MSW programs. It found that incide nts had occurred to students in 23 of the schools. Seven incidents involved students w ho required medical or police intervention as a result of client violence. These schools reported incidents of violence occurring to77 students. The most severe incident i nvolved a student being held hostage by a client. Though Ellison asked social work schoo ls to send copies of their safety policies, only 5 (2.5%) complied with the request. Of these five schools, only one had a published mechanism for reporting client violence. Ellison acknowledged that it was possible that many incidents could have gone unrepo rted, especially if the incident did not result in injury. Since this study did not inc lude contact with students, it is difficult to tell if the reported incidents are just the tip of the iceberg (p. 87). The same year that Ellisons study was published, a qualitative study was published by Wayne and Raskin (1996). This study in cluded 22 structured phone interviews with field directors in cities identifie d as having highest rates of violent crime. Ten (45%) of the directors reported that they knew of social work students who were victims of crime such as muggings. Ninety five perc ent said that they had no formal policies concerning students in high-crime areas, t hough many of them said that they were giving the issue serious thought. Several repo rted that they left it up to the field agencies to establish formal policies. Nearly all o f the respondents said that personal
37 safety was a concern for at least some of their stu dents, with 31.5% saying that it was a concern for most or all of their students. A questi on was asked about how the field directors responded to students who turned down pla cements because of the students fear of the neighborhood. Three (14%) said that the y would stand firm on their recommendation, since social work students are expe cted to work in such areas. An additional 18% of the directors said that they woul d replace the students, but they would question the students fit with the profession of s ocial work. While most of the respondents stated that their school environments e ncouraged the freedom of students and faculty to express concerns about personal safety, two said that some faculty members were embarrassed to express their fears. They furth er stated that they believed that some faculty perceives expression of fears about persona l safety as a lack of commitment to the profession. This study has limited generalizability as it includes only a small number of field directors and they represent only those socia l work schools in high-crime cities, yet it was useful in beginning to recognize the issue o f student safety, especially as it pertains to communities that may be more dangerous. Reeser and Wertkin (2001) sent questionnaires to al l 418 social work schools that were CSWE accredited. They received response from 2 58 schools. They found that 42% of the programs had at least one student to be thre atened by a client in the preceding two years and 13% had a student to experience physical assault. A large majority (86%) of the directors said that more attention needs to be give n to the issue of student safety in the field. This was a very thorough study that was larg ely focused on safety training needs of students and how these are met in social work progr ams.
38 A more recent random survey of 200 CSWE accredited BSW programs had only nineteen schools (10.5% return rate) to reply (Fari a & Kendra, 2007). Most of the schools that replied had programs with fewer than 100 stude nts, had only BSW programs and were at private institutions. Five of the programs (26%) stated that within the past five years, there had been incidents of verbal or physic al violence directed towards a student. The researchers acknowledged that the small respons e rate made it difficult to generalize the findings to all BSW programs. Characteristics of Violence towards Social Worker Types of Violence Direct exposure to client violence. All empirical studies on client violence towards social workers or social work students have documen ted direct exposure to client violence. Many studies document the prevalence of p hysical and verbal assaults. Physical violence may include grabbing, slapping, kicking, b eing hit with fists, and being pushed to the floor (Schultz, 1987). Psychological aggress ion may include the following acts by clients: insulting or swearing, shouting or yelling stomping away during disagreement, saying or doing something to spite, threatening to hit or throw something, destroying something that belonged to the social worker, or ca lling the worker fat or ugly. Ringstad (2005) found that 62% of her sample of more than 10 00 NASW members had experienced psychological aggression in the past 12 months. They reported 8,113 acts of psychological aggression in the past 12 months, com pared with 695 acts of physical aggression in the same time period. A study of juv enile probation officers, some of whom were social workers, found that verbal or psyc hological violence was much more
39 prevalent than physical violence (Rapp-Paglicci, 20 04). In a qualitative study on this subject, Littlechild (2005) did a content analysis of open ended questions on 21 questionnaires completed by child protection social workers in England and Finland and he completed seven in depth interviews with workers who had experienced violence in the field. He found that physical violence was rela tively rare, but that indirect violence (p.67) was common, less obvious, and more pervasiv e and insidious (p. 67) than physical aggression. In this case, Littlechild was referring to violence that occurred directly to the workers, but that did not consist o f overt physical actions. For example, threats of further actions from clients towards the social workers or their families seemed to have the greatest negative impact on the social workers. It was stated that the term incident was inadequate to describe some of the o ngoing experiences of workers. He suggested using the term developing violent scenar ios, which doesnt imply one incident, but an environment where threats are mad e and actions taken to frighten and disempower the worker (p.67). Indirect exposure to client violence. Social workers may be negatively affected vicariously by hearing of colleagues being harmed b y clients or threatened by clients. in a study of work related trauma effects in 273 child p rotection social workers, Horwitz (2006) found that vicarious events were more highly associated with trauma effects than were direct events. He studied the trauma effects o f negative workplace events, including being directly or indirectly exposed to physical as sault, property damage, or being placed in fear of safety by a client. He found verbal abus e and threats were significantly positively associated with workplace trauma effects He speculated that vicarious events
40 may be more toxic (p.14) because they reflect the social workers lack of control. Rapp-Paglicci (2004) found that probation officers who had been personally victimized had less concern about their safety than those who had not been victimized. She speculated that those who had not been victimized m ight have increased concern due to hearing about colleagues who had been victimized. Personal/Individual Factors Influencing Client Viol ence Gender. Several studies have indicated that males are at hi gher risk of client violence than females. National Institute of Social Work Workforce Studies in England found that 48% of male social workers had experienc ed violence versus 29% of females (McLean, 2000, as cited in Brockmann, 2002). In the U.S., a regional study of workers from California and Pennsylvania found that males w ere more likely to encounter client violence, though they were also more likely to be p racticing in areas where more client violence occurred (Newhill, 1996). A national rando m sample of social workers from the U.S. found that males experienced more physical thr eats and assaults than female social workers (Jayarante et al., 1995). Other national st udies of NASW members have similarly reported that male social workers were si gnificantly more likely than female social workers to be victims of client violence (Ja yarante, Croxton, & Mattison, 2004; Ringstad, 2005). Similarly, Beaver (1999) found mal es to experience client violence more often, but the results were not statistically significant. Two studies had contradictory findings. Song (2005) found mixed results, reporting that males had higher frequencies of phys ical assault, threat of physical harm, property damage and physical assault or threat agai nst family members, but females
41 experienced more verbal abuse. In Songs study, nei ther t tests nor chi square analyses indicated significant statistical differences betwe en male and female exposure to violence. Guterman, Jayarante, and Bargel (1996) fo und that male workers in the U. S. had higher levels of client victimization via physi cal threats and assaults, but this was not true for the Israeli workers also surveyed in this study. Age. Younger workers in agency settings have been found to be at greater risk of physical threat, threat of law suit, verbal abuse a nd sexual harassment than older colleagues (Jayarante et al., 2004). Additionally, it was found that younger workers have more fear of verbal abuse and physical assault, reg ardless of whether they were agency practitioners or private practitioners. Similarly, Jayarante and colleagues (1995) reported that younger workers experienced significantly more verbal abuse than older workers. In that study the younger workers were also significan tly more likely to know co-workers who had experienced verbal abuse and were also more likely to report knowing coworkers who had experienced physical threats and th reats of lawsuits. In his national study Beaver (1999) also found a significant negati ve correlation between the age of the social worker and experience with client violence. Likewise, Songs (2005) national study of child welfare social workers found a signi ficant negative correlation between age and exposure to client violence, indicating that yo unger workers experience more client violence. Another U.S. study indicated that social workers age was significantly statistically related to experience of physical vio lence, but the researcher stated that there was not a substantive difference between the mean a ge of those physically assaulted ( M = 53.4) and the mean age of non-assaulted social w orkers ( M = 55) (Ringstad, 2005).
42 Guterman and colleagues (1996) found that younger s ocial workers in the U.S. experienced more exposure to client violence, but t his wasnt the case for Israeli social workers. However, this finding was intermingled wit h a discussion on worker experience and it is not clear if the researchers were referri ng solely to age as a variable. It may be that these researchers implied that less experience d workers were younger. One social work student study found that younger st udents were somewhat more likely to express discomfort working in the inner c ity and they were more likely to view social work as a dangerous profession (Knight, 1996 ). In this study only BSW students were questioned, but the mean age was 29, indicatin g a mixture of traditional and nontraditional aged students. Experience. Less experienced workers may be the recipients of v iolent acts more than experienced workers. In one early study of psy chotherapists it was found that workers with 11 years or more experience were assau lted or threatened at a ratio of 1: 4 compared to those with less than 11 years experienc e (Bernstein, 1981). A more recent study similarly indicated that clinicians in a psyc hiatric hospital who had more years of mental health experience had statistically fewer in cidents of violent episodes (Privitera et al., 2005). Beavers (1998) national study indicate d that the least experienced social workers had the highest annual rate of client viole nce. The workers with the most years experience had the lowest mean of incidents. An ANO VA test indicated that experience was significantly related to client violence experi ence. Gutterman and colleagues (1996) also found that less experienced workers in the U.S had more exposure to client violence, but this wasnt the case for workers in I srael. The researchers stated that this
43 might be due to differences in the workplace for Is raeli and American social workers. It was explained that Israeli social work agencies are hierarchically flatter (p. 184); thus, social workers in those agencies might not have as many opportunities to advance in the workplace. The researchers surmised that as America n social workers advance in years in their agencies, they may have the latitude to impro ve their working conditions and to have greater control over the clientele with whom t hey will work. Social work student studies have found conflicting evidence regarding how experience affects exposure to client violence. In Knights (1996) study of BSW students in eight schools, almost two thirds of all the stud ents felt that personal safety would play at least a limited role in their career decisions; however, those with more practice experience were significantly less likely to feel t hat personal safety would be a factor in their career decisions. These feelings could have b een attributed to another finding that those with more experience were significantly less likely to believe that social work was a dangerous profession. The only U. S. study to show that more experienced workers suffered more client violence than less experienced social workers was a study of MSW and BSW social work students and field educators done at the University of Georgia (Tully et al., 1993). These researchers that professionally seasoned MSWs (p. 197) experienced more client violence than less experienced student clinicians (p. 197). Though this statement is not well explained in the study, it is presumed that th e researchers are referencing the experience level of students versus the experience level of the field instructors who were
44 also questioned. Tully and colleagues speculated th at less experienced students might be protected from violence by their status as students Ethnicity. There has been little research on how ethnic minori ties are affected by client violence. There is some evidence that worke rs of minority race/ethnicity in social service settings in the United Kingdom may be dispr oportionately affected by client violence because of being in lower status jobs and having more likelihood of being in residential work (Butt, 2000, as cited in Brockmann 2002). It should be noted that numerous social work studies on client violence in the U.K. refer to residential social workers, versus field social workers, and day care social workers (Balloch et al., 1999; Brown, Bute, & Ford, 1986; Norris, 1990; Rowett, 19 86). It is implied that residential social workers are those who work consistently with clients in residence. Another U.K. study found that staff from ethnic minorities seeme d to be concerned about receiving unsympathetic responses from management and/or coll eagues if they were victims of client violence (Rowett, 1986). Thus, they may not report client violence when it has occurred. Budd, Arvey, and Lawless (1996) found that non-Whit e participants worry more about future exposure to workplace violence even th ough they experienced workplace violence less than White participants. With some su pport and some dispute of this finding, a study where Jayarante, Croxton, and Matt ison (2004) took precautions to over sample minority social workers found no significant difference in the experiences of client violence between minority practitioners and White practitioners. However, African American social workers were significantly less lik ely to have fear of almost every type
45 of client violence that they studied. In their stud y of school social workers, Astor and colleagues (1998) found that members of ethnic mino rities were significantly more likely than White social workers to view the community aro und the school as dangerous. However, the researchers viewed this finding with c aution, because they felt it might stem from the fact that a disproportionate percenta ge (35%) of inner city school social workers was people of color. One epidemiology summa ry of several U. S. national workplace violence reports stated that in all the o ccupations studied, African Americans and Hispanics had higher frequencies and rates of h omicide in the workplace (Kraus, 1996). However, it is not known if this finding app lies to African Americans and Hispanics in social work agencies. Educational level. Most social work client violence studies have not r eported the educational level of the social workers. However, B eaver (1999) found that bachelors level social workers reported client violence durin g the past year at a rate that was almost twice the annual rate for the total sample, indicat ing that education level may have significant influence on experience with client vio lence. Likewise, Privitara, Weisman, Cerulli, Tu, and Groman (2005) also found that BSW s were endangered, threatened, and assaulted more than MSWs. In contrast, in their s tudy of social work students, Tully and colleagues (1993) found that there was no statistic ally significant difference between MSW and BSW students on exposure to verbal abuse an d threats of harm. The same researchers emphasized that violence is a part of t he field placement experience for both MSWs and BSWs.
46 Organizational/Practicum Related Factors Influencin g Client Violence Place of client violence. Only three social work studies have actually compar ed the experience of client violence in the office ver sus out of the office. One study found that probation officers were verbally threatened an d physically intimidated more often in the office than in the field (Rapp-Paglicci, 2004). There had been some incidents of being held hostage in both the office and the field. In a much larger random study of social workers in Pennsylvania and California, Newhill and Wexler (1997) similarly found that the place of violence was dependent on the type of violence. The home of the client was the place that child and youth workers were threate ned most often, whereas property damage occurred most often in the workers office. Actual attacks of child and youth workers happened in a wide array of locations. Two student studies found that the majority of client violence incidents (70% and 78%, respectively) occurred at the agency versus the home (Mama, 2001; Tully et al., 1993). Several social work studies have indicated that wor kers are at high risk for client violence while making home visits. A national study of school social workers found that 94% of the respondents saw home visits as an effect ive intervention; however, 74% thought that home visits were potentially dangerous (Astor, Behre, Wallace, & Fravil, 1998). In spite of this concern, 82% of these schoo l social workers continued to make home visits to aggressive children. Two studies with small sample sizes found that larg e percentages of social workers had experienced client violence while on ho me visits. One study of 31 child welfare workers found that 61.3% had experienced a threatening experience while on a
47 home visit and 94% knew a co-worker who had been in volved in a threatening situation during a home visit (Castellanos, 1998). In a quali tative study of 15 home visiting social workers in Los Angeles County, Vergara (2006) found that 93% of the participants said that they had experienced one or more types of viol ence, and none of them felt completely safe. Most (92.9%) had reported incident s of violence to their supervisor and 85.7% were satisfied with their supervisor/s respon se. A similar percentage (93.3%) was satisfied with the safety policies of the agency. In a study of in-home nurses, social workers, chil d management specialists, and behavior management specialists in Canada, 32.8% st ated that an in-home client had been verbally aggressive with them within the last six m onths (Barling, Rogers, & Kelloway, 2001). Some workers (4.5%) had been threatened wit h a gun and 3.9% had a client to try to hit them within the past six months while they w ere in the clients homes. One student study found that 48% of MSW students in dicate that they were concerned about their personal safety at least some times when making home visits (Schwarzmueller, 1998). Of this same group 30.1% ac tually encountered verbal threats during home visits during the practicum and 44.7% w ere physically threatened during home visits. Time of client violence. Few studies have inquired about the time of day tha t client violence took place. Working in evening hou rs has been shown to be a factor that increases anxiety about the possibility of workplac e violence. Knight (1999) reported that 39.1% of social work students who were questioned a t the end of their practicum stated that they were concerned about their personal safet y when working during evening hours.
48 Another study of MSW students indicated that at the beginning of their internship, 51.7% were at least sometimes concerned about their perso nal safety while working evening hours (Schwarzmueller, 1998). During their practic um, 15.6% actually encountered verbal threats during evening hours work and 33.3% were physically threatened during evening hours. An analysis of seven national studie s indicated that occurrence of homicide and nonfatal assault is more frequent in t he afternoon and evening hours than late morning or early afternoon hours (Kraus, 1996) Client violence per practice setting. Violence occurs to social workers in many practice settings, though some are more dangerous t han others. Child protection social work has been noted to be particularly hazardous (J ayarante et al., 2004; Newhill, 1996; Schultz, 1987). A national survey of 13,380 child w elfare workers in 10 states found that over 70% had been victims of violence or threats of violence (AFCME, 2007). A very small sample ( N = 31) of children and families workers in Californ ia similarly found that 77.4% had been verbally assaulted and 6.5% had been physically assaulted during their career (Castellanos, 1998). In Snows (1994) qualit ative study of 20 children and youth workers in Ontario, all said that they had been kic ked, spit upon, and hit by clients using weapons of various sorts. Ninety five percent had b een verbally assaulted, with 60% stating that this was a daily occurrence. Almost al l (90%) had been hit with an open hand and 75% had been hit with closed fist. Social workers in residential settings may be expos ed to a higher level of physical violence than other social workers (Brockmann, 2002 ). Leadbetter (2003) noted that childrens homes were emerging as the most violence prone setting. His study found that
49 social workers in three adolescent residential unit s in Scotland encountered 14 acts of moderate physical aggression ( including slapping, hair pulling, and the throwing of missiles (p.625) and 10 incidents of physical assa ult ( including kicks and punches) in 14 days, indicating that these workers can be at ve ry high risk of harm. Ringstads (2005) national study also found that workers in residenti al settings experienced significantly more physical assaults than other workers. According to a national study of NASW members, more than 4 out of 5 of licensed social workers in criminal justice agencie s have concerns about their safety (NASW, 2006b). Newhill (1996) also found criminal j ustice settings to among the most dangerous in her study of social workers in Califor nia and Pennsylvania. Littlechild (1997) found that 30% of the probation officers he interviewed had been victims of physical violence. Schultz (1987) similarly reporte d 25% of corrections social workers had been attacked with knives and that corrections settings were one of the two most likely places to experience threats of harm. Rapp-P aglicci (2004) found that 69% of the probation officers that she studied were concerned about safety at least some of the time. Beaver (1999) found that, by far, social workers in the criminal justice field experienced higher rates of client violence, with a 53.3% rate, though the trend noted was not statistically significant. Similarly, psychological violence to social workers was found to be significantly more prevalent in corrections sett ings than most other settings in Ringstads (2005) national study of NASW members. Institutional mental health settings may also be ha zardous for social workers. In a national survey of social workers, 82% of those in psychiatric hospitals reported concerns
50 about personal safety on the job (NASW, 2006). A na tional probability study indicated that workers in a mental health setting experienced the largest number of physical threats, when compared to workers in 6 other practice settin gs (Jayarante et al., 1995). Similarly, Jayarante and other colleagues (2004) found institu tional mental health settings to be one of the two most dangerous settings for social worke rs. In another national study, Ringstad (2005) found that workers in inpatient mental healt h settings experienced significantly more physical and psychological assaults. When all workers in a university psychiatric center were surveyed, 34% of the clinicians, includ ing persons with an MSW or BSW, had been assaulted by patients (Privitera et al., 2 005). In a study of injuries to staff in a large state psychiatric hospital, 70.7% of the inju ries incurred to staff were head injuries, which is a significant occupational hazard (Carmel & Hunter, 1993). School social workers may also be very vulnerable t o client violence. In a national study, Jayarante and colleagues (1995) fou nd that school social workers experienced the highest percentage of assaults and similarly, Jayarante, Croxton and Mattison (2004) found that 21% of school social wor kers had been physically threatened and 5% had actually been assaulted. School social w orkers may also experience significantly more psychological aggression, accord ing to Ringstads national study of NASW members (2005). A national random sample of sc hool social workers found that many feared for their safety, especially if they wo rked at inner city schools (Astor et al., 1998).
51 Race /Ethnicity Match of Field Educator and Student Historically, social work and counseling literature on the effects of ethnicity on supervision process has been sparse and the literat ure that does exist tends to be more theoretical (Page, 2003). To date, there have been few empirical studies on the effects of cross-cultural supervision and the studies that hav e been done are plagued with a lack of comparison groups, small sample sizes of minority g roups, and the use of inexperienced supervision dyads to measure the supervision relati onship (Leong & Wagner, 1994). Specifically, there have been no studies on how the race/ethnicity match in a supervision dyad may impact exposure to client violence. Furthe rmore, there are no studies that indicate that the race of the field educator may im pact a students exposure to client violence. However, there are studies on how the stu dents perception of cross-cultural supervision may impact how the student receives and utilizes supervision (Cook & Helm, 1988; Hilton, Russell, & Salmi, 1995; Lewis & Ginge rich, 1980; McRoy, Freeman, Logan, & Blackmon, 1986; Ramirez, 2003; VanderKolk, 1974). Cross cultural supervision. The few empirical studies that have been done in th e area of cross-cultural supervision shed some light on cross-cultural supervisory relationships, in spite of the studies limitations and datedness. There have been some indications that non-White students may be suspicio us of supervisors who are of a differing race/ethnicity. Cook and Helm (1988) stu died 225 minority students in psychology graduate programs, 89% of whom had White supervisors. Thus, a very large majority of the students in this sample were in unm atched racial/ethnic supervision dyads. Blacks, Hispanics, and Native Americans fel t significantly less liked by their
52 supervisors than did Asians. Native Americans were significantly more likely to feel that their supervisors had a high level of discomfort wh en relating to them. This was particularly significant in light of the studys fi nding that perceived supervisor liking explained almost 70% of the total variance in satis faction with cross-cultural supervision. Another study of 37 Native American graduate studen ts and 40 non-Native American graduate students may further explain the Native American students perceptions concerning their supervisors (Lewis & G ingerich, 1980). These researchers indicated that there were significant differences i n the way that the two groups perceived leadership. The Native American students believed t hat leaders are more sacred, personoriented, intuitive and honest. Seventy-six percent of the Native American students felt that characteristics of a person more defined them as a leader as opposed to the skills and knowledge they possessed. Conversely, the non-Nativ e American students largely believed that skills and knowledge were the most im portant basis for choosing a leader. Since this study only compared leadership perceptio ns of Native American students and non-Native American students, it is possible that d ifferential leadership perceptions may be unique to the Native American culture. However, it is also possible that some of these leadership perceptions may translate to other cultu res. This study did not report on whether the students responded differently to leade rs, depending upon their perceptions of the leader, but it is frequently the case that p erceptions and beliefs influence actions (Fishbein & Ajzer, 1975). Thus, it is possible that when students do not perceive qualities of leadership in a supervisor, they may not be as a ccepting of the leaders authority.
53 Other studies indicate that non-White supervisees m ay expect or experience a lack of understanding by supervisors from a culture othe r than their own. For example, Vander Kolk (1974) studied the relationship of race to anticipation of the supervisory relationship for counseling students. The 41 White students and 9 Black students were divided into two groups, depending on whether they expected the supervisory relationship to be facilitative. It was found that the Black students were significantly more likely than the White students to expect their supervisor to be less empathetic, less respectful, and less congruent. Yet another study included 323 professionals and paraprofessionals from 11 sites of the MHMR (no exp lanation of acronym in article) system in south Texas (Ramirez, 2003). At least 40% of the respondents in this study were in social work positions. Overall, the fluent Spanish speaking respondents were significantly more likely to believe that superviso rs needed more training/education about Hispanics and less likely to believe that the super visor promoted ethnic-sensitive practices. The Hispanic practitioners were signifi cantly more likely to view supervisors as marginal on cultural competence. One social work study looked at supervision of mino rity students to discover what potential problems can occur in cross-cultural supe rvision (McRoy et al., 1986). Overall, both the field instructors and the students identif ied far more problems with cross-cultural supervision than advantages. Both Hispanic and Blac k students felt that there could be problems with racism. Black students specifically m entioned that there could be problems with the power and authority of field educators in terms of grading. White students said that problems could develop from the field instruct ors heavy accent, differing values and
54 inability to be honest and direct. Several of the B lack field educators had experienced negative attitudes from White students when the stu dents questioned their competence and were unwilling to accept supervision. Like Bla ck field educators, Hispanic field educators also said that there can be problems asso ciated with power and authority in cross-cultural supervision dyads. White field educa tors additionally perceived that there were numerous potential problems in cross-cultural supervisory relationships, including lack of empathy, lack of knowledge of cultural diff erences, prejudice, student defensiveness and failure to recognize student stre ngths. Of seven students who had experienced problems as a result of cross-cultural supervision, only two had discussed the problems with the field educator. Even when student s perceived that field educators were knowledgeable and sensitive about cultural differen ces they still did not initiate conversation about any discomfort that they had. An empirical article discussed the effects of super visor race and level of support on perception of supervision (Hilton et al., 1995). The 60 undergraduate psychology supervisees that were recruited for the study were White. There were three Black supervisors and three White supervisors, each of wh om supervised part of their supervisees in a low-support condition and an equal part of the supervisees in a highsupport condition. The study found that supervisor race did not significantly affect the supervisees ratings of support by the supervisor. However, the participants who experienced high levels of support from their super visors evaluated the supervision more positively.
55 Supervision functions in social work. In his book about social work supervision, Kadushin (1976) delineated three functions for the social work supervision: administrative, educational, and supportive. For th e administrative function, authority has normally been given to the supervisor by agency adm inistration. This authority is the sanctioned use of power (p. 93). However, such aut hority must be endorsed by supervisees before it can be fully implemented. As Kadushin said, supervisees have the ultimate veto power (p. 106) concerning a supervi sors authority. The supervisory relationship can become filled with transference, ambivalence, and resistance as a result of earlier life experien ces. Kadushin (1976) stated that when supervisor and supervisee are of different races, s ome historical societal racial tensions may affect the supervisory relationship. Fong and L ease (1997) concurred, stating that a number of historic, cultural and role-related issue s affect white supervisors attempts to provide cross-cultural supervision (p. 389). These problems may include unintentional racism, unbalanced power dynamics, trust issues, an d misunderstood verbal and nonverbal communication. Furthermore, the White sup ervisor may accept the White cultures values as the standard, making minority s upervisees actions and thoughts seem comparatively resistant and deviant. Kadushin (1976) stated that the supportive function of social work supervisors is most essential because without support to deal with stresses, the students work may be seriously impaired. He said that supportive super vision provides the psychological and interpersonal supplies that enable the worker to mo bilize the emotional energy needed for job performance (p. 200). Furthermore, it was sta ted that:
56 In implementing the objectives of supportive superv ision the supervisor seeks to prevent the development of potentially stressful si tuations, removes the worker from stress, reduces stress impinging on the worker and helps her to adjust to stress. (p. 271) Integration of knowledge on cross-cultural supervis ion and supervision functions. Though literature is sparse, there are indications that cross-cultural supervision may present numerous potential problems, including feel ings on the part of the supervisee that the supervisor may be uninformed, uncaring, and eve n racist or biased. These perceptions, real or imagined, may make it difficul t for a student to respect the authority of the supervisor or to receive the necessary suppo rt needed to do the job. It may be difficult for the supervisee to even discuss issues in supervision. In the case of a social work student who believes that a given client situa tion may be dangerous or who has already been victimized by a client, a cross-cultur al supervision relationship may inhibit the discussion of such concerns. If there is a fail ure to appropriately supervise a student or if there is a failure of a student to appropriat ely utilize supervision, it is possible that high stress client situations may not make it to th e supervision dialogue and opportunities may be lost to protect the student. Admittedly, no empirical studies support such a claim, but the connections that can be made in the literat ure warrant a closer look at this possibility. Training on Safety and Client Violence Social work client violence studies have found that time spent in safety training may be limited. Newhill (1996) reported in her samp le of over 1000 social workers that only 4% recalled receiving training in their MSW or BSW program and 3% recalled receiving training at their field internship site. Most of the 59% of those who had safety
57 training had received it at their workplace. A majo rity (79%) said that they needed more training. Similarly, a national study of school soc ial workers looked at where workers received safety training, and found that only 5% ha d received training as a part of their graduate or undergraduate social work studies (Asto r et al., 1998). Sixty two percent had received training in the form of in-service trainin g from the school district and 70% had gotten training at conferences. In spite of the fac t that many had received training, 59% reported a need for immediate training focused on s chool violence. Both of these studies included many social workers who had been in practi ce for a substantial period of time and they relied on recall of their schooling to inf orm their responses. Rey (1996) reported that of the 175 social workers studied, 41% had completed training on clinical predictors of violence and onl y 2% had training on violence prevention. Only 22% had training on safety issues during home visits. A little more than half were aware of their agencys safety policies. Training Student Perceptions School social work studies that contain information about safety training are largely written from two perspectives. The first pe rspective is of those who primarily receive training, the student, and second is that o f those who may be responsible for providing the trainingfield directors, field educ ators and field agencies. From the student perspective, Knights (1999) study of 110 B SW students in one social work program included a survey pre-field and post-field. In the pre-field survey 68% of the students projected that participating in training o n how to verbally diffuse a potentially threatening situation would help to reduce any anxi ety about their personal safety in their
58 field placement. Also, 56.2% of the students felt that training in self defense would help and 42.6% thought that training in assertiveness wo uld help. In the post-field survey, less than 50% of the students had any training in the th ree subjects. About one half or more of those who had training on how to verbally deescalat e a situation felt that the strategy had effectively reduced their anxiety. One third of the students who had self defense training felt that the training reduced their anxiety. Training in social work program venues. In one of the few studies including MSW students, data from 36 MSW students in Californ ia found that over 90% of the students said that they felt their social work prog ram had either not prepared them at all or only somewhat prepared them to handle situations where they might feel concerned about their personal safety (Schwarzmueller, 1998). Tulley and colleagues (1993) received responses from 121 BSW and MSW students, s lightly over half of whom reported that they had received safety training in the social work curriculum. Most of them had received the training in practice classes. Another study indicated that 78% of the 37 BSW students from one program said that that the social work program had offered the most information about violence in the field (Mama, 2001). These students were not asked to differentiate where in the progra m they received the training. Elwood and Rey (1996) found that only 17.9% of the social workers had academic training on predictors of violence, whereas 41% had classroom t raining on violence prevention and 43% had learned about threats toward students in so cial work classes. Two social work studies from the student perspectiv e suggest that there is a deficit of safety training in field agencies. Mama (2001) reported that 32% of BSW
59 students surveyed said that they had received train ing on how to handle dangerous situations at the field agency. Similarly, Elwood and Rey (1996) indicated that almost 35% of the social workers in their study received a gency training about predicting violence and even less (18%) had received agency tr aining about safety on home visits. Data indicated that overall the students had more t raining in social work classes than practicum agencies. Field agency safety policies. Only two studies have asked students about their knowledge of field agency safety policies. Mama (20 01) reported that 51% of the 37 students surveyed stated that their field placement agencies had a policy regarding service to dangerous clients. Similarly, almost half (48.7% ) of the 78 social work students in Elwood and Reys 1996 study were aware of agency sa fety policies. Training content areas. The only study to ask for student reports of the co ntent of safety training they had received was Elwood and Re ys (1996) study of 60 medical students and 78 social work students from Universit y of Nevada. They included questions about the students training on safety and violence by asking about various content areas of training on client violence. The questions about content areas were divided into content areas received as a part of formal educatio n and content areas received as a result of training in hospitals, clinics, and placement ag encies. Eleven of the formal education content areas were related more to circumstances wh ere violence may occur. For instance, questions were asked as to whether studen ts had received training on suicide, child abuse, domestic violence, and homicide. Three questions were related to training students had received to recognize and/or prevent v iolence. Specifically the three content
60 areas were predictors of violence, violence prevent ion, and threat toward students. Seven questions were asked about training in the field ag ency. While these content areas were generally more oriented towards recognition and/or prevention of violence they included control by medication and use of safety equipment, which may not be relevant in every social work setting. The social work students in th e study had more safety training than medical students and the social worker students enc ountered less violence than the medical students. Additionally, the type of trainin g may have affected exposure to violence. For example, medical students academic p reparation in the area of violence frequently included training on suicide, post traum atic stress disorder, child abuse, and managing violence with the use of medication. In co ntrast, more social work students had training on clinical predictors of patient violence and de-escalation techniques. This has been the only article to attempt to relate numerous content areas of safety training to a reduction in incidents of client violence and as su ch, it has strong merit. TrainingField Director and Field Educator Perspec tive Student training. Ellison (1996) reported that 61% ( n = 89) of the field directors at 147 social work schools said that they had safet y and violence training for students. The median length of time for students to receive t raining on safety issues was two hours. In contrast, Reeser and Wertkin (2001) found that o f 258 social work programs only 38% offered some type of safety training to students. T he majority (69%) of those mandated student safety training. Of those programs that of fered training, only 37% felt that their training was adequate. Seventy seven percent of the field directors felt that safety content should be integrated across the curriculum. In a mo re recent study, Faria and Kendra
61 (2007) found that 68% of the 19 social work schools taught safety education in their undergraduate curriculum. The six schools that did not teach safety content in their curriculum commented on why they did not. Some of t he comments included not having enough room to put it in the curriculum and one sch ool replied, I have never heard anyone mention it (p.146). A major issue with the Faria and Kendra study was there was only a 10% response rate, so it was not possible to generalize the results to all social work programs. Ellison (1996) generally reported training content areas covered in the 147 social work programs, but it is unknown what percentage of programs focused on each content area. The top five training components in social wo rk training programs were: assessing situations for potential danger, how to work with d ifficult clients, identifying escape routes, de-escalating volatile situations, and asse ssing when to have a co-worker present. Reeser and Wertkin (2001) asked field directors spe cifically if they offered safety training in particular content areas. The majority said that they offered training in awareness of danger (93%), assessing situations for danger (86%), and de-escalation of potentially threatening situations (73%). About one quarter (28%) offered self defense skills. Faria and Kendra (2007) reported in detail about the safety content covered by BSW programs. They listed 31 content areas and aske d if the programs had offered training in those areas. Ninety percent ( n = 12) of the programs that taught safety content in the curriculum said that they included content o n the following 9 topics: characteristics of high risk situations, creating safe office space s, highrisk practice settings, maintaining a confident, secure demeanor; verbal de -escalation of a clients rage; how to
62 behave with an angry client; how to dress; and how to sit when interacting with a client. Fifty four percent ( n = 7) said that they covered 75% or more of the 31 content areas. Two studies of social work programs have addressed where field training is received. Faria and Kendra (2007) reported that of the 13 BSW programs that taught safety in the curriculum, 92% included it in the up per division courses. Specifically, 69% ( n = 9) taught safety content in practice courses and 54% ( n = 7) taught it in both practice classes and field seminar. Reeser and Wertkin (2001 ) found that 50% of safety training was incorporated into social work practice classes, whereas 17% of programs offered safety training as a special workshop. Other progra ms offered a combination of the two. Social work schools may rely on field agencies to t rain on students safety concerns. Reeser and Wertkin (2001) reported that 96% of 258 program respondents felt that the field instructor had the primary responsib ility for discussing safety issues with student, yet field directors were unaware if traini ng was being done in the agencies. Many programs were unable to estimate how many of their field agencies were offering safety training. Faria and Kendra (2007) similarly found t hat those programs that did not teach safety education in the curriculum relied on field agencies to do safety training with the students, yet of the 19 schools, only one program s aid that safety was taught in all of its field agency sites and four others said that it was taught in most of their sites. Seven (37%) said that they didnt know if safety was taug ht in the field agencies. Field educator training. Field educators have reported a need for more train ing on client violence. Tulley and colleagues (1993) fo und that 94% of the field educators indicated that they had a need for more education o n the issue of violence. Mama (2001)
63 reported that of 17 field educators surveyed 59% sa id that their agency didnt have sufficient staff training for workplace violence or safety and Ellison (1996) found that only 30% ( n = 45) of social work programs offered safety train ing to field educators. Additionally, the time spent in training was relati vely brief. Social work programs reported exposing field educators and field liaison s to a median of 45 minutes of safety training. Social work program policies. Three studies surveyed social work programs to ask about social work program safety policies. Elli son (1996) reported that of 147 programs, 38(26%) had some type of safety policy. R eeser and Wertkin (2001) found that of 258 social work programs only 12% ( n = 31) had a formal written policy to address client violence. MSW programs were significantly m ore likely to have written policies than BSW programs. Additionally, the programs in wh ich a student had experienced assault and/or threat were more likely to have form al policies. An additional 41% of the programs had informal policies (unwritten policies that had not been formalized), though a large amount of those stated that the informal po licy consisted of relying on safety training in field seminar, student orientations, an d social work practice classes. A few of the programs stated that they did not intend to dev elop policies because they thought this was the responsibility of the field agency. Faria a nd Kendra (2007) found that 21% ( n = 4) of the BSW social work programs that responded t o their survey had written safety policies. Forty two percent ( n = 8) had no policies and 32% ( n = 6) did not know if they had policies.
64 Field agency safety policies. Studies conflict over the number of field practicum agencies that have safety policies, partially depen ding on who is asked. When field educators are asked, answers are varied. Tulley and colleagues (1993) received surveys from 96 field educators who were affiliated with th e University of Georgia and found that two thirds of the field educators stated that their agencies had safety policies though less than half of them actually knew the policies. In co ntrast, Mama (2001) found in her survey of 17 field educators that 17% ( n = 3) had an agency policy on providing services to clients in potentially dangerous areas and 41% ( n = 7) said that their agencies had adopted policies on providing services to dangerous clients. Only one study reported surveying field agencies di rectly about field agency safety policies. Lyter and Martin ( as cited in Lyt er & Abbott, 2007) found in the year 2000 that of 200 field agencies in one geographic r egion on the East Coast, only 18% reported that they had a formal written agency safe ty policy. Since this study was reported within another study, nothing is known abo ut the methodology, nor to whom was the safety survey addressed. Summary on training. In spite of numerous recommendations to implement safety training for students, only one study with a small sample of social work students found a possible relationship between training and reduction of incidents of client violence (Elwood & Rey, 1996). Even in this study, no statistical significance is mentioned between having more training or training on specific content and having less incidents of violence. One additional study (Schat & Kelloway, 2003) addressed the possibility of training reducing the effects of wor kplace violence. While the inclusion of
65 training may be helpful in reducing incidents and e ffects of client violence towards social work students, its effectiveness needs to be furthe r evaluated. Additionally, field directors and program directors may believe that their programs have offered more safety training than stu dents recall receiving. More information is needed to understand more about the student perspective of how much safety training is received and where it is receive d. Finally, it would be helpful to know more about the content areas where students have re ceived training in order to assist social work programs in designing effective trainin g on safety and violence. Effects of Violence on Social Workers Short Term Effects of Violence on Social Workers Researchers indicate that social workers are affect ed in various ways when they have been victimized. Newhill (1995) presented thre e case vignettes in an article summarizing findings on client violence toward soci al workers. In all three cases, the social workers were highly skilled MSWs who had bee n harmed by client violence. It was reported that the typical immediate response to the violence was a numb feeling of unreality (p. 636). The numbness later turned into a realization that it was possible that they could be mortally harmed while doing social wo rk. When the social workers began to realize that they had been subjected to life thr eatening harm, they developed feelings of helplessness and demoralization. Jayarante and colleagues (1995) reported that worke rs who had experienced verbal abuse, physical threats, threats of lawsuits or sexual harassment reported significantly higher levels of depression. Addition ally, the victimized workers were
66 significantly more likely to depersonalize their cl ients. Newhill and Wexler (1997) found that 23% of those who had property damage were sad or depressed, but less social workers were depressed when they were threatened (1 4%) or actually attacked (5%). Newhill and Wexler (1997) found that many social wo rkers experienced strong emotional reactions immediately following the clien t violence incident. The reactions varied with the type of violence. When workers expe rienced property damage, 63% reported anger as their most prevalent immediate em otion. Social workers were also angry when threatened (48%) or actually attacked (4 3%), though the feeling wasnt as prevalent. When workers were actually attacked, the y were much more likely to experience feelings of helplessness and inadequacy, being drained or emotionally exhausted, and shocked, surprised, or shaken-up. In Newhill and Wexlers study, 69% of child and you th workers who had been threatened were scared or fearful and 62% were anxi ous. This coincides with Jayarante and colleagues (1995) finding that social workers w ho were victims of client violence were more likely to be anxious and irritable. Inter estingly, in Newhill and Wexlers study, a smaller percentage of those who were actua lly attacked reported feeling scared (39%) or fearful (43%). These findings resonate wit h those of Rapp-Paglicci (2004), who found that probation officers who had been personal ly victimized had less concern about their safety than those who had not been victimized To compound these initial reactions to trauma, Schu ltz (1989) points out that it is often assumed that social workers who have experien ced this trauma will be able to effectively cognitively manage the burnout and stre ss. In Schultzs study, 83 social
67 workers reported that neither staff networking nor peer review of violent experiences was available to aid in resolving the effects of the tr auma. Social workers who experience client violence may feel that they should have been able to predict the assault and that the attack could have been avoided (Guy, Brown & Poelst ra, 1991). Guy and Brady (1998) state that this increases feelings of guilt, shame, and sense of failure in the worker, which can be particularly difficult for students: When combined with the emotional turmoil inherent i n being a graduate student, such as financial hardship, academic pressures, fea r of the unknown, and loss of a prior support network, the emotional impact of pati ent attack on the well being of the trainee may be completely overwhelming and debi litating. (p. 405) Long Term Effects of Violence on Social Workers When short term effects of violence are not adequat ely addressed, some social workers may proceed towards longer term trauma effe cts. Snows 1994 study found that 75% of the child welfare workers she interviewed me t criteria in three categories of Post Traumatic Stress Disorder: re-experiencing the trau ma, avoidance or numbing of responsiveness, and increased arousal. Social worke rs may also suffer negative effects on their health when trauma symptoms have not been add ressed. Balloch and colleagues (1998) did the only longitudinal study on client vi olence and found that even two years after an incident of client violence, social worker s had significantly higher scores on the General Health Questionnaire (a scale that detects minor psychiatric disorders in the general population) than home care workers or resid ential workers. When social workers have been physically assaulted, they often react instinctively with self protective responses that can include agg ression toward the client. When this
68 has occurred, they may worry for a long period of t ime about eventual litigation or malpractice claims by the client (Guy and Brady, 19 98). Littlechilds (2005) qualitative study reported tha t a core issue in terms of longer term effects and severity of effects is whether the social worker is personally threatened. Several people in this study had experienced person al threats of clients saying that they would get even or kill them or their families. Su ch veiled or blatant threats contributed to workers having increased fear of personal violen ce. In fact, Littlechild found that almost half of participants experienced anxiety, fr equently mixed with anger and fear, both during the client violence and when they later thought about the situations. This finding was echoed in another qualitative study by Smith and Nursten (1998) in the comments of female participant about a male client: He said he was going to get me, beat me up, follow me home I would be sorry the thr eats were against me, not against the department (p. 357). Norris (1990) found that respondents tended to be m ore cautious in practice after experiencing violence. They had less ability to con front clients. Some felt that violence had reduced their skills and they felt less able to help other clients. Additionally, there was a tendency to retreat to more administrative po sitions in order to avoid direct client contact. Some felt that they might decide not to vi sit some clients alone and they might refuse to work certain shifts alone. An additional long term effect is the tendency for social workers who have been victimized by clients to contemplate leaving the fi eld of social work. Lyons, Lavalle, and Grimwood (1995) stated that one of the primary reas ons practitioners leave the profession
69 is the threat of or experience with violence. Of t he social workers who suffered abuse or threat, 24.6% had left the profession of social wor k and 12% were contemplating leaving the field. Similarly, Newhill and Wexler (1997) fou nd that 19% of their sample of child and youth social workers who had experienced violen ce reported not wanting to return to work. One practitioner stated that: It has reinforced my belief that the profession doe s not take the threat of violence seriously, and I have found that many people feel u ncomfortable admitting fear or lack of control. It adds to other reasons to leave social work. (p.207) It is possible that social workers contemplate lea ving the profession because experiencing violence is an assault to their profes sional competence. Bibby (1994) found that social workers have an unrealistic view that t hey should have been able to stop the violence. Effects of Violence on Social Work Students It has been found that students who are physically assaulted state that they would be uncomfortable engaging in all practice activitie s (Knight, 1999). Furthermore, one half of all of the students questioned felt that social work was a dangerous profession. Almost two thirds of the students thought they might be in private practice at some point. Two thirds also stated that personal safety needs would play some role in their career decisions. In contrast, an early study of students exposed to violence indicated that at times, they began to view the environment as less threaten ing than it really was (Mayer & Rosenblatt, 1979). They believed that the school of social work would not send them to a neighborhood that was truly dangerous.
70 Fear of Future Client Violence When workers have experienced workplace violence, t hey have a tendency to fear returning to the site (Waters, Lynn, and Morgan, 20 02). Guy and Brady (1998) reported that fear of future victimization is the most frequ ently reported emotional consequence for psychotherapists who were victims of physical a ttack. Budd, Arvey, and Lawless (1996) stated that experiencing violence is associa ted with a significantly higher likelihood of worrying about violence. In Snows ( 1994) qualitative study of child welfare workers, 90% reported that they felt fearfu l of imminent physical danger at work. Atkinson (1991) likewise found in a qualitative stu dy of 8 clinical workers that they tended to have fear and anxiety about being in situ ations that were reminiscent of the assault. These fears were predominantly contained i n the workplace and they generally did not discuss them at home. They largely looked f or support from their colleagues as administrative support was almost non-existent. Som e commented that agency and system influences made the effects of the assault l ess personally frightening. Making sense of the assaults helped to ameliorate the nega tive effects from the assault. Rogers and Kelloway (1997) found supportive evidenc e that fear of future violence mediated the relationship between exposure to workplace violence and negative outcomes such as psychological well being, somatic complaints, and intent to leave the organization. Using a sample of in-home service wo rkers, Barling, Rogers, and Kelloway (2001) found that having an experience wit h workplace violence caused a fear of a recurrence of violence in the workplace. Fear of future violence then predicted lower organizational commitment and increased withdrawal intentions. A national study of 296
71 children and family social workers echoed this find ing. Song (2005) found that fear of future victimization fully mediated the relationshi p between client violence and burnout and the relationship between client violence and af fective organizational commitment. In other words, experiences with client violence direc tly and significantly increased fear of future violence. Fear, in turn, was the factor that directly and significantly increased burnout and decreased affective organizational comm itment. Fear did not significantly predict career turnover intentions. These two studi es looked at the effect of fear on organizational commitment and career turnover inten tions rather than on occupational commitment, but they provide support for the idea t hat fear may be a result of client violence and this fear may have negative effects on social workers. Compromised Occupational Commitment When a social work practitioner or student has been exposed to client violence, it is possible that their commitment to the profession of social work may be compromised. Prior to the early 1990s, the terms occupation, pr ofession, and career were used interchangeably in the literature. Blau, Paul, and St. John (1993) suggested that a tighter work referent, such as occupation, be used in futur e research after they found that some responses to items referring to career were related to some organizational commitment facets as opposed to occupational commitment facets Lee, Carswell, and Allen (2000) defined occupation as an identifiable and specific line of work that an individual engages in to earn a living at a given point in tim e. It is made up of a constellation of requisite skills, knowledge and duties that differe ntiate it from other occupations and typically, is transferable across settings (p.800) Earliest references to occupational
72 commitment actually involved research of career com mitment. Hall (1971) defined career commitment as the strength of ones motivation to work in a career field (p. 59) and Blau (1988) defined it as ones attitude toward on es vocation, including a profession.(p. 295). It has also been defined as a persons belief in and acceptance of values of his or her chosen occupation or line of w ork and willingness to maintain membership in that occupation (Vandenberg & Scarpe llo, 1994, p. 535). Correlates of occupational commitment. Lee, Carswell, and Allen (2000) completed a meta-analytic review of occupational co mmitment, aggregating correlations of variables in 76 studies on the subject. They fou nd that demographic variables did not correlate with occupational commitment. Additionall y, occupational commitment was positively associated with work related attitudes s uch as job involvement, job satisfaction, satisfaction with work itself, cowork er satisfaction, pay satisfaction, work ethic endorsement, and career satisfaction. Negativ e correlations included burnout, reduced accomplishment, and depersonalization. Add itionally, job stress showed a moderate negative correlation with occupational com mitment. Occupational commitment was most strongly and positively related to occupat ional turnover intention. Based on the observations in the metanalysis, Lee and colleagues felt that maintaining occupational commitment might depend on job design variables, as reaction of respondents to the job greatly affected occupational commitment. In a sample of 237 nurses in Western Canada, it was found that career (occupational) commitment was negatively correlated with job tensions (Cohen, A., 1999). Job tensions were measured using a scale tha t measured tensions and pressures
73 growing out of job requirements. This is an indicat ion that as job tensions increase, as can happen with experiences such as client aggression, career commitment decreases. Turnover intentions have been found to be negativel y correlated to all forms of commitment (Meyer & Allen, 1991; Meyer et al, 1993) Brierly (1996) found a significant negative correlation between profession al commitment and professional turnover intentions. Blau (1988) found almost the s ame correlation between career commitment and career turnover intentions. Occupational commitment in social workers. Landsman (2001) studied occupational commitment in 990 public child welfare workers in the Missouri Department of Social Services. He found that both j ob satisfaction and occupational commitment were significantly and positively associ ated with intent to stay in the occupation. Additionally, having a degree in social work was positively and significantly related to occupational commitment, indicating that professional preparation in the social work profession may increase the probability of a w orker having stronger commitment to the profession. Landsman used Structural Equation M odeling with this large sample and established good evidence of causal ordering in the areas mentioned. Gifford (2003) questioned 207 social service employ ees in three multi-service social service organizations in a suburban communit y in New York State about organizational and professional (occupational) comm itment. It was found that professional commitment was positively related to h aving an administrative position in the organization. It was also found that workers in public agencies had less professional commitment than workers in non-profit agencies and proprietary agencies. Gifford
74 speculated that line staff social service workers h ave less commitment to the profession since they have less responsibility and less flexib ility with decision making. This study did not differentiate between persons with professi onal training, such as social workers, and other social service workers who do not have pr ofessional training, so it is unknown how professional training influences professional c ommitment in this study. With his sample of 179 professional social workers in Israel, one researcher looked at the willingness of the social worker to s eek help, along with gender, education, religiosity, seniority in social work, and ethnicit y as possible predictors of professional commitment (Cohen, B. 1999). He found that willingn ess to seek help was the most powerful predictor of professional commitment. This finding possibly relates to the present study since social workers who have been ha rmed by client violence may maintain their professional commitment to the exten t that they are willing to seek help when an incident has occurred. An additional study of organizational and career (o ccupational) commitment was done by a lecturer from the social work school at U niversity of Haifa in Israel (Freund, 2005). The sample included 220 employees in welfare organizations providing community service. He found that career commitment was significantly related to intention to leave. Additionally, he found job sati sfaction to be the most meaningful factor influencing withdrawal intentions. The respo ndents in this study largely occupied executive positions. Two thirds were males and at l east half had graduate degrees. It is unknown how many, if any, were professionally train ed social workers.
75 Occupational commitment in students. Occupational commitment in social work students has been measured in only one study. Profe ssional (occupational) commitment in social work students was studied by researchers in Israel, utilizing data collected from 227 senior BSW students at 5 Israeli universities ( Lazar, Cohen, and Guttman, 1995). These researchers used Allen and Meyers (1990) org anizational commitment scale, changing the wording to reflect commitment to the p rofession of social work. The strongest predictor of professional commitment was the variable that measured how likely it was that the students would stay in their profession. Researchers speculated that those students who stated that they would not chang e their profession were likely to be satisfied with their jobs and believed that social work jobs would continue to offer challenges and opportunities for advancement. Meyer, Allen, and Smith (1993) examined career comm itment in 355 student nurses in Ontario. They used their own scale to mea sure occupational commitment on three subscales: affective, normative, and continua nce commitment. They found that affective commitment and normative commitment were related to positive experiences, such as satisfaction with the job or training exper ience. They speculated that positive experiences could lead to affective attachment to t he profession and a sense of obligation to the profession. Inversely, it might be expected that negative experiences might lead to less affective and normative career commitment. Add itionally, both affective commitment and normative commitment were positively and significantly related to intent to stay in the career. Continuance commitmen t correlated negatively with intention to establish a long term career. It was also relate d to variables that reflect an increased
76 investment made by the student, such as year in the nursing program, years in nursing, and employment status. Thus, the researchers observ ed that by the end of the nursing program, students may be staying in the profession more because of the time, money, and energy they have put into the program (continuance commitment), and less because of desire (affective commitment) and obligation (norma tive commitment). In this study measures were taken at two points in time during th e nursing programthe first, after 2 weeks of school and the second, near the last week of class in March. It is of interest to note that both affective and normative commitment d eclined as students progressed through their program. Career commitment was examined in 92 nurses in thei r first year of training in England (Arnold, 1990). A series of multiple regres sion analyses were performed in order to establish the most effective predictors of caree r commitment at Day 1, 4 months, and 1 year. On Day 1, commitment was higher when the deci sion to enter the nursing field was perceived to be an important decision that had been publicly discussed with people close to them. At one year, experiences during nursing t raining predicted career commitment. The more negative the experience, the lower the com mitment to the career of nursing. Also, unmet expectations during the fourth through twelfth months predicted career commitment, but only if those expectations had cons tituted reasons to enter the profession of nursing. Overall, the mean level of c areer commitment was higher during the first four months of training, but dropped slig htly between four months and a year. In the three student studies, negative experiences in their academic setting and negative experiences in the practicum setting were noted to lessen career commitment.
77 Conversely, positive experiences in the practicum o r training settings led to stronger career commitment. Measurement of occupational commitment. In 1993, Meyer, Allen, and Smith created a new three component Occupational Commitme nt Scale, based on their previous work with an organizational commitment scale. In Al len and Meyers (1990) review of organizational commitment literature, they found ev idence of three distinct themes. They called these affective commitment, which involves a n employees decision to stay because he/she wants to stay; normative commitment, which involves a decision to stay because they feel obligated to do so; and continuan ce commitment, which involves a decision to stay because they need to do so. Meyer and Allen believed that one could better understand a persons commitment to an organ ization if the three areas are differentiated. Factor analysis has indicated that the three areas of commitment are distinct concepts. The concepts of affective, normative, and continuan ce commitment were first used to measure occupational commitment in 1993, us ing modified versions of the organizational commitment scales. They demonstrated good discriminant validity and acceptable reliability scores. These scales have be en demonstrated as effective with professional students. A modified version of the or iginal organizational commitment scale has also been used with social work students in Israel (Lazar, et al. 1995). Career Withdrawal Intentions Evolution of the variable. A literature review on employment turnover stated t hat the earliest studies of employment turnover appeare d in the early 1970s and
78 predominantly used single explanatory variables tha t did little to help understand the turnover process (Mobley, Griffith, Hand, and Megli no, 1979). These researchers stated that a linking mechanism was needed that took into account an individuals perception and evaluation of the individuals alternatives to the present work position. They proposed a complex model of employment turnover and stated that the best predictor of turnover was the intention to quit. Turnover intent ions have alternately been referred to as withdrawal cognitions, which Mobley and colleagues defined as the extent to which an individual has thought about quitting his or her jo b. Blau (1985) was one of the first to differentiate b etween organizational withdrawal cognitions and career withdrawal cogniti ons. He theorized that career withdrawal cognitions would be different than organ izational withdrawal cognitions and that the hypothesized relationships with career com mitment and organizational commitment would help to justify the conceptualizat ion of career commitment. His study of 119 registered nurses indicated that career with drawal cognitions are different than organizational withdrawal cognitions. Career withdr awal cognitions were significantly negatively correlated with career commitment, indic ating that the stronger a persons career commitment, the less they will be likely to think about changing careers. Career withdrawal intentions in other fields. In their study of 244 nursing professionals, Bedeian, Kemery, and Pizzolatto (199 1) found that turnover (withdrawal) intentions had a direct effect on turnover. This wa s verified by obtaining hospital personnel records of the nurses 6 months after they completed their questionnaires and recording whether the nurse had actually left the p osition. Additionally, career
79 commitment interacted with expected utility of pres ent job (feeling that the job would lead to future attainment of career goals) to predi ct turnover intentions. Neither career commitment nor career growth opportunities were dir ectly related to turnover when turnover intentions were held constant. This model begins to look at possible effect of a mediator (turnover intentions) and a moderator (car eer commitment) in determining turnover. Career withdrawal intentions and career stages. Career stage research assumes that individuals move through distinct occupational stages. Supers (1957) theory was that careers progress through four stages: 1) trial 2) stabilization; 3) maintenance; and 4) decline. In the trial stage, individuals are ident ifying interests, strengths, and abilities. They assume an apprentice role, learning from super visor and coworkers. In this stage, individuals achieve a sense of mastery. Aryee, Chay and Chew (1994) measured the effect of job characteristics on career commitment across career stages and found a positive correlation with career commitment across all career stages. Hierarchical regression indicated that organizational commitment was the most significant predictor of career commitment, accounting for 16% of the varian ce. While being in the maintenance stage of the career was a significant predictor of career commitment, it only accounted for 3% of the variance. Correlations in the same st udy indicate that career commitment negatively and significantly relates to career with drawal intentions across all career stages. This is significant to the study of career commitment and career turnover intentions in social work students, as it indicates that social work students who are
80 presumed to be in the trial phase of their career m ay be as likely to change their careers as social workers who have been in the field for a lon ger period. Career withdrawal intention in social work. One of the few studies to address career withdrawal intentions in social work was a n ational study of 923 direct care providers and supervisors who were providing psycho social rehabilitation in community mental health centers (Blankertz & Robinson, 1997). There was no mention of types of professionals represented in this study, so it is u nknown if any were professionally trained social workers. This study found that youn ger workers and those with bachelors or masters degrees were more likely to have intent to leave the field. Additionally, supervisors were more likely to leave, as opposed t o direct service providers. Finally, there was a positive significant correlation betwee n intent to leave the field and the emotional exhaustion scale. A study of 259 mental health workers in New York St ate was completed to determine the effect of organizational conditions o n job satisfaction and intention to leave their job (Acker, 2004). Hierarchical multiple regr essions were completed, entering demographic variables in the first stage and workpl ace variables (such as role conflict, role ambiguity, caseload, and type of work activiti es) were entered in the second stage. Twentyseven percent of the variance was accounted by the workplace variables, indicating that organizational conditions are stron g predictors for job satisfaction and intention to leave the organization. A recent article measured school social workers in tent to stay in the field of school social work, using a single item measure of whether the 48 school social workers
81 had intent to stay or intent to leave school social work (Caselman & Brandt, 2007). Moderate effect sizes were found between intent to stay and collaboration with other school personnel and self-efficacy, indicating that school social workers who felt knowledgeable and qualified to do their jobs and th ose who interact with principals, teachers, and guidance counselors as part of a team are more likely to stay in the field. Measurement. Career withdrawal intentions (or cognitions) have g enerally been measured using a three item scale that was initiall y devised to measure job withdrawal cognitions. The original scale was based on organiz ational withdrawal cognitions that immediately precede job turnover: 1) thoughts of qu itting a job; 2) intention to search for a job; and 3) intention to quit a job (Mobley et al ., 1978). Blau (1985) first utilized these same items in a c areer withdrawal cognitions scale by changing the referent of job to profess ion. He reported internal consistency of .87 and .85 on two occasions 7 months apart. Tes t-retest reliability was .57. When he used the scale again with a population of field off ice personnel in an insurance company, the internal consistency reliability was .93 (Blau, 1988). He demonstrated discriminant validity of the scale when comparing it to the job withdrawal cognitions scale. Aryee, Chay, and Chew (1994) used the same scale with mana gerial and professional employees and the Cronbachs alpha score was .91. Blau, Tatum, and Ward-Cook (2003) provided supporti ng evidence for the discriminant validity of professional (career) with drawal cognitions versus organizational withdrawal cognitions. They utilized a different re search design that included using pretest and post-test for each type of cognition. They looked at variables in an organizations
82 culture (i.e. concern for employees) and variables in a professional culture (i.e. a sense of calling and expertise), surmising that there would be differences in the variables relationships to professional withdrawal cognitions and occupational withdrawal cognitions. Again, they demonstrated that career wi thdrawal cognitions was a different construct than organizational withdrawal cognitions Though a handful of studies have addressed social workers intention to leave their agency, only three have looked at career with drawal intentions (or intent to leave the field). More study is needed on social workers car eer withdrawal intentions and whether those intentions actually lead to a career (profess ion) change. Summary Client violence towards social workers is an aversi ve subject that many prefer not to discuss. The literature on client violence inclu des studies with different definitions of violence and varied methodologies, so it is difficu lt to compare results of the studies. Many samples are too small to generalize findings. Other studies are from distinct geographic areas, also making it difficult to gener alize findings to social workers across the U.S. Some studies have focused on social worker s in particular practice settings. Of those, two have used national random samples. Six studies have been conducted with social workers from the field of child welfare/chil d protection. A total of four studies have looked at social workers at large in the Unite d States using samples from the national membership of NASW. There has been little research on client violence o n social work students. Four studies have used samples of social work field dire ctors and seven U.S. studies have used
83 student samples. Of those that have used student sa mples, all have had relatively small samples, with the largest sample size to date being 258. The student studies have also been regionalized studies, making it impossible to generalize the results of the studies. Additionally, the smaller sample sizes prohibit mor e complex statistical analysis. Only three of the studies report how students are affect ed by client violence. When social work practitioners and students have en countered client violence, it is possible that they will fear future occurrences of violence. Two social work studies have specifically explored the relationships betwee n violent incidents and trauma effects. Other qualitative and quantitative studies have dis covered in their search for effects of client violence that social workers may have fear a nd anxiety concerning future contacts with clients. Only one social work study of child w elfare workers has examined the mediating relationship of fear of future violence b etween exposure to violence and organizational commitment (Song, 2005). No student studies have specifically addressed fear of future violence. For social work students fear of future violence co uld contribute to reduced occupational or professional commitment. Most studi es on occupational commitment have been done in other fields such as organization al psychology. However, some social work studies indicated that having a social work de gree, being in an administrative position, and working in a proprietary agency incre ase occupational commitment. Israeli social work studies found that having a willingness to seek help and being satisfied with ones job increases occupational commitment. Studen t studies in other professions have shown that negative experiences in the academic set ting and in the practicum setting
84 lessen occupational commitment. The only social wor k student study on occupational commitment was done in Israel. It indicated that th e strongest associations with occupational commitment were satisfaction with care er studies, determination not to change professions, and having an intention to work in social work. No social work study has specifically looked at the effect of client vio lence on occupational commitment. Studies indicated that when occupational commitment is lessened, a person may begin to have career withdrawal intentions, or thou ghts of leaving the occupation. Though this has been studied in other fields, it ha s been almost untouched by social work researchers. Only two studies discussed intentions of leaving the field of practice. Both surveyed groups of workers that likely included soc ial workers, but it was unclear if any of the workers from either study were professionall y trained social workers. One other study with a small sample size asked specifically a bout school social workers intent to stay in the field of school social work. No studie s have focused on career withdrawal intentions of social work students. Figure 1 (Appendix A) illustrates a model for the c urrent study. It is particularly important to understand what factors may help to pr edict client violence in social work students, so that we may help students to avoid vio lent encounters altogether. Additionally, there is a need to understand what ki nds of training and venues of training may help to reduce violence toward social work stud ents. Finally, we need to understand how students are affected when they have encountere d violence. Specifically, might their commitment to the profession of social work be shak en? Might they be lost to the profession of social work? It is the intent of this study to seek answers to these questions.
85 As stated in a report on recent national study of s ocial workers, a profession cannot successfully retain its workforce when issues of pe rsonal safety go unaddressed (NASW, 2006a, p. 35). It is hoped that the findings of this study will ultimately contribute to the health and safety of social work students and that it will aid in retaining these newcomers in the profession of social work.
86 Chapter III Methodology Research Design and Methods This research was a descriptive, nonexperimental, c ross-sectional study. The primary purpose was to explore the prevalence of cl ient violence toward social work students. The relationship between certain demograp hic characteristics and experience with violence was explored, as well whether these d emographic characteristics may predict client violence. Additionally, the study ex amined whether fear of future violence mediates the relationship between exposure to viole nce and career commitment and career withdrawal cognition. The study was a nonman ipulative design, in that it examined naturally occurring variables, which were not be co ntrolled by the researcher. Expedited approval (IRB#: 106461) to proceed with t he study was received on December 19, 2007 from University of South Florida, Office of Researchs Division of Research Integrity and Compliance (Appendix B). Ini tial study approval dates were December 18, 2007 through December 16, 2008. On Nov ember 25, 2008, the Institutional Review Board extended the approval da te to November 24, 2009. Study Sample The sample for this study was drawn from an accessi ble population consisting of the 2008 National Membership Directory of the Natio nal Association of Social Workers (NASW). As the largest membership organization of p rofessional social workers in the
87 United States, NASW had approximately 138,000 membe rs, as of June, 2007. Rental of NASW membership direct mailing list was secured fro m in Focus Association of Marketing Experts, a company designated by NASW to assist with research needs (Appendix C). A conservative expectation of response rate was 30% based on a review of social work education studies utilizing national NASW stud ent membership as a sampling frame. This researcher reviewed methodologies of al l articles published in two major social work education journals in the past six year s. A total of five studies used the national NASW membership as a sampling frame for a mailed survey. The response rates for the mailed surveys were predominantly between 2 4%-36%, with one group of BSW students having a 54% response rate. In order to determine an adequate sample size, pros pective power analysis information was used, according to the directions o f previous researchers (Green, 1991; Nakagawa & Foster, 2004). In factoring together sta tistical power of .80, alpha of .05 and a medium effect size, Green suggests sample sizes, based on a rule of thumb. In the present study, the most predictor variables that we re used at a time for any multiple regression analysis was 18. According to Green, whe n the number of predictors is 18, the recommended sample size for a medium effect size is approximately 149. In order to insure an adequate number for analysis, a random sample of 1500 was requested. This number was predicted to yield appro ximately 450 responses, presuming a 30% return rate. This number was well above the rec ommended sample size, even if there was a lower response rate.
88 The sample was identified and selected by a compute rized procedure from approximately 6,554 BSW student members and 13,579 MSW student members of NASW. The sample was stratified evenly between MSW and BSW members in order to insure equal numbers of graduate and undergraduate students in the study, thus the mailing list consisted of 750 BSW students and 750 MSW students. An attempt was made to oversample various ethnic an d racial groups because ethnic and racial groups have been underrepresented in most social work client violence studies. In social work student client violence stu dies, only Knight (1996, 1999) has given any information about racial/ethnic groups. O nly in the latter study did the researcher analyze effects of race/ethnicity on exp erience of client violence. Normally, the skewed distribution of White social workers ove r other ethnicities would lead to a skewed distribution in the study sample. The small numbers of persons from minority ethnic and racial groups make it difficult to draw conclusions about the experience of minority social work students with exposure to viol ence. From the study sampling frame, 8.98% ( n = 12,386) of approximately 138,000 members of NASW identified themselves with an ethni c group other than White. It was surmised that a similar percentage of students woul d report minority. This is significantly less than the 35.6% of BSW students and 34.4% of MS W students who are reported to be in social work programs in the U.S. (CSWE, 2006). I n order to increase the possibility of having the number of respondents from non-White gro ups closer to the percentage actually in programs, a request was given to in Foc us to draw 33% of the MSW students and 33% of the BSW students from minority groups. I n the request, percentages of
89 specific ethnic groups were requested in the same p roportion that they were represented in the larger sampling frame. For example, African Americans made up 50% of the minority ethnic membership of NASW, so it was reque sted that 50% of the 33% of ethnic minorities drawn for the current study be African A merican. In response, In Focus replied that only 231 BSW students had listed their ethnicity: 164 African Americans, 17 Asian/Pacific Islanders, 25 Chicanos/Mexicans, 17 N ative Americans, and 8 Puerto Ricans. All of these were drawn for the sample and an additional 519 White students made up the BSW population. In contrast, 7009 of t he MSW students had identified race, so it was possible to randomly draw the 33% request ed ( n = 250) with the remaining two thirds of the MSW student sample being White. Once the sample was drawn, in Focus emailed the names and addresses of all in the sampl e and it was possible to proceed with data collection. Data Collection Data were collected through an 82 item paper and pe ncil self-administered questionnaire (See Appendix D). The questionnaire a ppeared on pages divided into two columns, making the instrument five pages long. It was developed using strategies of the Tailored Design Method (Dillman, 2000). Additionall y, several of the mail procedures suggested by Dillman were used in an attempt to inc rease the response rate. Both the initial and follow up cover letters were printed on University of South Florida letterhead (See Appendix E & F). The cover letters were perso nally signed in blue ink. The cover letter from the researcher, questionnaire, and stam ped business reply envelope were
90 folded together. The envelopes containing the resea rch materials were individually stamped with a 58 cent stamp and sent by regular ma il. Dillman (2000) recommends using a four contact proc ess, plus a final contact by phone in order to improve response rates. In this s tudy, the phone call was not possible because phone numbers of the students were not acce ssible. Also, a decision was made to make three contacts instead of four due to cost con siderations. A pre-survey contact was made via an announcement postcard, which was mailed on March 11, 2008, approximately one week before th e first survey instrument was mailed (See Appendix G). This post-card notified t he social work students that they had been randomly selected to participate in a national study of client violence against social work students. The students were thanked in advance for their assistance with the study. The initial mailing took place on March 11th & 12th, 2008. The mailing intentionally took place the third week of March. This was projected to be slightly past the midpoint of the spring semester at most univers ities. This allowed sufficient time for a follow up mailing without getting too close to fi nal exams, which normally occur toward the end of April. This timing helped to insu re that the majority of students in placements would be nearing the end of their field practicum placements regardless of whether they were in concurrent placements or sprin g semester block placements. To maintain anonymity and confidentiality, each qu estionnaire was assigned a number to identify nonrespondents for follow up mai lings. An Excel control register was updated as questionnaires were returned. The names that were marked off the register
91 were deleted from the mailing labels. Only those pa rticipants whose questionnaires were not received got the second mailing. The second mailing contained a cover letter which reiterated the importance of the study and expressed appreciation to any student s who had already assisted with the study by returning their questionnaire. Social work students who had not returned their questionnaires were encouraged to do so. Questionna ires were included with this mailing in the event that students had misplaced the origin al questionnaire. This mailing occurred on April 8, 2008, three weeks after the initial mai ling. After this mailing the control register was destroyed. Instrumentation A survey instrument was developed for the purpose o f measuring the variables in the study. Where existing scales and questions exis t, these were used. The questionnaire includes questions about five personal demographic factors; five practicum site/organizational demographic factors; direct and indirect encounters with client violence; reporting client violence; fear of future violence; three types of occupational commitment; and career turnover intentions. A total of 17 factors were studied. The factors are identified here, with some description. For a list of all possible responses in each category, refer to the questionnaire in Append ix D. Personal Demographic Variables Demographic information was obtained from the respo nses to ten questions. Five questions pertain to personal information regarding the participants background. Those demographic categories were: gender, age, years of paid social work experience,
92 ethnicity, and current social work program enrollme nt (BSW/MSW). The personal demographic variables are discussed below. Gender. Students were asked to check their gender. Response items are male and female. Age. Students were asked to list their age as of their l ast birthday in number of years. Traditional aged BSW students were consider ed to be those under age 25 and traditional aged MSW students were considered to be aged 25-30. All students over the age of 30 were considered to be non-traditional age d students. Years of social service experience. Students were asked the total number of years of paid social work experience they have had prior to their current degree program. A space was provided for their response in number of years. Less experienced students were considered to be those students who have had l ess than 2 years experience. Ethnicity. Students were asked to identify their ethnic/racial origin by checking the category that most clearly describes them. Cate gories were derived from NASW membership ethnicity categories. BSW/ MSW. Students were asked to check whether they are pursu ing a BSW or MSW degree. Practicum/ Organizational Demographic Variables Information about five practicum/organization relat ed variables was solicited. These include place of client violence, time of day of incident of violence, amount of safety training both within the social work program and outside the social work program,
93 practicum site practice setting, and field educator /student race match. These factors include those listed here: Place of client violence. Students were asked to indicate where their most re cent incident of violence took place for each type of vi olence listed. If they had not experienced a particular form of client violence, t hey checked not applicable. Time of day of client violence. Students were asked to indicate the time of day tha t their most recent incident of violence took place f or each type of violence listed. If they had not experienced a particular form of client vio lence, they circled not applicable. Practicum practice setting. Students were asked to identify the type of social work setting where they had completed or were compl eting their practicum. They were asked to check only one category. Categories were t he same as those used by NASW. Field educator/student ethnicity/race match. Students were asked to identify the race of their field educator. The categories were d erived from the categories used by NASW to identify student race. Students were then d ivided into four categories, according to how their race matched with the race o f their field educator. The categories consisted of students and field educators whose rac e/ethnicities matched, White students whose field educators were of another race, non-White students whose field educators were of an other race, and students of minority race/ethnicity whose field educators were of differ ing minority race/ethnicity. Training. Students were asked to answer two questions about s afety training that they have received during their practicum experienc e. The first question asked students to indicate whether or not they have received train ing in particular content areas.
94 Additionally, they were asked to check all of the p laces that they had received training in the various content areas regarding safety and viol ence issues. This subscale contained 19 content areas that have been identified by previous researchers (Faria & Kendra, 2007; Reeser &Wertkin, 2001) as areas that field director s perceive that they have offered training to students. Total numbers of content area s in which the student had received training was totaled for each venue of training. Th e second question asked students to what extent they felt prepared to effectively deal with situations in which they may be concerned about their personal safety. Subscales: Experience with Client Violence Direct violence. Client violence questions were based on a modified version of Newhills (2003) definition of client violence. In Appendix H, Dr. Newhills permission to use portions of the questionnaire used in her st udies can be found. The following five categories of client violence were measured: actual physical assault, threat of physical assault, verbal abuse, threat of lawsuit, and damag e to personal or agency property. Participant social work students indicated the numb er of times they had experienced each type of violence during the practicum. If they did not experience a particular type of the violence during practicum, they indicated this with a zero. Indirect exposure to violence. Indirect exposure to violence was measured concerning the same types of violence that were mea sured for direct violence. Participant social work students were asked to list the number of times they had heard about and/or witnessed each of the types of violence occurring t o coworkers in the practicum or to
95 student colleagues while they had been in their pra cticum. If they had not experienced a particular type of violence, they indicated this wi th a zero. Subscale: Fear of Future Violence Fear of recurrence of violence in the workplace was measured using the Fear of Future Violent Events at Work scale. This scale was revised from a scale that was developed for use initially by Rogers and Kelloway (1997). It has also been used in other studies (Barling, et al, 2001; Schat & Kelloway, 20 00). Permission to use the scale was given by Kevin Kelloway (See Appendix H). It is a s ix item scale that evaluates the degree to which a participant is fearful of becomin g a victim of client violence in the next year (e.g. I fear that I was be physically assault ed by a client.). Participants answer on a 5-point scale from 1(strongly disagree) to 5(stro ngly agree). High scores are indicative of increased fear of future client violence. The sc ale has been found to have good validity. Initial test of the scale showed that fea r of future violence differed between persons who had been robbed and persons who had not been robbed at a bank (F (1, 185) = 5.29, p < .05) (Rogers & Kelloway, 19 97). Internal consistency for the scale was initially measured at .94. Other studies utiliz ing the same scale have found similar internal consistency scores of .95, .96, and .97 (L eblanc & Kelloway, 2002; Schat & Kelloway, 2000; Schat & Kelloway, 2003). Subscale: Occupational Commitment To measure the degree of commitment to the social w ork profession, the Occupational Commitment Scale (OCS) developed by Me yer, Allen and Smith (1993) was used, as presented in the Client Violence Quest ionnaire (Appendix C). The OCS
96 consists of 18 items to assess three subscales: Aff ective commitment (ACS), Continuance commitment (CCS), and Normative commitment (NCS). A ffective commitment refers to the participants attitudes and feelings toward the occupation and his/her identification with the professionstaying because they want to d o so. Continuance commitment refers to staying in the occupation because the participan ts need to do so. It involves tangible items such as salary and job security. Normative co mmitment refers to remaining with the occupation because the participant feels that t hey ought to do so. This scale was based on the earlier work of Meyer and Allen (1990) with an Organizational Commitment Scale. Meyer, Allen, and Smith (1993) modified the scale to be used to measure commitment to the occupation. The OCS includes six items for each subscale. The i tems are arranged on a 5point Likert scale ranging from strongly disagree scored as 1, to strongly agree, scored as 5. A higher degree of occupational commit ment is indicated by higher scores. The minimum score on each subscale is six, and the maximum score is 30. The Occupational Commitment Scale was initially pub lished by the authors in the Journal of Applied Psychology and as such, is publi c domain. The initial testing of the scale was on nurses and nursing students and wordin g of the items specifically reflected this. For the purposes of this research, all refere nces to nursing were changed to social work. Internal consistency scores on the affective commit ment subscale were as follows on the initial study: student pretest, .87; student posttest, .85, and practicing professionals, .82 (Meyer, Allen, & Smith, 1993). Normative commit ment scale internal consistency
97 scores were .73 and .77 on the two tests with the s tudents and .83 on the tests with professionals. On the continuance commitment subsca le, Cronbach alpha scores were .77 and .83 with the students pretests and posttests a nd .74 with the professional. The three components of commitment demonstrate good discriminant validity, in that they were differentially related to occupation ally related behaviors and appeared to measure three distinct domains. Subscale: Career Withdrawal Intentions Turnover intentions were initially measured by Blau (1985), using a three-item scale that had initially been used to measure organ izational withdrawal cognitions and changing the wording to reflect profession rather than job. Identical three item preand post-measures were collected on the updated ver sion of the scale in 1996 and 2000 (Blau, Tatum, and Ward-Cook, 2003). For the purpose s of this study, the words social work profession were substituted for other profess ions that Blau measured. A sample item is I intend to leave the profession of social work as soon as possible. The items were measured on a five-point scale (1 = strongly d isagree, 5 = strongly agree). Item responses were linearly summed to create a summed s core. The minimum score was 3 and the maximum score was 15. Initial internal cons istency reported was .87 and .85 (1985). Test-retest reliability was .57. Blau repo rted internal consistency of .93 in 1988. Coefficient alpha reliability estimates for the upd ated scale were .91 in 2000. Instrument Pilot Test All of the items discussed were compiled into a que stionnaire that included 64 questions on the 17 factors. The instrument was com pleted by 15 undergraduate social
98 work students in a private university in central Fl orida. All of the students were female seniors who had been in their practicum placements for 80-100 hours. Most were traditional aged college students, with two non-tra ditional aged students in their early 40s. The questionnaire was given for validity purp oses only, gleaning information about its readability, question clarity, completion time, and overall assessment. Though there was no intention to analyze the data, effort was ma de to help insure the students anonymity. Because the researcher had personal know ledge of the students, they were asked not to complete information about their age, ethnicity, and primary area of service in field placement. The time to complete the instrument ranged from 7 m inutes to 11 minutes, with a mean time of 8.73 minutes. During the discussion f ollowing the completion of the instrument, several of the participants expressed d ifficulty in knowing how to answer a question about the amount of time they had spent in safety/violence training. They felt that this question might not really convey good inf ormation about the training they had received. They believed that it would be easier for them to identify specific areas of safety/content that they had learned about. As a re sult of this discussion, the initial question was removed from the questionnaire and it was replaced by a longer question asking for feedback on specific areas of safety con tent that they had covered. The students completed the replacement question the fol lowing week and commented that it was much easier to answer the question and they wer e able to recall safety training that they had not identified the previous week.
99 Statistical Procedures Variables The dependent variables for the first part of this study were five different types of direct exposure to client violence. Exposure to cli ent violence was measured using a subscale that was a modification of a survey that w as developed for the purpose of exploring client violence to social workers (Newhil l, 2003). Social work students were asked to indicate the number of times they had expe rienced physical assault, threats of physical harm, verbal abuse, threats of lawsuit, an d damage to personal or professional property during their practicum experience. Other dependent variables included fear of future v iolence, occupational commitment, and career withdrawal intentions. These variables were measured using subscales that had established reliability and vali dity. All of these subscales were included in the questionnaire that was used for thi s study. Independent variables for predicting client violenc e included gender; age; experience; ethnicity/racial background; current so cial work program enrollment (BSW/MSW); place of violence; time of violence; pra ctice setting; racial/ethnic match of field educator and student; and amount of safety tr aining in social work practice classes, field seminar, field agencies and other places. Cl ient violence became an independent variable when measuring its effect on fear of futur e violence. Training on reporting client violence was an independent variable in reference t o actual reporting client violence. Fear of future violence was an independent variable when measuring its impact on occupational commitment and career withdrawal inten tions. Additionally, occupational
100 commitment was an independent variable when measuri ng its impact on career withdrawal intentions. Statistical Procedures Statistical methods used with the initial hypothese s included: chi square test of independence, t test for independent samples, analysis of varianc e (ANOVA), and multiple regression. To determine if fear of violence mediated the relat ionship between exposure to client violence and occupational commitment and car eer withdrawal intentions, a four step process was proposed (Baron and Kenny, 1986; K enny, 2008). Each subsequent step was contingent upon significant findings in th e preceding step. In this study, the condition for the first step required that there wa s a relationship between the dependent variables (types of direct and indirect client viol ence) and the independent variables (occupational commitment). However, since no signif icant relationships were found, the latter three steps of the mediation analysis were n ot conducted. Data were analyzed utilizing the SPSS statistical p rogram, version 16. Table 1 (Appendix A) presents the hypotheses, dependent var iables, independent variables, and the type of statistical analysis that was used for each of the hypotheses.
101 Chapter IV Results This chapter presents analysis and results of the s tudy data. Descriptive statistics will be discussed initially. The findings concernin g subscales and scales will then be discussed. The results of hypotheses testing will b e the third major section of the chapter. Finally, there is a discussion of analyses on predi ctors of client violence. Descriptive Statistics Study Sample of Social Work Students The total sample for this study consisted of 1500 s ocial work students. Cover letters and questionnaires were mailed to a stratif ied random sample of the current national student membership of National Association of Social Workers. Twenty-six of the 1500 mailings were returned undeliverable. Of t he 1474 potential respondents remaining in the sample 667 returned their survey i nstruments, yielding a response rate of 45.25%. Among them, 72 did not meet the inclusion c riteria for the study, in that they had not yet begun field practicum. There remained 5 95 questionnaires that met inclusion criteria and thus were used for this study. Personal Demographics of Social Work Student Sample Personal demographic data from this sample of socia l work students were collected and analyzed. Table 2 (Appendix A) presen ts frequencies of the demographic
102 characteristics of the sample as a whole and by BSW and MSW program. The following sections include discussion concerning gender, age, experience level, ethnicity, and current social work degree program of the social wo rk students Gender. Of the 593 social work students who responded to th is survey item, 88.2% ( n = 525) were female and 11.4% ( n = 68) were male (See Appendix A, Table 2). Two (.3%) social work students did not respond to t his question. When broken down into the social work program in which they were enrolled 91.9% ( n = 227) of the BSW students were female and 8.1% ( n = 20) were males. In its most recent summary of so cial work education statistics, CSWE (2006) reported tha t 15% of fulltime BSW students and 12% of part-time BSW students are male. Thus this s ample yielded a slightly lower percentage of BSW males than the national statistic s. This study was additionally comprised of 86.4% ( n = 293) female MSW students and 13.6% ( n = 46) male MSW students. In this case the percentage of males more closely aligned with the national percentages of male MSW students, reported to be13. 3% (full time) and 15.5% (parttime) (CSWE, 2006). Age. Five hundred fifty eight (93.7%) social work studen ts listed their age. Thirty seven (6.3%) respondents declined to re port their age. Data was compared from the cases that did not list age to cases where age was reported. T -tests and chi squares were performed as appropriate for the varia ble and it was found that there were no significant differences between those who did an d did not list age. For quantitative variables with more than 5% of the cases missing it is recommended that a method of replacement be used (Mertler & Vennatta, 2005). The option that is most commonly used
103 for replacing missing data is substituting the seri es mean, which is the mean of all available cases for the variable. In this case, the mean age was 33.1 ( SD =10.67) and the median age was 29, with the ages of the students ra nging from 20 to 64. The mean age was substituted for all missing cases. Prior to substituting the mean for missing cases, t he skew value for the age variable was 11.46 and kurtosis was 132.07. After s ubstituting the mean the skew was .884 and the kurtosis was -.244. These were closer to 0 and less than the absolute value of 2, which are considered acceptable values for skew and kurtosis (Heppner & Heppner, 2004). Prior to substituting the mean age for missing case s, social work students were grouped into age ranges in order to view data on tr aditional aged students vs. nontraditional aged students. The first age group incl uded students from the youngest to 24, which would typically be considered to be a traditi onal age range for a bachelors level college student. The next group included those aged 25-30, which might approximate the age of a masters level college student. Beyond thes e two groupings, students were grouped in ten year age spans for those in their 30 s and 40s and the final category was for those respondents 51 and up. As Table 2 (Append ix A) indicates, one hundred forty five students (24.4%) were aged 24 and below. An ad ditional 27.4% ( n = 163) of the respondents were aged 25-30, making this the larges t age group. As might be expected in a student population, the age group frequencies dim inished as age increased, with 18.8% ( n = 112) aged 31-40, 14.5% ( n = 86) aged 41-50, and 8.7% ( n = 52) over the age of 51.
104 When divided into social work degree groups, the la rgest percentage (44.5%) of BSW students was in the 24 and under group. There w ere only 13.0% of MSW students in this group. However, in the next category of tho se aged 25-30, the percentages were reversed, with 17.0% of the BSW students and 38.2% of the MSW students. In the remainder of the age groups, there were only slight ly more MSW students than BSW students. Experience. Four percent ( n = 25) of the respondents did not list years of experience. Analyses were completed to determine if there were any patterns indicated for those who did not document experience level. In all of the tests, a variable was created to indicate those reporting and those who d id not. A chi square test showed that BSW students were significantly less likely to repo rt years of experience than MSW students, 2 (1, N = 588) = 5.36, p =.02, however the effect size was small ( V = .08). Additionally, t -tests were used to check for differences between t hose who reported experience and those who did not. There were two si gnificant findings. Those who had more safety training content in field seminar ( M = 7.45, SD = 5.87) reported their experience level less often than those who had less training content ( M = 3.79, SD = 4.55), t (569) = 3.67, p .001, d = .73. Similarly, those who had more safety training content in social work classes ( M =12.45, SD = 5.74) were less likely to report experience level than those who had less training c ontent ( M = 7.38, SD = 5.23), t (570) = 4.44, p .001, d = .94. Since the percentage of missing cases was close to 5%, the series mean of years of experience was substitu ted for all missing cases. The mean was 2.75( SD = 4.64) with a range of 1 to 29 years of experienc e.
105 Prior to substitution of the mean, the skew value o f the experience variable was 23.46 and the kurtosis was 132.07. After mean subst itution the skew was 2.58 and the kurtosis was 8.50. While these were closer to the a cceptable absolute value of 2, there was concern that the kurtosis value was higher than acceptable. The variable was screened for outliers and with the use of boxplots, it was discovered that there were 16 very extreme outliers and 16 other less extreme out liers that were skewing the data. All of the significant outlier cases had over 10 years exp erience thus a decision was made to delete the 32 cases with over 10 years experience i n major analyses. When these cases were eliminated, the skew was 1.38 and the kurtosis was .925. The majority of the social work students in the sur vey had little paid social work experience. Of the 570 who responded to this questi on, 49.4% ( n = 294) had no experience. An additional 15.1 % ( n = 90) of the total sample had 1-2 years of experience. Thirty three percent of the students ha d more than 3 years experience, with 6.4% ( n = 38) having more than 10 years experience. When d ivided into degree categories, 77.7% of the BSW students had no experi ence, compared to 33.6% of the MSW students. Ethnicity/race. White social work students ( n = 425) comprised the majority (71.4%) of the respondents. In spite of efforts to oversample students of minority ethnicities, only 27.4% responded. This is slightly less than the 35.6% of all BSW students and 29.4 of all MSW students enrolled in s ocial work programs in the U. S. reported to be from minority ethnicities (CSWE, 200 6). Ninety-nine Black students (16.6%) responded to the question about ethnicity. The remaining ethnic groups were
106 represented by less than five percent of the total number of students. There were 4.3% ( n = 25) Latino/Hispanic students, 3.6% ( n = 21) of mixed heritage, and 2.2% ( n =13) Native Americans. The least frequently reported rac e/ethnicity (.8%) was Asian/Pacific Islanders ( n = 5). Current social work degree program. A total of 340 (57.1%) of the survey respondents were MSW students, whereas 248 (42.2%) were in the process of earning a BSW degree. Seven (1.2%) students did not list thei r current degree program. Of the MSW students who responded to the survey 31.6% had 250 to 749 hours in field practicum. An additional 37.8% had 500 or more hour s at their field sites. Few MSW students (7.4%) had less than 250 hours of experien ce in the field. In contrast, 75% of the BSW students had between 250 and 749 hours in their practicum and 18.9% had less than 250 hours in the field. Social Work Students Practicum Characteristics Home visits. Because there was a hypothesis concerning the possi bility that violence occurred in differing amounts in different settings, a preliminary question was asked concerning the number of home visits the stud ents made during their practicum. As Table 3 (Appendix A) indicates, almost half (46.4%, n = 274) of the respondents said that they did not make any home visits. Of the students who made home visits, the majority (56%, n =177) made 11 or more home visits while eighty fou r (26.6%) made between one and five home visits and 17. 4% ( n = 55) made between six to ten visits. Similar percentages of MSW and BSW students made home visit s. Half (49.6%, n =167) of the MSW students and 42.3% ( n = 105) of BSW students had made no home visits. On the
107 opposite end of the continuum, almost one third of MSW students (28.8%, n = 97) and BSW students (31.5%, n = 78) made 11 or more home visits during their pra cticum. Students making 11 or more home visits were signifi cantly more likely to be working in children/youth/child protection settings 2 (30, N = 582) = 1.25, V = .27. Twenty nine percent ( n = 51) of those making 11 or more home visits were in children/youth/child protection settings. This was followed next by 15.3% ( n = 27) of the student workers who were in mental health/psychiatr ic settings, and 13.6% ( n = 24) who were in family service settings. These percentages contrast greatly with the 2.3% ( n = 4) of students in alcohol/substance abuse work who mad e 11 or more home visits during their practicum. Of those students in children/youth/child protecti on settings, only 22.2% ( n = 23) had never made a home visit, whereas 49.5% ( n = 51) had made 11 or more visits. Those students in developmental disabilities work (33.3%, n = 5) and family services work (34.6%, n = 18) were the next smallest groups of students to make no home visits. In contrast, 78.6% ( n = 33) of students in alcohol/substance abuse setti ngs had never made a home visit. Work during evening hours. One of the hypotheses stated that students would experience differing amounts of violence depending on the time of day. Therefore, an initial question was asked about the percentage of time the students worked during evening hours. One third of the respondents (33.3%, n =198) stated that they did not work any evening hours, but the largest category of students (42.7%, n = 254) indicated that they completed 1-25% of their practicum during evening hours (See Appendix A,
108 Table 3). A much smaller percentage (23.3%, n = 139) of the students worked evening hours more than 25% of the time. Among those who worked at least some evening hours, almost thirty five percent ( n = 8) of students in corrections/criminal justice w orked 51-100% evening hours. This group was followed by 31.3% ( n = 10) of students in alcohol/substance abuse setti ngs, and 27.3% ( n = 12) in family services settings. By way of contr ast, only one (5%) of the school social worker students worked more than 50% of hours in the evening. Of those who completed 76-100% of their practicum in the eve nings, 80% ( n = 24) were MSW students and 20% ( n = 6) were BSW students. All of these students were female. The majority (62.1%, n = 18) were White. Practicum practice settings. Social work students were asked to select one practice setting that best described where they wer e completing their internships. The three largest practice settings were children/youth /child protective services, with 17.5% ( n =104) of the students, psychiatric/mental health w ith 17.0% ( n = 101), and medical/healthcare services (11.9%, n = 71). Six settings were designated by between 5 to 10 % of the students. They were: family services (8.7%, n = 52); school social work (7.9%, n = 47); alcohol/drug/substance abuse services (7.1% n = 42); service to the aged (6.9%, n = 41); community center/organization/planning (6.6 %, n = 39); and corrections/criminal justice (6.4%, n = 38). Four categories were listed by less than 5 % of the students: developmental disabilities (2.5%, n =15); occupational/vocational (1.0%, n = 6); group services (.7%, n = 4); and public assistance/welfare (.3%, n = 2). Twenty
109 five (4.2%) listed other as their practice settin g. Eight (1.3%) people did not complete this question. In order to do further analysis with this variable, three practice settings were merged with the other category, because the small numbers in those categories prohibited further analysis: group services ( n = 4); occupational/ vocational ( n = 6); and public assistance/welfare ( n = 2). A chi square was then done to determine the percentages of MSW and BSW students in the various practice areas (See Table 3, Appendix A). Some significant differences were foun d, 2 (10, N = 580) = 31.477, p .001, V = .23. For instance, higher percentages of the MSW students could be found in hospital/medical care settings, alcohol/drug/sub stance abuse work, schools, community organizations, schools and psychiatric/mental healt h settings. The difference was most marked with psychiatric/mental health work, where 2 2.3% ( n = 75) of MSW students and 10.7% ( n = 26) of the BSW students worked. Phrased differen tly, 74.3% ( n = 75) of those in mental health work were MSW students and 25.7% ( n = 26) were BSW students. Higher percentages of BSW students could be found i n work with developmental disabilities, corrections/criminal justice, childre n and youth/child protection, and family services. Work with children and youth/child protec tion work was the area with the most marked difference with 23.0% ( n = 56) of the BSW students in this practicum setting versus 14.0% ( n = 47) of the MSW students. Race/ethnicity of the field educator. One of the research hypotheses related to having the same or different race field educator, s o a question was asked about the ethnicity of the field educator. The majority (63.4 %, n = 377) of the field educators were
110 White, while 26.1% ( n =133) were of minority ethnicities. Black field ed ucators comprised the largest minority group (14.1%, n = 84) with Latino/Hispanic instructors making the second largest group (3.0%, n =18). The remaining ethnic groups were represented by smaller numbers: 13 (2.2%) mixed her itage; 10 (1.7%) Native American; and 6 (1.0%) Asian/Pacific Islanders. Eighty seven (14.6%) students did not list their f ield educators ethnicity. This is very near the 15% break point at which consideratio n should be given to dropping the variable from the analysis (Mertler & Vannatta, 200 5). However, the variable was retained to assess the effect of field educator rac e on exposure to violence as no other client violence studies have asked this question. A variable was created to trace patterns of those who reported field educator race versus th ose who did not. This variable was then used in t -tests and chi square, as appropriate per variable. A chi square test showed a significant difference between the race of the stud ents who did and did not report race 2 (1, N = 588) = 4.069, p .05, V = .08. Thirty (18.4%) of the minority students di d not report race of field educator versus 51 (12.0%) Whi te students. Additionally, independent groups t -tests revealed that those students who listed the race of their field educator were significantly more likely to report that changing careers would be hard, t (584) = .178, p .05, d = .24, (Race missing: M = 2.96, SD = 1.32, Race listed: M = 3.28, SD = 1.33), changing careers would be costly, t (103.999) = -2.04, p = .026, d = .27 (Race missing: M = 3.24, SD = 1.46, Race listed: M = 3.59, SD = 1.28), and changing careers would be a sacrifice t (583) = -2.13,
111 p = .033, d = .25 (Race missing: M = 2.86, SD = 1.46, Race listed: M = 3.20, SD = 1.36). These are three of the five items on the continuanc e commitment scale. Not surprisingly then, the same pattern was found with the summed co ntinuance scale. Those who listed field educator race ( M = 16.48, SD = 5.11) had stronger continuance commitment than those who did not list field educator race ( M = 15.23, SD = 4.97), t (583) = -2.071, p = .039, d = .25. A final pattern emerged concerning missing data pat terns for field educator race. Concerning amount of safety training received, an i ndependent groups t -test showed that those who listed race of field educator ( M = 3.72, SD = 4.19) had significantly less training in social work field seminar than those wh o did not list race ( M = 5.15, SD = 5.81), t (94.45) = 2.205, p = .038, d = .32. Similarly, those who listed the race of field educator had significantly less training at t heir field agencies ( M = 6.03, SD = 5.14) than those who did not list the race of their field educators ( M = 7.50, SD = 5.39), t (570) = 2.37, p = .018, d = .28. Field educator/ Student race match. A new variable was created to reflect the race match between the field educator and student. The categories consisted of those field educator-student dyads that were of the same ethnic ity, those dyads where the supervisor was White and the student was non-White, those where the supervisor was non-White and the student was White and those where the field educator was of one minority racial/ethnic and the student as of a differing rac ial/ethnic groups. Of those reporting student and field educator race, 57.5% ( n = 342) were supervision dyads that were
112 matched in ethnicity. Almost a third (27.7%, n =165) had a mixed ethnicity dyad. Supervision dyads with a White supervisor and a non -White student made up 12.6% ( n = 75). A very similar percentage (12.9%, n = 77) of the dyads had non-White supervisors and White students. The smallest group was field educators and students who were of mixed minority ethnicities (2.2%, n = 13). There were 4 Latino students, 1 Asian student, 5 Black students and 3 student of mixed et hnic heritage represented in these dyads. This variable contained 88 (14.8%) missing cases. M issing cases would be those not reporting student race and those not reporting field educator races, most of whom would be those not reporting field educator race, a s discussed in the previous section. Safety Training Training in safety and client violence was measured by constructing a list of 17 areas that were mentioned in previous studies where field directors were asked about content areas in which they offered training. Addit ionally, students were asked whether they had knowledge about field agency safety polici es and social work program safety policies. For each of these 19 training content are as, students were asked if they had training in social work practice classes, social wo rk field seminar, field agencies, or in places another place. There were 572 students who addressed every content area in each training venue. Five others responded generally that they had eithe r had the content or had not had the content, but they did not indicate any areas where they received the training. Therefore it was not possible to include these responses in the totals per training venue. Eighteen did not respond at all to the questions about training. This combined to a total of 23 (3.9%) who had missing data about training. to examine pos sible patterns in cases with missing
113 data, variables were created to reflect those who r esponded about training in various venues and those who did not respond. Ttests and chi squares were competed as appropriate with all of the demographic variables a s well as all variables in the hypotheses. Across all training venues, there was o nly one significant finding. Chi square tests indicated that there was a significant relati onship between students race and those who did not report training in any area 2 (1, N = 588), p = .01, V = .14. Of those who did not report training 59.1% ( n =13) were of minority ethnicity and 40.9% ( n = 9) were White. Totals of training content areas. For each training venue, the numbers of content areas in which the students had received training w ere totaled. The mean number of content areas covered in social work practice class es was 7.63( SD = 5.38). The median number of content areas was 7. Forty two (7.2%) of the students had received no safety training in social work practice classes. On the op posite end of the continuum, seventeen (2.9%) had received training in every content area. When numbers of content areas were grouped to view the data more comprehensively 33.3% ( n =190) of respondents had training in five or less of the content areas in so cial work practice classes. As numbers of content areas increased, every category showed prog ressively fewer respondents who had experienced training. Almost 30 percent ( n =171) of students had covered six to ten training content areas, 19.8% ( n = 113) had training in 11 to 15 areas and 9.5% ( n = 54) had experienced training in 16 to 19 content areas. Overall, fewer students stated that they had receiv ed safety training in field seminar, with a mean number of content areas covere d being 3.92 ( SD = 4.66) of 19. Two
114 was the median number of content areas covered. A t hird (33.0%, n =190) reported no training in field seminar in any of the safety cont ent areas. When categorized, a large majority (72.2%, n = 413) of students remembered covering five or les s training content areas in field seminar. The decline of percentages of respondents in categories with higher amounts of training was notable, with 16.6% ( n = 95) having six to ten areas of training, 7.5% ( n = 43) with 11to 15 areas of training, and 3.5% wit h 16 to19 areas of training covered in field seminar. Only 4 students (.7%) had training in all 19 areas in field seminar. At the field practicum agency, 12.9% ( n = 74) reported receiving no safety training at all. The mean number of safety training content areas covered by the agencies was 6.24 ( SD = 5.20). The median of incident was 5. Only 19.9% ( n = 131) received training in more than 11 content areas and 24.7% ha d training in six to ten content areas at their field agencies. In contrast, 52.4% ( n = 300) had training in five or less content areas. Slightly more than one percent (1.5%, n = 9) had training in all 19 areas. Slightly more than half of the students (54.4 %, n = 311) had received training from other sources. The median number of content ar eas covered was 1 and the mean was 3.68 ( SD = 5.18). Two hundred sixty six (45.6%) received n o training from other sources and 27.6% ( n = 158) had training in one to five content areas. Much lower percentages of participants received training from more than five content areas. Almost 12% ( n = 68) received training in six to ten content areas; 10.0 % ( n = 57) had training in 11 to 15 areas; and 4.9% ( n = 28) had 16 to 19 areas of training from other so urces.
115 Total training in the social work program per cont ent area. For each content area, responses were summed across the training ven ues to see how many parts of the social work program (social work practice classes, field seminar, field agency) addressed the content area. The intent of this analysis was t o determine if safety content was being reinforced across the social work program. The con tent areas of knowledge of the social work programs safety policies and knowledge of t he field agencys policies were not included in these analyses because it was presumed that there would not be training in these two areas by all three parts of the social wo rk program. For example, it was not presumed that training in field seminar would cover the safety policies of individual agencies. Table 4 (Appendix A) reflects the percent ages of students who received training in given content areas in more than one pa rt of the social work program. The content area that was received by the most stud ents (17.5%, n = 100) in all three parts of the social work program was maintai ning a confident, secure demeanor. The next highest content area covered by all three parts of the program was keeping supervisor informed of ones itinerary, which was received by 16.4% ( n = 94) of the students. There were seven other content areas that were received from all three parts of the program by between10% to 16.4% of the students. Eight content areas were received by less than10% of the students from all three part s of the social work program. The content areas least likely to be received by studen ts in all three places were physical techniques for self protection (3.5%, n = 20), characteristics/life experiences of people more likely to commit violent acts (4.5%, n = 26), and physical signs that attack is imminent (5.2%, n = 30).
116 Next, percentages were calculated for students who had training in both field seminar and the field agency and secondly, in both social practice classes and field seminar. The patterns were similar, though the perc entages of students who had training in both practice classes and the field agency were higher than percentages of those who had training in both field seminar and field agency Percentage of students having training in both field seminar and field agency ran ged from 5.1% ( n = 29) to 21.5% ( n = 29) per content area. Content areas received by less than 7% of students from field seminar and the field agency were the same three co ntent areas least received by three parts of the program, as reported in the previous p aragraph, in addition to forms of mental illness associated with violent behavior (6 .1%, n = 36) and assessing history of violence in clients (6.5%, n = 37). When looking at the reverse side of this tr aining question, those students who received no training i n either field seminar or field agency in a given content area training ranged from a low of 40.0% (keeping supervisor informed) to a high of 76.7% (physical techniques for self protection). Percentages of students receiving training in both social work practice classes and their field agency were the highest percentages of the three possible combinations of training within the social work program. Yet even h ere the highest percentage of students who had received training in both places was 23.8% ( n = 136). The lowest percentage of students who received training in practice classes and the field agency was the 6.1% ( n = 35) who had received training in physical techni ques for self protection. When the percentages are viewed in terms of students who rec eived no training in both social work practice classes and their field agencies, there we re four content areas where more than
117 50% of the students had received no training in bot h areas. Additionally, there were 13 content areas where at least 27% of the students re ceived no training in both places. Individual training content in social work practice classes. Individual training content areas were explored to see the percentages of students who had training in each content area for each venue (See Appendix A, Table 5 for a complete listing and Figure 2 for depiction of the same). In social work practic e classes, the content area taught most often was where to sit when interacting with a clie nt (58%, n = 345). Only three other content areas were experienced by at least 50% of t he respondents in their social work practice classes. They were: forms of mental illnes s associated with violent behavior (55.5%, n = 330); identifying and managing feelings that can arise when working with victims and perpetrators of violence (51.1%, n = 301); and maintaining a confident, secure demeanor (52.8%, n = 214). The content area taught the least was phys ical techniques for self protection (14.5%, n = 86). Only 40.5% ( n = 241) of the students said that they had received training on the social work programs safety policies. When the percentages are reversed to reflect those who received no training in particular content areas, there were eight areas wh ere more than 60% of the respondents received no training in social work practice classe s (See Appendix A, Table 6). In ascending order they are: physical techniques for s elf protection (81.7%, n = 486); recognizing physical signs that attack is imminent (71.3%, n = 424); recording incidents of violence (68.6%, n = 408); debriefing and support after an incident / reporting the incident (64.9%, n = 386); assessing history of violence in clients (6 2.2%, n = 370); home visit safety (60.7, n = 361); and keeping supervisor informed of ones itinerary (60.5%,
118 n = 360). Individual training content in social work field s eminar. In social work field seminar, no content area was experienced by more th an 35% of the respondents. The five content areas in which the students received the mo st training in field seminar were: keeping the supervisor informed of ones itinerary (30.1%, n = 179); maintaining a confident, secure demeanor (26.7%, n = 159); where to sit when interacting with a clien t (26.6%, n =158); home visit safety (25.2%, n =150); and characteristics of high risk situations such as being in non-public, isolated pl aces (24.4%, n = 145). Less than 10% of the respondents reported having content in the foll owing areas during field seminar: assessing history of violence (8.6%, n = 51); characteristics/ life experiences of people more likely to commit violent acts (8.9%, n = 53); and physical techniques for self protection (8.7%, n = 52). As with content in social work classes, percentages of those having training in content areas in field seminar were inverted to emp hasize how many respondents did not have training in each area. In all but two of the n ineteen content areas, more than 68% received no training. In the two areas that remaine d, 62.5% ( n = 372) had not received training in understanding student rights and 65.2% ( n = 388) had not received training concerning the safety policies of the social work p rogram in field seminar. Individual training content in field practicum age ncy. At the practicum agency, the content area received in training most frequent ly was keeping supervisor informed of ones itinerary (49.1%. n = 292). Additionally, 67.3% ( n = 385) of the respondents had training on the field agencys safety policies. Stu dents also had training more frequently
119 at their agencies on debriefing and support after a n incident takes place (38.7%, n = 230) and recording incidents of violence (37.5%, n = 225) than in field seminar and social work practice classes. Between 17.5% and 33.3% of s tudents reported that they had training at the practicum agency in each of the oth er content areas. When percentages were inverted and reported as thos e who did not receive training in the practicum agency, 62% or more respo ndents did not receive any training in 15 of 19 content areas. The content area where the largest percent (76%, n = 452) of students did not receive training in the field agen cy was recognizing characteristics/life experiences of people likely to commit violent acts Other training sources. The content areas in which the highest percentage o f students received the most training outside of prac tice classes, field seminar or the field agency were verbal de-escalation (25.2 %, n = 150) and physical techniques for self protection (25.9 %, n = 154). Other areas with percentages over 20% were : characteristics of high risk situations (23.9%, n = 142); recognizing verbal acts of violence (23.5, n = 140); forms of mental illness associated with vi olent behavior (23.0%, n = 137); maintaining a confident, secure demeanor (22.5%, n = 134); and physical signs that attack is imminent (21.8%, n =130). Several students added in comments on where they received other training, which included: prior or p resent social work employment sites outside of the social work program; speakers; volun teer work; psychopathology class; personal therapy sessions; police seminars; and con tinuing education units. Additional analyses concerning training. Additional analyses were completed to compute the level of training according to some of the demographic variables. A t -test
120 was used to test whether BSW students or MSW studen ts received more training. This revealed a significant difference between the two g roups in amount of training in field seminar, t (466.441) = 3.78, p .001, d = .32 and social work practice classes, t (563) = 5.23, p .001, d = .44. LSD post hoc tests revealed that BSW studen ts received more training than MSW students in field seminar (B SW: M = 4.79, SD = 4.93; MSW: M = 3.28, SD = 4.34) and social work practice classes (BSW: M = 9.00, SD = 5.50; MSW: M = 6.65, SD = 5.07). In contrast, MSW students received more training than BSW students from sources other than social work practice classes, field seminar and the practicum agency, other training so urces t (562.95) = -5.753, p .001, d = .46, (MSW: M = 4.66, SD = 5.66; BSW: M = 2.30, SD = 4.09). ANOVA computations were performed to check for diff erences in training per age category, using the categories of below 25, 2530, and above 30. Students aged 25-30 had significantly less training ( M = 6.41, SD = 4.85) in social work practice classes than students under 25 ( M = 8.22, SD = 4.74) and students over30 ( M = 8.05, SD = 5.86), F (2, 567) = 5.84, p = .003, 2= .02. Additionally, students under 25 had signific antly less training ( M = 1.69, SD = 3.11) in places other than the social work progr am than students 25-30 ( M = 3.56, SD = 4.85) or students over 30 ( M = 4.79, SD = 5.86), F (2, 569) = 17.76, p .001, 2 = .06. Total training in different training venues was al so compared according to level of experience. Students with no experience had signifi cantly more training ( M = 4.15, SD = 4.70) in field seminar than students with 1-2 ye ars ( M = 2.60, SD = 3.74), or those with 5-10 years experience ( M = 2.92, SD = 3.81), F (3, 510) = 3.63, p = .013, 2 = .05.
121 Students who received more training from other trai ning venues were significantly more likely to have 5-10 years experience ( M = 6.62, SD = 6.37) than no experience ( M = 2.20, SD = 3.96), 1-2 years experience ( M = 3.64, SD = 4.77), or 3-5 years experience ( M = 4.99, SD = 6.37), F (3, 511) = 18.92, p .001, 2 = .17. Males were significantly more likely to receive tr aining in field agencies ( M = 8.44, SD = 5.26) than females ( M = 5.93, SD = 5.06), t (512) = -3.44, p = .001, d = .17. Additionally, males were significantly mor e likely to have more total training in all venues ( M = 23.76, SD = 12.81) than females ( M = 20.31, SD = 12.36), t (509) = -1.937, p = .05, d = .28. Practice settings were compared as to the amount o f field agency training that was received by the students. Students in criminal just ice settings ( M = 8.09, SD = 5.98) and students in children/youth/ child protection settin gs ( M = 7.55, SD = 5.17) were significantly likely to have increased training at the field agency than students in alcohol/substance abuse settings ( M = 5.44, SD = 4.78), community organization settings ( M = 4.44, SD = 4.77), and school social work settings ( M = 5.57, SD = 4.48). Additionally students in mental health settings ( M = 6.93, SD = 5.02) had significantly more training at their field agencies than students in community organization settings, F (10, 497) = 3.03, p = .001, 2 = .06. Extent prepared to deal with violent clients. A final question related to training was whether students felt prepared to deal with vio lent or potentially violent clients. Less than half (44.2%, n = 247) of the students felt mostly or fully prepare d. Forty two (7.1%) said that they felt fully prepared and 36% ( n = 205) related feeling mostly prepared. In
122 contrast, a slightly larger percentage (55.9%, n = 312) said that they felt somewhat or not at all prepared. Two hundred thirty four (41.9%) sa id that they felt somewhat prepared, while 14% ( n = 78) said that they were not at all prepared to d eal with violent clients. There were 36 (6.1%) students who did not answer th e extent prepared question. A variable was constructed to measure tho se who responded to this question versus those who did not respond. Chi square and t -tests were performed as appropriate to the variable to determine any missing data patte rns. The only significant t -test indicated that those who did not respond to the ex tent prepared question ( M = 4.77, SD = 3.88) had less safety training content in social work classes than those who did respond to the question ( M = 7.64, SD = 5.35), t (13.087) = -2.613, p =.021, d = .53. Extent prepared relationships with other variables Chi square tests were performed with categorical variables and extent pr epared to check for significant relationships. The first significant test showed th at males feel significantly more prepared than females to deal with violent clients, 2 (3) =16.087, p = .001, V = .017. Sixty five percent ( n = 42) of males felt fully or mostly prepared, vers us 41.5% ( n = 205) of the females. Another test indicated that MSW students f eel more prepared than BSW students, 2 (3) = 8.971, p = .030, V = .13. One hundred fifty five (48.6%) MSW students felt at least mostly prepared, whereas 38.2% ( n = 89) of the BSW students felt equally as prepared. A further chi square test suggested that students who are 24 years of age and below feel the least prepared to deal with violent clients, 2(15) = 25.164, p = .048, V = .12. As students ages progressed, so did their p erception that they were prepared to deal with violent or potentially violent clients. T he exception to this was that the students
123 who were over the age of 50 felt less prepared than students 25-50, though more prepared than the youngest group. A final significant chi square test indicated that those students with more paid social work experience had progressively more feeli ngs of being prepared to deal with violence, 2 (15) = 56.63, p .001, V = .19. For example, 77.4% ( n = 24) of the respondents who had 11 or more years experience fel t fully or mostly prepared, whereas of those students who had no experience 34.4% ( n = 97) felt prepared at the same level. Chi square tests on practice setting and student ra ce were not significant. Pearsons correlations were performed between total s of training content areas received in each training venue and the extent to w hich students felt prepared. For every venue of training, total training units was negativ ely related to the extent of feeling prepared. The extent prepared variable was scored w ith a 1 indicating being fully prepared, so lower scores indicate more preparednes s. Therefore, feelings of preparedness to deal with violence is significantly correlated with training in social work practice classes ( r = -.291, p .001), field seminar ( r = -.198, p .001), field agency ( r = -.246, p .001), other training ( r = -.300, p .001) and total training in all venues ( r = -.408, p .01). Another Pearsons correlation was completed between types of occupational commitment and extent prepared. The only correlatio n among the types of commitment was affective commitment, the desire to be part of a profession. Affective commitment is significantly negatively correlated with perception s of preparedness ( r = -.137, p .01). With the reminder that lower scores on extent prepa red mean feeling more prepared, the
124 implications is that as students feel more prepared to deal with violence, they have higher affective commitment, which is commitment due to de sire to be part of a profession. Prevalence of Direct Client Violence against Social Work Students Client violence was the primary dependent variable for most of the hypotheses in this study. Social work students were asked to writ e the number of times they have directly experienced five types of violence during their practicum. The numbers were then summed for a total number of direct encounters with any type of violence. Prevalence rate of direct incidents of client viol ence. Of the 589 social workers who responded to the questions on direct client vio lence, 248 had experienced client violence, meaning that the prevalence rate for soci al work students experiencing violence during their practicum was 41.7% (Appendix A, Table 7). Conversely, 57.8% ( n = 347) of the students had not directly experienced violen ce. Prevalence rate by types of violence. The most common type of violence experienced directly by this sample of social work students was verbal abuse (37.5%, n = 223), as indicated in Table 8 (Appendix A). The next most prevalent type of violence was threat of physical harm, with 84 (14.1%) being victimized in this manner. Fifty six (9.4%) of the students had been threatened with a l awsuit and 43 (7.2%) had experienced damage to property. The smallest category of client violence was physical assault, with 21 (3.5%) of the students reporting assault. Total incidents of direct exposure to client violen ce. Two hundred forty eight social work students (41.7%) recorded 1591 incident s of direct exposure to client violence (Appendix A, Table 9). The mean number of incidents was 2.69. The standard
125 deviation was 21.17, indicating wide variance in th e number of incidences. One person recorded 500 incidents of verbal abuse, which contr ibuted to the maximum number of total violent incidents being 506, since this indiv idual also experienced other types of violence. Transformation of direct violence variables. Because of a few students reporting extreme high numbers of various types of violence, the mean number of incidents was not indicative of the overall samples experience w ith direct violence and the skewedness and kurtosis values were unacceptably high. Six me thods were attempted to reduce the effect of the outliers. Log transformations and inv erse transformations were attempted, as is sometimes recommended for markedly positive skew values. Though the skew and kurtosis were substantially reduced, the skew and k urtosis values were still over 2 and the violence variables continued to deviate from normal distributions. Attempts were also made to eliminate cases over 3 and then 4 standard deviations. These transformations also reduced the skew and kurtosis, but they were still above the absolute value of two, the distributions were still not normal, and cases were lost that could have related valuable information. Another attempt was made to truncate a ll of the direct violence variables at given number of incidents, as had been done by at l east two other researchers (Beaver, 1999; Song, 2005) of social work violence. There co ntinued to be unacceptable values of skewedness and kurtosis. Since all of the methods f ailed to help the distributions achieve normality and acceptable skew and kurtosis, a decis ion was made to choose a method where the skew and kurtosis was reduced substantial ly by changing the deviant scores so that they were not as deviant. It should be noted t hat though this effort decreased the
126 deviancy of the skew and kurtosis, they continued t o be slightly higher than a favorable skew and kurtosis. This method had the benefit of r eflecting the essence of what the extreme cases reported while reducing the impact of these outliers (Tabachnick & Fidell, 2007). Mertler and Vannatta (2005) suggest that in sample sizes over 100, outliers may be considered as those scores that are more than 4 standard deviations from the mean. Ultimately, a decision was made to reduce the numbe r of client violence incidents in all cases with incidences in excess of 4 standard devia tions to a number that was the highest number before 4 standard deviations plus one. For e ach type of direct violence, a new variable was created to reflect this transformation The adjusted types of violence variables were summed to create the adjusted total direct violence variable. These variables were used for all other analyses througho ut the study, unless otherwise noted. After reducing the number of incidents as appropria te for each type of violence, the total number of direct violence incidents was 1104, with a mean of 1.86 and standard deviation of 4.66 (Appendix A, Table 9). Quantities of direct incidents by type of client v iolence. Total numbers of incidents per type of violence were recorded (See A ppendix A, Table 10). The numbers for each type of violence were divided into the tot al number of incidents overall. For example, for physical assault, there were 46 incide nts which were divided by the 1591 total incidents previously discussed. This gave a p ercentage (2.9%) of total incidents that were accounted for by physical assault. This type o f violence occurred the least frequently ( M =.08, SD =.86). With a range of 0-7 incidents, threat of la wsuit comprised 5.22% ( = 83) of the total incidents. Likewise, damage to p ersonal and professional
127 property made up 5.5% ( = 87), with incidents ranging from 0-10 ( M =.15, SD = .72). There were 167 incidents of threats of physical har m, ranging from 0 to 10 incidents person. This made up 10.50% of the total overall in cidents ( M =.28, SD =.90). Finally, the largest category of direct incidents of violenc e was verbal abuse, with 1208 incidents recorded, ranging from 0 to 500. This high number o f incidents made up 75.92% of all of the direct violence incidents, with a mean of 2.04 and standard deviation of 20.75. Again, because of a few high numbers of incidents i n each category of client violence, the total number of incidents, means, and standard deviations were inflated and there was a possibility that they might not accurat ely reflect the experience of the total population. When variables were adjusted by reducin g numbers of incidents over 4 SD the percentage of verbal abuse was reduced slightly to 68.75% ( = 68.75), while all other categories of violence had slightly higher pe rcentages of incidents than those percentage with the non adjusted numbers. Threat of lawsuit accounted for 7.15% ( = 79) of incidents, while damage to property accou nted for 6.97% ( = 77), so these two reversed order, but other than this, the types of incidents with adjusted variables were in the same order as with non adjusted variables. T able 10 (Appendix A) contains number of incidents, means, standard deviations, and range for the adjusted variables. The percentages of incidents accounting for each type o f violence are also illustrated in Figure 3 (Appendix A). Prevalence of Indirect Client Violence towards Soci al Work Students Indirect exposure to client violence was an indepen dent variable in the section of hypotheses that dealt with establishing fear as a m ediator. Social work students were
128 asked to write the number of times they had heard a bout or witnessed five types of client violence occurring to a coworker or a fellow studen t during their practicum. The numbers were then summed for a total number of indi rect exposures to any type of client violence. Prevalence rate of indirect exposures to client vio lence. The majority of social work students recorded that they had either heard a bout and/or witnessed violent or aggressive events committed by clients toward their co-workers or fellow students. Since the number of students who responded to this series of questions ( N = 587) was almost identical to the number who responded to questions on direct occurrence of violence ( N = 589), it was possible to relate the percentages on each of the types of violence for direct and indirect violence. As the Table 11 (Appe ndix A) indicates, 60.2% ( n = 361) experienced violence vicariously. This was almost 2 0% more than those who experienced direct violence. Prevalence rate by types of indirect exposure to vi olence. With every type of violence, indirect exposure was higher than direct experience of violence. As Table 12 (Appendix A) shows, the highest occurrence of indir ect exposure was verbal abuse which occurred to 54.1% ( n = 322) of the students. Similar to the pattern of direct experience with client violence, the next highest category of indirect exposure was threat of physical harm (36.8%, n = 219). This is two and half times the rate of dir ect exposure to threat of physical harm (14.1%). Indirect exposure to threat of lawsuit and physical assault were experienced by the same amount of students (23.4%, n = 139). Indirect exposure to physical assault occurred almost 7 times more often than direct physical assault, whereas
129 indirect exposure to threat of lawsuit occurred 2 times more than direct threat of lawsuit. Finally, 20.5% ( n =122) of the students had been indirectly exposed to personal or professional property damage. Total incidents of indirect exposure to client vio lence. There were 4110 incidents of vicarious exposure to client violence by 361 stu dents. Total incidents per person ranged from 0 to 151. The mean number of indirect i ncidents of the entire sample was 7.05( SD = 19.97). This was almost 3 times the mean of dire ct violence ( M = 2.66, SD = 21.15). As with direct incidents of violence, there were so me students who recorded extreme numbers of indirect violence exposures, whi ch caused an extreme positive skew of 9.429 and kurtosis (128.49). Since these outlier s were likely to inflate the numbers of incidents and the mean number of incidents, the cas es that exceeded 4 standard deviations were kept in the sample, but were reduced. This all owed them to contribute to the data, while reducing their extreme effect. With these cas es reduced there were 3603 total incidents of indirect exposure to violence, with a mean of 6.18 ( SD = 12.64) and a range of 0-97 (See Appendix A, Table 13). Quantities of indirect client violence by type of c lient violence. Social work students in the sample reported that they had heard of or witnessed more incidents of verbal abuse than any other type of violence (See A ppendix A, Table 14). The total number of verbal abuse incidents was 2060, which wa s 50.63% of the total number of indirect incidents of violence. The mean number of incidents was 3.56 ( SD =11.15) and incidents per person ranged from 0 to 200. The next highest number of indirect violence
130 incidents was threat of physical harm which ranged from 0 to 100 incidents per person. There were a total of 857 incidents, comprising 20. 85% of the total indirect incidents ( M = 5.31, SD =12.26). The other three types of indirect exposur e occurred in similar quantities. There were 425 incidents of physical as sault (10.34% of total incidents, M =.72, SD = 2.21), with one student knowing about 24 inciden ts. Almost ten percent of the incidents (9.7%, = 399) were vicarious knowledge of damage to person al or agency property ( M =.68, SD = 2.75) and 8.98% ( = 369) were indirect exposure to threats of lawsuit ( M =.63, SD = 2.07, Range 0-20). After reducing the numbers of cases over 4 standard deviations, as discussed on p. 126, the pe rcentages of each type of violence were very comparable to the percentages prior to reducin g the cases. (See Table 14 and Figure 4, Appendix A). Fear of Future Violence Reliability and sampling adequacy. The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy and the Bartletts test of spheri city (BTS) were conducted to check for the appropriateness of a factor analysis. The K MO evaluates for the sampling adequacy which should be greater than .50 for satis factory factor analysis to proceed. The KMO on this scale was .839. In addition, the Ba rtletts test should be also significant to ensure the adequacy for factor analysis. The BTS was significant at the .000 level (2 = 1824.32). Thus both scores indicate that there i s a good factor and the factor consists of appropriate attributes. After explorato ry factor analysis (see below), this scale was tested for internal consistency and had Cronbac hs alpha score of .91.
131 Results of exploratory factor analysis. To test for unidimensionality of the fear of future violence scale, an exploratory factor analys is was completed. Principal components analysis was conducted using Oblimin rot ation, which is recommended if there is a prior belief that the underlying factors are correlated (Mertler & Vannatta, 2005). The fear construct was measured by five item s (FFV1 to FFV5). The analysis retained one factor which accounted for 70.71% of t he variance. The component matrix showed that all of the five items loaded onto one f actor, with the lowest score being .901(fear of future threat of harm) and lowest be ing .785(fear of future lawsuit). Thus none of the factors loaded at less than .40. A dditionally, communalities scores ranged from .617 to .811, thus none were below the accepted value of .50. Therefore all of the items remained in the scale for all bivariat e and multivariate analyses. Fear of violence descriptive information. Fear of future violence was measured by asking students if they had fear that each of the t ypes of violence would occur in the next year and then totaling the scores in all areas for a total fear of violence score. For mean scores of each type of violence, skew, and kurtosis see Table 15 in Appendix A. Overall, students demonstrated that they have moder ate amounts of fear of future client violence. Students responses indicated that they had more fear of verbal abuse than other types of violence. Almost half (47.9%, n = 285) noted that they were at least moderately afraid that this would occur to them in the next year. The next highest concern was fear of future threat of harm, with 28. 5% ( n = 170) in agreement that this might occur. Nearly one quarter (23.2%, n = 138) were fearful that they might be threatened with a lawsuit within the coming year. O ne hundred four (17.8%) of the
132 students strongly or moderately agreed that they ha d fear of being physically assaulted within the next year and a similar amount (16.5%, n = 98) strongly or moderately agreed that they were fearful of damage to their property. A total of 264 students (36.2%) either strongly or moderately feared that they might exper ience some type of client violence in the next year. Occupational Commitment Reliability and sample adequacy. The KMO test for measuring sampling adequacy resulted in an acceptable value of .88, wh ich is greater than the cut off score of .60. Additionally the BTS was significant at the .0 00 level (2 (153) = 4819.80). These scores indicate that there is a good factor and the factor consists of appropriate attributes. Therefore, it was appropriate to proceed with an ex ploratory factor analysis. Exploratory factor analysis. To test for dimensionality of the occupational violence scales, an exploratory factor analysis was completed. Principal components analysis was conducted using Oblimin rotation with the eighteen items in the occupational commitment scale. As discussed previou sly, the Oblimin rotation was most appropriate to use because of prior belief that the re were correlations between three first order factors. The analysis initially retained thre e factors which accounted for 59.38% of the variance. This raised some question about accep ting only three factors because a general rule of thumb is to retain factors that acc ount for 70% of the variance. However, the scree plot also confirmed retaining three facto rs. Additionally, after the rotation, the pattern matrix demonstrated a unique relationship w ith no overlapping among the three factors. Prior to the analysis, Cronbachs alpha wa s obtained to check for internal
133 consistency of the scales. The affective commitment (AC) subscale had a score .65, the normative commitment (NC) subscale had .74, and alp ha for the third subscale, continuance commitment (CC), was .83. Thus affectiv e commitment and normative commitment subscales have lower than the ideal reli ability, though acceptable. Affective commitment (AC). The affective commitment construct was measured by six items (AC1 to AC6). The principal components factor analysis showed that the communality of AC1 (Social work is important to my self image) was .302, which it lower than suggested .50. After dropping this item, Cronbachs Alpha for this scale was .91. This is .26 higher than the original Alpha sc ore, thus substantially improving the reliability of the subscale. Normative commitment (NC). The normative commitment construct was measured by six items (NC1 to NC6). The exploratory factor analysis showed that lower than the suggested .50 communality was found for 2 items: NC1 (I am in social work because of sense of loyalty to it), communality = 48 and NC6 (I believe people who have been trained in a profession have a responsibi lity to stay in that profession for a reasonable amount of time), communality = .42. Aft er dropping these two items, Cronbachs Alpha was reassessed at a level of .82. This is .11 higher than the original Alpha and substantially improves the reliability of the subscale. Continuance commitment (CC). Six items (CC1 to CC6) measured the construct of continuance commitment. Exploratory factor analy sis revealed that one item, CC5 (There are pressures to keep me from changing prof essions), had a communality of .42, which is lower that the suggested .50. After droppi ng this item and recalculating the
134 scale, reliability was measures at .81. Though this is slight drop from the original reliability of.83, it remains an acceptable level o f internal consistency. As indicated in Table 16 (Appendix A), after adjust ing the three subscales, a new principal component analysis was run to check again for dimensionalities of the occupational commitment subscales. Again, three fac tors were extracted. However, the three factors now accounted for 66.77% of the varia nce, which was closer to the 70% that is the general rule of thumb for retaining factors. Because of this and the fact that the two of the three subscales had substantially higher Cro nbachs Alpha scores when recalculated, a decision was made to report scores for the recalculated subscales, with a total of four items dropped from the original occup ational commitment scale. Additionally, the recalculated subscales scores wil l be used for all bivariate and multivariate analyses. Occupational commitment descriptive information. Three subscales measured occupational commitment. The first was affective co mmitment, which is a type of commitment that expresses pride in the profession. In this case, it would demonstrate a desire to be a social worker. Each of five items wa s scored on a 5-point Likert scale with a 5 indicating that the student strongly agreed wit h the statement. Means, standard deviations, skew and kurtosis for each scale can be seen in Table 17 of Appendix A. The strong affective commitment to the profession of so cial work was striking. The mean score was 23.35 ( SD = 2.93) out of a total score of 25. Sixty percent ( n = 356) of the social work students scored a 5 on every question. For each of the following individual statements, students said that they strongly agreed : 73.0% ( n = 432) satisfied with their
135 career choice; 77. 2% ( n = 457) proud to be in the social work profession, 80.0% ( n = 476) like being a social worker, 74.8% ( n = 443) identified with the social work profession, and 73.8 ( n = 437) enthusiastic about being a social worker. The second type of occupational commitment was norm ative commitment. With this type of commitment students were asked about t heir level of agreement with statements that indicated that they should or ought to be committed to the social work profession. In short, they have feelings of obligat ion. With two items deleted from the scale, four items were left, with a maximum score o f 20. The mean score was 11.69 ( SD = 4.18). In contrast with affective commitment, no rmative commitment had the lowest scores of the three types. Only 4.0 ( n = 24) scored a 5 on every question. With this type of commitment, students were more prone to dis agree with the statements. For three of the four categories, at least two thirds of the students either disagreed or were neutral with the statements. The statements included I fee l obligated to remain in the social work profession, I would feel guilty if I left th e social work profession, and Even if it were to my advantage, I do not feel it would be rig ht to leave social work now. Only one of the four statements received more agreement than disagreement or neutrality. It was I feel a responsibility to the social work profession to continue. Continuance commitment was the third type of occupa tional commitment. This type of commitment speaks of staying in a professio n because the alternative would be too costly in every way. One has come too far to tu rn back now. With one item dropped from this subscale, five items remained, for a top score of 25. The mean score for continuance commitment was 16.30 ( SD =5.10), thus continuance commitment fell
136 almost halfway between affective commitment and nor mative commitment. For each of the five statements, the largest category of respon ses was moderately agree. Students stopped short of strongly agreeing with statements. For instance, for the following two statements, students either strongly or moderately agreed: I have put too much into the social work profession to consider changing now. ( 52.2%, n = 321) or Changing professions now would require considerable personal sacrifice. (45.9%, n = 173). Overall, students conveyed a higher than a medium l evel of continuance commitment. Career Withdrawal Intentions Career withdrawal intentions were measured with thr ee questions. The first indicated thoughts about leaving the profession, th e second indicated an intention to look for a new profession, and the third indicated an in tention to stay in the social work profession for an extended time. Disagreement with the first and second statements indicated a stronger likelihood of staying in the p rofession whereas disagreement with the third statement indicated a high likelihood of leav ing. The third item was reverse scored before the items were totaled. Therefore, lower sc ores on this scale indicate low intentions to seek another profession. With 5 point Likert questions, the maximum score on this scale is 15 and the minimum is 3. The mean score on this scale was 4.40 ( SD = 2.49), indicating very low levels of career with drawal intentions. Mean scores of the individual items in the scale can be seen in Ta ble 18 (Appendix A). Three hundred seventy seven (63.4%) scored a one on each item, in dicating the maximum insistence on staying in the social work profession. In contrast, 18 students (3%) scored in the top third of scores, indicating plans to leave the profession
137 Findings Related to Prevalence of Client Violence H ypotheses The first section of hypothesis testing addresses the prevalence of client violence as affected by ten independent variables or series of variables. In all of these hypotheses, the dependent variables are the five types of viole nce which are physical assault, threat of physical harm, verbal abuse, threat of lawsuit, and property damage. The last dependent variable is a sum of the five types of direct viole nce. The initial findings from each hypothesis analysis are reported according to commonly accepted values for probability levels. Th at is, findings that are at the .05 and .01 probability levels are reported as trending tow ards significance. Since numerous hypotheses were investigated, Bonferroni adjustment s were warranted (Montcalm & Royse, 2002). Bonferroni procedures typically requi re the researcher to divide the .05 probability level by the number of statistical test s that are conducted. Because there were 13 hypotheses tested in this study, .05 was divided by 13, yielding a new probability level of .004 required to declare statistical significanc e. Hypothesis 1-1 Male social work students will experience more of e very type of client violence than female students. Male social worker students experienced direct viol ence at a higher rate than females in every type of client violence. In fact, for every type of violence, the rate of male social work students experiencing direct clien t violence was almost twice the rate of female social work students. For the overall preval ence rate during practicum, 39.3% ( n = 205) of the female students encountered violence whereas 61.6% ( n = 42) of the male students were directly affected by client viol ence. This was a significantly higher
138 rate of violence for males than females, 2 (1) =12.50, p .001, V = .15. Mantel-Haenszel common odds ratio estimate is 2.498, indicating tha t males are 2 times more likely to encounter direct client violence. Results of the ch i square analyses that yielded the rates of violence for each type of violence can be viewed in Table 19 (Appendix A). Males and females were compared as to means of inci dence and total incidents. Males had a higher mean of total incidents of 4.35 ( SD = 8.81) whereas females had a mean of 1.55 ( SD = 3.70). This is portrayed on a chart in Figure 5 (Appendix A). For every type of violence, males had a higher mean of incidents. Females experienced a total of 807 incidents, accounting for 73.1% of the total of incidents, whereas males encountered 296 incidents, comprising 26.8% of all the incidents. The number of incidents was again noticeably higher for females i n the area of physical assault, where they experienced 20 incidents, versus males experi ence of 8 incidents. T-tests for independent samples were used to check for differences in males and females direct encounters with violence. The tests showed a probability level less than .05 for direct verbal abuse, t (69.65) = -2.20, p =.031, d = .54 with males experiencing more verbal abuse than females (Males: M = 3.09, SD =7.57, Females: M = 1.05, SD = 2.93). An additional finding at less than the .0 5 probability level was found for the total of direct violence incidents. Males experienc ed more total violence than females, t (70.11) = -2.45, p = .017, d = .63, (Males: M = 4.32, SD = 8.27, Females: M = 1.59, SD = 3.91). These two findings approached significance After the Bonferroni adjustment was completed, stat istical significance was found for threats of physical harm, t (71.95) = -3.24, p = .002, d = .64, with males ( M =.69,
139 SD = 1.21) experiencing more threats than females ( M =.21, SD = .66). The t -tests did not show a statistically significant difference at the .004 level between males and females experiences with physical assault, verbal abuse, th reat of lawsuit, damage to property, or total of client violence. Full results of the t -tests are given in Table 20 of Appendix A. Hypothesis 1-1 is accepted for the following types of violence: Direct threatened physical harm Hypothesis 1-1 is rejected for the following types of violence: Direct physical assault Direct verbal abuse Direct threatened lawsuit Direct damage to property Total direct client violenceHypothesis 1-2 Younger social work students will experience more o f every type of client violence than older social work students. To study prevalence of client violence by age, youn ger students were defined as traditional aged students. Since it was found that significantly more BSW students are aged 24 and under and MSW students are significantl y more likely to be aged 25-30, these two categories of age were used separately to demonstrate traditional aged students. All students over 30 were in the third group. When direct violence incidents are summed across all categories of violence, those students w ho are aged 25-30 have the highest rate of violence (45.7%, n = 74), with 41.8% ( n = 119) of students 31 and older experiencing direct violence, and students 24 and under having t he lowest rate of violence, at 37.9%
140 ( n = 55). Students aged 25-30 had the highest rate of violence for threat of physical harm, verbal abuse, and threat of lawsuits. The rate of p hysical assault was highest for those 24 and under, whereas property damage occurred most of ten to those who were 31 and over. Results of the chi square tests that yielded these rates of violence are listed in Table 21 (Appendix A). An ANOVA was performed to further assess the differ ences between these three groups. As indicated in Table 22 (Appendix A), the only type of violence that approached significance was threat of lawsuit, F (2/ 589) = 4.007, p = .019, 2 = .013. LSD post hoc tests indicate that more students in the 25-30 age range ( M = .22, SD = .64) experienced threat of lawsuit than students 31 and over ( M = .11, SD = .45) and students aged 24 and under ( M = .08, SD = .31). Additional indicators of the frequency of violence were the number of incidents per age group and mean number of incidents per grou p. For this measurement the first two categories of age remained the same, but in ord er to better discriminate the experiences of those 31 and over, this group was di vided into three groups: those 31-40, 41-50 and 51 and over. The highest category of summ ed direct violence incidents was in the 25-30 ( n = 162) group, with 386 incidents totaling 34.1% of the total incidents of violence. The mean of the 25-30 group ( M = 2.38, SD = 5.87) was the same as the 41-50 group ( M = 2.38, SD = 7.10). Total incidents in the 41-50 group was 20 2, occurring to 85 people. The lowest mean of incidents ( M = 1.36, SD = 2.87) occurred in the 24 and under group ( n = 145), with a total of 198 incidents occurring.
141 A chart was constructed to demonstrate the differen ces in means of client violence incidents between those under 25, those 25-30, and those over 30. As shown, the students in the age 25-30 group experienced the highest mean of violence (Appendix A, Figure 6). The hypothesis was tested through the computation o f the correlation between age and incidents of violence. There was no statistical significance for any type of violence. Hypothesis 1-2 is rejected for every type of violen ce when measuring age continuously. Hypothesis 1-3 Less experienced social work students will experien ce more of every type of client violence than social work students with more paid social work experience. Rates of violence were calculated for each type of violence per years of experience. The experience variable had missing dat a substituted with the series mean. Additionally, per boxplots, all cases with over 10 years of experience were outliers that could affect the analysis, so cases with over 10 ye ars experience were eliminated from the analyses. to discover rate of violence, experie nce was grouped by years, with the first group having no experience, the second group having 1-2 years, the third 3-5 years, and the fourth 6-10 years. As indicated in Table 23 (A ppendix A), those students having 3-5 years of experience had the highest rate of violenc e (41.5%, n = 34). Rates of violence decreased progressively as years of experience decr eased. Those with no experience ( n = 118) and 1-2 years of experience ( n = 36) had a client violence rate of 40.4%. Client violence per experience level was additional ly measured by comparing means and numbers of incidents. Students with 3-5 y ears of experience ( n = 66) had the highest mean of incidents ( M = 2.85, SD = 7.24), experiencing a range of 0 to 57 incidents for a total of 234, or 25.7% of the total incidents. Students with no experience
142 ( n = 292) had the lowest mean of incidents ( M = 1.29, SD = 2.66). This group experienced 376 incidents which accounted for 35.2% of all the incidents. This is portrayed graphically in Figure 7 (Appendix A). Pearson correlation coefficients were computed to a ssess the relationship between years of experience and every type of direct incide nts of violence. As seen in Table 24 (Appendix A), experience was found to be positively related to verbal abuse ( r = .087, p =.04) and the sum of direct client violence ( r = .09, p = .03), with probability levels of less than .05 trending toward significance. Years o f experience did not approach significance with any other type of violence. To further analyze if specific levels of experience contribute to client violence, one way ANOVA tests were performed to check for sig nificant differences in incidents of violence among the four levels of experience. Ta ble 24 (Appendix A) shows that those students who had 3-5 years of experience ( M = 2.07, SD = 6.22) had more verbal abuse than those with no experience ( M = .88, SD = 1.87), F (4/525) = 3.01, p = .03, 2 = .02, with the probability level of .05 approaching significance. Similarly, students with 3-5 years experience ( M = 2.85, SD = 7.24) had more total violence than students with no experience (M = 1.29, SD 2.26), F (4/525) = 3.45, p = .02, 2 = .02. After the Bonferroni procedure, there were no statistically significant differences between the experience groups. To further explore characteristics that may be asso ciated with having 3-5 years experience, further chi square analyses were conduc ted. Experience and practice settings were found to be significantly associated, 2 (30, N = 525) = 49.07, p = .02. With 3-5
143 years experience, the highest percentages of studen ts were in schools (28.6%, n = 12), medical/health care settings (24.6%, n = 15), alcohol/substance abuse settings (20%, n = 7), and mental health psychiatric settings (19.6 %, n = 18). Additionally students degree program was strongly significantly related t o years of experience, 2 (3, N = 529) = 1.01, p .001, V = .44. Almost 21% ( n = 63) of MSW students had 3-5 years experience versus 7.9% ( n = 18) of BSW students. Experience categories were a lso strongly and significantly related to age categorie s, 2 (6, N = 532) = 79.29, p .001, V = .27, with those students who are aged 25-30 (27% n = 43) being much more likely to have 3-5 years experience than students under the a ge of 25 (4.3%, n = 6) or students over the age of 30 (14.5%, n = 34). Though no statistical significance was fo und, a substantially higher percentage of males (22.8%, n = 13) had 3-5 years experience than females (14.8%, n = 70). Also, higher percentages of Asians, Native Americans, Black students, and students of mixed heritage had 3-5 ye ars experience than White students. Additionally, the largest percentage of those makin g the most home visits and working the most evening hours had 3-5 years experience. Hypothesis 1-3 is rejected for every type of client violence. Hypothesis 1-4 There will be no difference in exposure to all type s of client violence by students of various ethnicities/racial backgrounds. The occurrence of client violence among social work students in the sample was analyzed to determine the rates per student race. T he highest rates of total violence were all with non-White ethnicity groups. The highest wa s 63% ( n = 12) in the mixed heritage group, followed by 60.0% ( n = 3) of Asian/Pacific Islanders, and 53.8 ( n = 7) of Native
144 Americans. This contrasted with the 40.9% ( n = 174) Whites, 40.8% ( n = 40) Blacks, and 36% ( n = 9) Latinos/Hispanics. This pattern was consisten t for all types of violence, as is indicated in Table 26 (Appendix A). This table disp lays the chi square analyses results that yielded the client rates for each type of viol ence. When non-White students other than Black students are clustered together in one g roup, they still have the highest rate of violence (49.3%, n = 34). Figure 7 (Appendix A) shows that Native American so cial work students had the highest mean number of client violence incidents ( M = 7.69, SD = 16.47). Students with mixed heritage had the next highest mean of 3.16 ( SD = 3.92), followed by Asian students ( M = 2.20, SD = 3.80). The students with the least mean of incid ents were Latino/Hispanic with a mean of 1.24 ( SD = 2.01). Black and White students had a similar mean number of incidents with means of 1.68( SD = 3.67) and 1.72 ( SD = 4.15) respectively. Again, since the minority groups with small representation made up the groups with highest means, all non-White students e xcept Black students were combined into a group to measure them with White and Black s tudents. Even when combining the groups ( n = 64), they had a mean of 2.98 ( SD = 7.76) incidents, which was a higher mean than Black students or White students. The largest percentage of incidents occurred to Whi te social work students, who experienced 66.4% ( = 733) of the total incidents. This was followed by non-Whites other than Black with 18.7% ( = 206), then Black students, who had 165 incidents (14.9%).
145 An ANOVA test was used to test the hypothesis. Race /ethnicity groups were compared regarding their encounters with violence. Results showed that the relationship between damage to personal and professional propert y and the ethnicity of students trended toward significance, F (5/589) = 2.59, p = .025, 2 = .02. According to LSD post hoc tests, property damage occurred more often to s tudents with mixed heritage than Black students, Latino students, or White students. Furthermore, Native American respondents had more property damage incidents than Black respondents. After the Bonferroni adjustments, results indicated that the ethnicity of the students and incidents of violence were significant ly related for each of the following types of violence: verbal abuse, F (5/579) = 5.426, p .001, 2=.05; threat of lawsuit, F (5/579) = 3.54, p = .004, 2 = .03; and total of direct violence, F (5/ 579) = 4.26, p = .001, 2 = .06. The results of all of the ANOVA tests conce rning student ethnicity are displayed in Table 27 (Appendix A). Further LSD pos t hoc tests were computed to determine which groups are significantly different. For verbal abuse, Native Americans had a significantly higher mean of violence than al l other groups. Social work students with mixed heritage had significantly more threats of lawsuits than all other groups except Asians. Finally, for the total number of dir ect violence incidents, Native American students experienced significantly more client viol ence than every other ethnic group except Asians. Because Native American students and students of m ixed ethnic heritage were found to experience significantly more of various t ypes of client violence, further analyses were completed to discover more about othe r characteristics that might be
146 related to student race. Two factors were found to be significantly related to student race. Age was the first, 2 (10, N = 588) = 28.99, p = .001, V = .167. A majority (76.9%, n = 10) of Native American students were over the ag e 30, versus 15.4% ( n = 2) who were aged 25-30, and 7.7% ( n = 2) who were under the age of 25. More than half of students (57.1%, n = 12) with mixed ethnic heritage were found to be over 30, compared to 23.8% ( n = 5) aged 25-30, and 19.0% ( n = 4) under the age of 25. Comparatively, less White students were over the age of 30(41.9%, n = 178) and more were under the age of 30. The second statistically significant finding wa s that Native American students (23.1%, n = 3) were more likely to work 51-75% evening hours than students of all other races, most particularly White students (5.2%, n = 22), 2 (20, N = 584) = 31.51, p = .049, V = .116. Though none of the other analyses were stat istically significant, it was found that a higher percentage of males (4.5%) were Nativ e American, versus 1.9% of females who were Native American. Additionally, a higher pe rcentage of BSW students (2.9%) were Native American, compared to 1.8% of MSW stude nts who were Native American. Finally, higher percentages of students with mixed heritage were in alcohol/substance abuse settings (21.1%) and children/youth/child pro tection settings (21.1%). Native American students were more commonly in medical/hea lth settings, alcohol/substance abuse settings, child/youth/child protection, schoo ls, and mental health settings, with 15.4% in each setting.
147 Hypothesis 1-4 is accepted for all racial minority groups for the following types of violence: Physical Assault Threat of physical harm Damage to personal and professional property Hypothesis 1-4 is rejected for the following types of violence per racial/ethnic group: Verbal abuse Native Americans Threat of lawsuit Mixed heritage Total direct violence Native Americans Hypothesis 1-5 BSW social work students will experience more of ev ery type of client violence than MSW students. MSW students (24.6%, n =144) experienced a higher rate of violence than B SW students (17.4%, n =102). Additionally MSW students experienced a hig her rate of threats of lawsuit than the BSW students, 2 (1) = 4.431, p = .035, V = .02, with this probability level approaching significance. The Ma ntel-Henszel common odds ratio is 1.902, meaning that MSW students are almost two tim es more likely to experience threats of lawsuit than BSW students. MSW students additionally experience higher rates of verbal abuse and threats of physical harm, while BSW students experience a higher
148 rate of physical assault and property damage. The r esults of all the chi square analyses that yielded rates of violence are displayed in Tab le 28 (Appendix A). Comparison of the means of incidents revealed that MSW students experienced a higher mean of violent incidents ( M = 2.20, SD = 5.89) than BSW students ( M = 1.42, SD = 3.23). BSW students recorded a total of 370 inci dents, ranging from 0 to 31 incidents per person and accounting for 32.8% of th e total incidents. MSW students recorded more than twice the number ( =757) of incidents that BSW students did. Number of incidents per person ranged from 0 to 62 and totaled 67.2% of the total incidents. This is shown in a chart form in Figure 8 (Appendix A). A t -test was computed to test the hypothesis, looking for significant differences in the mean score of MSW and BSW students. After the B onferroni adjustment, it was found that MSW students experienced significantly m ore threats of physical harm than BSW students, t (535.70) = -2.77, p = .004, d = .22, (BSW: M = .17, SD = .98, MSW: M = .33, SD = .98). Additionally, the total direct violence i ncidents were higher for MSW students ( M = 2.20, SD = 5.57) than BSW students ( M = 1.42, SD = 3.02), t (547.72) = -1.96, p =.031, d = .17, with the probability level below.05 approac hing significance. Verbal abuse also approached signifi cance at the .05 level, t (532.82) = -1.94, p = .053, d = .10, with MSW students experiencing more verbal abuse ( M = 1.52, SD = 4.60) than BSW students ( M = .96, SD = 2.39). No other type of violence trended toward statistically significant f indings. All t -tests results are shown in Table 29 (Appendix A). Hypothesis 1-5 is rejected for every type of type o f violence.
149 Hypothesis 1-6 There will be a significant difference in numbers o f client violence incidents occurring to social work students according to the place of social work practice (home visit, office, other). For each type of violence, respondents were asked t o indicate where the most recent incident occurred. The choices of places wer e home visit, office and other. As noted in Table 30 (Appendix A), social work student s reported the highest rate of physical assault in other places (47.1%, n = 8). The other four types of violence occurred at highest rates in an office setting. A l ittle more than 54% ( n = 44) of threats of physical harm, 51.4% ( n = 109) of verbal abuse, 58.8% ( n = 30) of threats of lawsuit, and 70.3% ( n = 26) of damage to personal or professional proper ty were all reported in an office. For these four types of violence the rates of violence in an office setting was more than twice the rate in home visits or other places. When all of the places are summed across all types of violence, the rate of most rec ent violent incidents taking place in an office was 54.43% ( n = 148), with an additional 23.36% ( n = 64) taking place at home visits, and 22.62% ( n = 62) taking place in other places. To calculate the total number of incidents per plac e of violence, the numbers of most recent violent incidents were summed for the p lace that the violence occurred. It was necessary to create a new variable for each pla ce of violence category (home visit, office, and other), as the SPSS program was not c apable of creating one variable in which data was summed for each of three categories. Once the three places of violence variables were calculated the data was then manuall y summed across the three variables to get a total amount of incidents and to calculate percent of recent incidents that
150 occurred at each place. There were 274 students wh o recounted 416 recent incidents in all of the categories of violence. In other words, 274 students recorded that they had a recent incident of at least one type of violence. T he other 143 had experienced a recent incident of 2 or more types of violence. One hundre d eighty four students recalled 227 incidents that had occurred in an office setting, a ccounting for 72.15% of all incidents reported and a mean of 1.53( SD = .73) incidents. Students recounted 64 total rece nt incidents that occurred at home visits, which made up 23.35% of the total incidents. The mean of incidents that occurred at home visits was 1.47 ( SD = .69). Similarly, 62 students reported experiencing 96 incidents at places other than an office or home visit setting. The mean of incidents occurring at other places was 1.55 ( SD = .78). The wording of the questions about place of violenc e was such that the students recorded only the place for the most recent inciden t of a particular type of violence. With each student listing one incident per type of viole nce, there was no variability in responses. Because of this only descriptive analysi s could be conducted for the place of client violence of each type of violence. This mean t that it was not possible to use inferential statistics to analyze the data using th e variables that were initially designed to compare the place of client violence. Though the hypothesis could not be directly explore d, a related question was analyzed with an ANOVA test to determine if social work students who make more home visits experience a higher mean number of violent i ncidents. The groupings of home visits were 1-5 home visits made during practicum, 6-10 visits made, and 11 or more
151 home visits. As Table 31 (Appendix A) suggests, thr eats of lawsuit occur more often when students make 11 or more home visits, F (2/313) = 4.57, p = .011, 2 = .03, ( M = .25, SD = .70) compared to those making 1-5 visits ( M = .06, SD = .24) or those making 6-10 visits ( M = .07, SD = .33). Total incidents of client violence were al so higher for those students who made more home visits F (2/313) = 3.646, p = .027, 2 = .01. The same pattern existed. Of those who make home visits, groups of students who make increased numbers of visits experience hig her total incidents of violence. For both of these types of violence variables, the find ings trended toward significance. However, after the Bonferroni adjustments, none of the types of violence were significantly related to making increased home visi ts. Hypothesis 1-6 could be neither accepted nor reject ed for any type of client violence. Hypothesis 1-7 There will be a significant difference in numbers o f client violence incidents occurring to social work students according to the time of day. To address this hypothesis, students were asked to check what time of day their most recent incident of violence had occurred for e ach type of violence. If they had not experienced that type of violence, they marked not applicable. Rates of violence were calculated for each type of violence as to the time of day that the most recent incident had taken place. Rates of violence per time of day are reflected in Table 32 (Appendix A) Category choices for time of day were daytime hours (8 a.m.-5 p.m.), evening hours (5p.m.-12a.m.), or early morning hours (12a.m.-8a.m .). Those students who had not experienced violence were omitted from this analysi s. For every type of violence, the largest percentage of violence had by far occurred during the daytime hours. The
152 percentages of violence in the daytime ranged from 71.4% ( n = 25) of the property damage to 82.4% ( n = 42) of the threatened lawsuits. Violence rates i n the early morning hours ranged from 0% of physical assaults to 3.9% ( n = 2) of threatened lawsuits. The highest rate of violence in the evening hours occur red with verbal abuse (26.5%, n = 56). Evening violence rates for other types of violence, in descending order, were: 25.7% ( n = 9) of property damage, 24.4% ( n = 19) of threatened harm, 23.5% ( n = 5) of physical assaults and 13.7% ( n = 7) of threatened lawsuits. To calculate the total number of incidents per time of day, the numbers of most recent violent incidents were summed for the time o f day in which incidents were reported to occur. It was necessary to create a new variable for each time of day category (daytime, evening, and early morning), as the SPSS program was not capable of creating one variable in which data was summed for each of t hree categories that are contained within one variable. Once the three times of day va riables were calculated the data was then manually summed across the three variables to get a total amount of incidents and to calculate percent of recent incidents that occurred at each time of day. There were 264 students who recounted 416 recent incidents in all of the categories of violence. In other words, 264 students recorded that they had a recent incident of at least one type of violence. The other 152 had experienced a recent in cident of 2 or more types of violence. One hundred eighty four students recalled 298 incid ents that had occurred in the daytime, accounting for 72.15% of all incidents reported and a mean of 1.62 incidents. Students recounted 69 total recent incidents the evening hou rs, which made up 24.69% of the total
153 incidents. The mean of incidents in the evening hou rs was 1.48 ( SD = .76). Comparatively, the rate of violence that occurred i n the early morning hours (4.16%, n = 11) was more than 6 times less than violence in the evening hours and less than 17 times the violence in the daytime. The wording of the questions about time of day of v iolence was such that the students recorded only the time of day for the most recent incident of a particular type of violence. Because of this only descriptive analysis could be conducted for the time of day of each of the types of violence. This meant that it was not possible to use inferential statistics to analyze the data using the variables that were initially designed to compare the time of day of incidents. Though it was not possible to inferentially compare times of day of most recent client violence per type of violence using the time of violence variables, another item in the questionnaire allowed a limited analysis of tim e of day of violence. This question asked the percentage of time that the student worke d during evening hours. It should be noted that for this question evening hours were not specifically defined. There were five possible categories: no practicum hours completed d uring evening hours; 1-25% of practicum hours during evening hours, 26-50% in eve ning hours, 51-75% during evening hours, and 76-100% during evening hours. Though thi s question did not directly address the hypothesis, it indirectly addressed the premise of the hypothesis because of the implications of the data. For example, if a student worked 1-25% of their practicum hours during evening hours, it was implied that they work ed 26-100% of their hours during hours other than evening hours. As can be seen in T able 33 (Appendix A), using this
154 variable, the rates of violence were very similar r egardless of the amount of evening hours the students worked. Those who worked 76-100% of their time during evening hours had the lowest rate (36.7%, n = 11) of summed direct violence. The next lowest rate of violence (38.1%, n = 75) was among those students who worked no eveni ng hours. The highest rate of violence, 46.5% ( n = 20), was among students who worked 51-75% of their practicum hours during the evening. For every type of violence, slightly different patterns existed. Those students who work ed 51-75% evening hours had the highest rates of physical assault, verbal abuse and property damage, while those that worked 26-50% evening hours had highest rates of th reatened physical harm. Threats of lawsuit were more prevalent for students working 76 -100% of time in evening hours. ANOVA tests were performed to determine if each typ e of violence was more likely to take place with groups of students who wo rked no practicum hours in the evening, 1-25% of their practicum hours in the even ing hours, 26-50% in evening hours, 51-75% in evening hours or 76-100% of their time in the evening. Only one type of violence, physical assault, had a probability level that was barely at the p .05 level. Students who worked 51-75% evening hours ( M =.16, SD = .57) experienced more physical assaults than those working no evening hou rs ( M =.04, SD = .23), those working 1-25% evening hours ( M = .05, SD = .28), those working 26-50% evening hours ( M =.02, SD = .12) and those working 76-100% evening hours (0% ), F (4/583) = 2.34, p = .054, 2 =.02.
155 After Bonferroni adjustment, there were no statisti cally significant findings for any types of violence. Results of all ANOVA tests f or types of violence per percentage of evening hours worked can be seen in Table 34 in App endix A. Hypothesis 7 is rejected for every type of violence Hypothesis 1-8 There will be a significant difference in numbers o f client violence incidents occurring to social work students according to the practice setting. Social work students at three practice settings had a total violence rate over fifty percent. Students in psychiatric/ mental health pra ctice settings had a 54.5% ( n = 15) rate of violence, those in developmental disabilities ha d a 53.3% ( n = 8) rate of violence and those in alcohol and substance abuse services had a 52.5% ( n = 22) rate. The lowest rate of violence (20.5%, n = 8) was in the community planning/community organ izing practice settings. The chi square results on the ra tes of violence per type of violence in practice settings are reflected in Table 35 in Appe ndix A. There were differing rates of violence in practice settings depending on the types of violence. For physical assault, the highest rat e of violence was in developmental disabilities (13.3%, n = 2), followed by services to the aged (7.3%, n = 3), and alcohol/substance abuse services (7.1%, n = 3). Threatened physical harm occurred to 23% ( n = 23) of the students in mental health services. S tudents in developmental disabilities followed closely with a 20% ( n =3) rate and school social work students had a similar rate of 19.6% ( n = 9). Verbal abuse occurred to the most students i n alcohol/substance abuse service (50%, n = 21), mental health services (47%, n = 47), and developmental disabilities services (46.7%, n = 7). Two practice settings had rates of
156 threatened lawsuits over 10%. They were alcohol/sub stance abuse services (16.7%, n = 7) and children and youth/child protective services (1 4.4%, n = 15). Property damage occurred to the most students in alcohol/substance abuse services (14.3%, n = 6) and mental health services (14%, n = 14). Five practice settings had a mean of total client v iolence exceeding 2.0. Criminal justice students had the highest mean of violent in cidents ( M = 3.21, SD = 9.46). The next four settings clustered near the same mean: school social work students had a mean of 2.66 ( SD = 8.73); developmental disabilities social work st udents had a mean of 2.60 ( SD = 4.46); mental health service students had a mean of 2.58 ( SD = 4.36); and alcohol/substance abuse student workers had a mean of 2.57 ( SD = 5.54) incidents. Figure 10 (Appendix A) depicts the means of total v iolence for each practice setting. For these five areas, the percentage of total incidents was not proportionate to the percentage of students who worked in the practice settings. Fo r instance, 17.2% of students served in mental health settings, but they encountered 23.8% of the total amount of incidents. The highest mean of incidents for specific types of vio lence occurred in some of the practice settings already mentioned. Physical assault ( M = .27, SD = .80) and threats of harm ( M = .47, SD = 1.30) occurred most frequently in developmental disabilities settings. Verbal abuse ( M = 2.60, SD = 8.41) and property damage ( M = .24, SD = .91) occurred most often in criminal justice settings. The highes t mean of threatened lawsuits was in alcohol/substance abuse settings. ANOVA tests were used to test this research questio n. As Table 36 (Appendix A) indicates, one finding, practice setting and exposu re to threatened lawsuit, trended
157 towards significance at the .05 probability level, F (10/574) = 1.88, p = .046, 2 = .03. LSD post hoc tests indicated that the mean of incid ents was higher in alcohol/substance abuse services ( M = .38, SD = .96) than every other practice setting in the st udy. A chi square test was also used to answer the resea rch question. The analysis of the sum of direct violence per practice setting app roached significance at the .05 probability level, 2 (10) = 19.17, p = .038, V =.18. After Bonferroni adjustments were made, no statistical significance was found for any type of violence per practice setting. Hypothesis 1-8 is rejected for every type of client violence. Hypothesis 1-9 There will be a significant difference in numbers o f client violence incidents occurring to social work students according to the race/ethnicity match between the student and the field educator. A preliminary ANOVA test explored whether field edu cators of any race/ethnicity supervised students who experienced more violence o f any type. There were no statistically significant results for this question In other words, the race of the field educator did not make any significant difference in the amount of client violence experienced by the student. To more specifically address the question of whethe r the race match between the field educator and student affects the students ex posure to violence, a variable was created with four categories. The first had student and field educator dyads that were matched in race/ethnicity. The second had White stu dents and non-White field educator pairs, while the third category was made up of dyad s where the student was non-White and the field educator was White. The final categor y combined students and field
158 educators from unmatched minority groups. The fourt h group, unmatched minority dyads, incurred the highest rates of violence in ev ery category (See Appendix A, Table 37). This was a small category ( n = 13), so the percentages represent a small number of students. The lowest rate of physical assault was e xperienced by both the matched dyads (2.6%, n = 9) and the non-White student/White field educato r dyads (2.6%, n = 4). Twenty five percent ( n = 3) of the students in unmatched minority dyads e xperienced threatened physical harm, whereas students from mat ched race dyads had the next highest rate of threatened harm (14.7%, n = 50). Students from unmatched minority dyads had the highest rate of verbal abuse (50%, n = 6) whereas the next lowest group had a rate 10% less than this. Two (16.7%) of minority unmatch ed supervision dyads experienced threat of lawsuit, which was the highest rate of th reat of lawsuit among the four groups. With property damage, students from mixed minority dyads encountered at least 3 times the amount as any other group. In total, students w ho were a part of mixed minority supervision dyads experienced a rate of 53.4% ( n = 8) exposure to violence. When the three unmatched race supervision categorie s were combined to compare matched groups with unmatched groups, the matched g roups had higher rates of violence in every type of violence except physical assault. Generally, the rates of the matched race dyads were within two percentage points of the unma tched dyads. The means of client violence incidents were compare d across students in the four different groups. Students from the mixed minority dyads had the highest mean of incidents of summed violence, with a mean of 2.76 ( SD = 3.96). The next highest group was students in matched race supervision dyads, wit h a mean of 1.83 incidents
159 ( SD = 4.41). The lowest means of total incidents of vi olence were with the other two groups: students who were White with non-White fiel d educators ( M = 1.20, SD = 2.43) and students who were non-White with White field ed ucators ( M = 1.34, SD = 2.28). The means of total violence per match of student/field educator race can be found in Figure 11 (Appendix A). Supervision dyads where the stude nt and field educator were of differing minorities had the highest mean of incide nts for every type of client violence. When the three unmatched race dyads were combined t o compare means of incidences with matched dyads, students from matched race supe rvision dyads had higher means of threatened harm and verbal abuse, whereas students from the unmatched groups had higher mean of physical assault, threats of lawsuit s, and property damage. to test the hypothesis, one way ANOVA tests were c alculated to consider differences between the four groups as to incidents of violence. Results of the analyses are presented in Table 38 (Appendix A). After Bonfe rroni adjustments, the match of race between student and field educator and incidents of property damage in the sample were significantly related, F (3/501) = 4.69, p = .003, 2 = .000. LSD post hoc tests indicated property damage was experienced by significantly mo re students in mixed minority race supervision dyads( M =.69, SD = 1.49) than students with matched race supervisio n dyads ( M = .13, SD = .55), White students with non-White supervisors ( M = .13, SD = .55), and non-White students with White supervisors ( M = .05, SD = .36). Hypothesis 1-9 is accepted for the following circum stance: Students who are in mixed minority supervision dyad s who encounter property damage. Hypothesis 1-9 is rejected for all other types of v iolence.
160 Hypothesis 1-10 Social work students who have more safety training offered by their social work program will experience less of every t ype of client violence than social work students who have had less safety train ing in their social work program. This hypothesis was addressed using information abo ut the training that students had received in field seminar, social work practice classes, and the field agency. In addition a fourth category of data concerned traini ng content which students had received in places other than the social work program. Initi al analyses calculated the rate of violence for each type of violence for each trainin g venue and the mean of incidents for each type of violence per training venue. Overall, the rate of violence increased as training content increased. Rate of violence per training in social work classe s. The rate of violence for total violence incidents was highest for those who had 11 -15 content areas of 19 covered in their social work practice classes (43.3%, n = 49). The lowest rate of violence was 28.2% ( n = 12) for students who had no training in their so cial work classes. For physical assault, the highest percentage (4.3%, n = 8) of incidents was in students with 1-5 training content areas covered in social work class es. The highest rate of threats of physical harm occurred (15.5%, n = 29) with students with 1-5 training content area s and 11-15 content areas. Verbal abuse, threat of lawsui t and property damage all occurred at the highest rate in students with 11-15 content are as in practice classes. The means of incidents were compared between the gr oups having different amounts of training. For the sum of client violence incidents, those who had 11-15 safety training areas covered in practice classes had the highest mean( M = 2.37, SD = 6.41).
161 Those with 16-19 content areas had slightly lower m ean of incidents. The lowest mean of incidents ( M = 1.12, SD = 2.32) was among those students who had no safety training in social work practice classes. Rate of violence per training in field seminar. In field seminar those with the most training, 16-19 content areas (out of 19), had the highest rate of totaled violence incidents (55%, n = 11). Again, the group with the lowest rate was students who received no safety training in field seminar. Threats of physic al harm and property damage occurred to the most students in the 11-15 training content group. Verbal abuse occurred at the highest rate with students with the most training ( 50%, n =10). Only physical assault and threats of lawsuits occurred at highest rates to st udents with less training hours. Physical assault occurred at the highest rate (3.7%, n = 7) to students with no training in field seminar and threats of lawsuits happened at the top rate (19.0%, n = 8) with students who had 6-10 training content areas covered in field se minar. When means of incidents were compared, social work students had the highest mean of incidents ( M = 2.83, SD = 6.80) when they had 11-15 safety content areas i n field seminar. The lowest mean of violence incident s was with students who had no safety training in field seminar ( M = 1.52, SD = 3.24). Rate of client violence per training at the field a gency. Social work students who received no training at their agency had the lowest rate of client violence (33.8%, n = 25). As content areas of safety training increased the t otal violence rate increased, with a top rate (51.6%, n = 16) in the group with 16-19 content areas covere d in training. Physical assault and verbal abuse occurred at highest rates to students who had 16-19 safety
162 training content areas at their field agencies. Thr eats of harm, threats of lawsuit and property damage occurred at highest rates to studen ts who had 11-15 content areas covered by the field agency. Students who had 16-19 safety training content area s covered by the field agency had the highest mean of incidents ( M = 2.97, SD = 5.78). Once again, the group of students with the lowest mean of incidents was stud ents who had no safety training at the practicum agency ( M = .97, SD = 1.57). Rate of client violence per training in other place s. For those who had training in places other than the social work program, a simila r pattern existed. The highest rate of total violence, threats of harm, and verbal abuse h appened to those with the 16 to 19 of 19 possible training areas. Physical assault, threa ts of lawsuit, and property damage occurred to students in the next to the highest cat egory of training, those with 11-15 training areas. The highest mean of incidents was 2.81( SD = 7.82) for those who had received training in places other than the social work progr am. This occurred in the next to highest group among those students who had training in 11-1 5 content areas. Hypothesis testing. To test the hypothesis Pearsons correlations were calculated between the five types of violence and the summed v iolence and total training content areas covered in social work classes, field seminar field agency and places other than the social work program. All of these correlations are recorded in Table 39 (Appendix A). For training in social work classes there were no s ignificant correlations. Though correlations werent significant, the correlations between threatened harm and threatened
163 lawsuits were negative as predicted. The other type s of violence all had positive correlations with training in social work classes. Correlations between the six categories of violence and training in field seminar were all positive, indicating that as students had more training in field seminar they experienced more violence. Two categories of violen ce trended towards significance. The first was verbal abuse ( r = .104, p = .013) and the second was the sum of violence ( r = .109, p = .009). Thus it is indicated that when training i n field seminar increases, verbal abuse and total client violence significantl y increases. Types of violence and training in field agencies ei ther had positive statistical significance or approached statistical significance for every type of violence. The following correlations trended toward significance: physical assault ( r = .109, p = .009); verbal abuse ( r = .092, p = .029); threatened lawsuit ( r = .084, p = .044); and property damage ( r = .108, p = .010). After the Bonferroni adjustment, the summ ed client violence had a positive statistical significance ( r = .128, p = .002) when correlated with total training in the field agency. The correlation betw een threatened harm and training in the field agency was also statistically significant ( r = .140, p = .001). For training in places other than the social work p rogram there were three correlations that trended toward significance, with a .05 probability level. Verbal abuse was positively correlated with total training in ot her places ( r = .084, p = .046). Likewise, property damage and training in other places had a positive correlation ( r = .096, p = .022). Finally, the summed total of violence was a lso correlated with other training ( r = .093, p = .027).
164 When all of the types of training were totaled, the re were positive correlations that were either statistically significant or trending t owards significance with every type of violence except physical assault. Two correlations approached significance at the .05 probability level. Threats of lawsuit ( r = .087, p = .039) and threats of harm ( r = .083, p = .049) had slightly lower correlations than the o ther types of violence. After Bonferroni adjustments, three types of viole nce were statistically significant when correlated with total training con tent. The totaled violence correlated with total training at .155 ( p .001). The next strongest correlation was .139 ( p = .001) between verbal abuse and total training. This was followed by property damage ( r = .120, p = .004). Additional testing was done to determine if having training in more venues per content area affects exposure to violence. Pearson s correlations were calculated for numbers of sources of training within the social wo rk program (social work practice class, field seminar, and field agency) and the six categories of violence (five types, plus summed violence). The correlation table in Appendix A, Table 40 contains all correlations between types of violence and total tr aining per content area. Of 17 content areas, five areas were positively and significantly correlated at the .004 level with total client violence: physical techniques for self prot ection ( r = 1.52, p = .000), characteristics/life experiences of people more li kely to commit violent acts ( r = .131, p = .002), verbal de-escalation techniques ( r = .128, p = .002), physical signs that attack is imminent ( r = .121, p = .004), and forms of mental illness associated w ith violent behavior ( r = .120, p = .004). One content area was correlated at the 01 level,
165 approaching significance: characteristics of high risk situations (i.e. nonpublic, isolated places) ( r = .108, p = .010). Four other areas were positively correla ted at the .05 level, again nearing significance: recognizing verbal act s of violence ( r = .087, p = .039), knowledge of office safety (arranging work spaces to maximize safety) ( r = .100, p = .017), where to sit when interacting with a cli ent ( r = .087, p = .038), and keeping supervisor informed of ones itinerary ( r = .095, p = .023). These positive correlations indicate that as training occurs in more training v enues per content area overall client violence increases. Analyses were similarly conducted for each type of violence. Verbal abuse was the type of violence with the most significant corr elations. One area of training was positively and significantly correlated with verbal abuse at the .004 level: physical techniques for self protection ( r = .136, p = .004). Correlations between verbal abuse and three other content areas trended toward signif icance. In descending order of strength of correlation they were: characteristics/life exp eriences of people more likely to commit violent acts of violence ( r = .118, p = .005), knowledge of office safety ( r = .115, p = .006), and forms of mental illness associated w ith violent behavior ( r = .110, p = .009). Four other areas additionally approached s ignificance, with a probability level of .05. They were: verbal de-esca lation techniques ( r = .106, p = .012), physical signs that attack is imminent ( r = .099, p = .019), characteristics of high risk situations ( r = .095, p = .023), and keeping supervisor informed of ones itinerary ( r = .090, p = .031).
166 Property damage had two significant correlations at the .004 level, including training on forms of mental illness associated with violent behavior ( r = .151, p .001) and physical techniques for self protection ( r = .135, p = .001). Two additional content areas trended toward significance with a probabilit y equal to or less than .01. This training was on characteristics/life experiences of people more likely to commit violent acts ( r = .118, p = .005) and physical signs that attack is imminent ( r = .106, p = .011). Two additional content areas approached significanc e, with a .05 probability level: recognizing verbal acts of violence ( r = .097, p = .020) and verbal de-escalation techniques ( r = .096, p = .022). Again, these findings indicate that when there is training in more training venues on these subjects, there is a greater likelihood that students would have experienced client violence. Threat of physical harm had no statistically signif icant correlations with training content areas at the .004 level. However, a positiv e correlation that trended toward significance, with a .01 probability level, was fou nd in one content area, physical signs that attack is imminent ( r = .111, p = .008). Training on student rights ( r = .100, p = .017), verbal de-escalation ( r = .096, p = .022), and understanding characteristics/life experiences of people more likely to commit violent acts ( r = .090, p = .032) were positively related to threat of physical harm at a .05 level that approached significance. Physical assault was positively correlated with tra ining content on physical signs that an attack is imminent( r = .109, p = .009), approaching significance at the .01 probability level. Two additional areas trended tow ard significance with a .05 probability
167 level or less: verbal de-escalation ( r = .092, p = .022), and physical techniques for self protection ( r = .101, p = .016). The type of violence with the least amount of corre lations with training content areas that approached significance was threat of la wsuit. One training area neared significance with a .01 probability level: understa nding high risk situations ( r = .111, p = .008) and one other area, keeping supervisor inf ormed of ones itinerary ( r = -.091, p = .031), approached significance with a .05 probabil ity level. Extent of preparation. A question related to safety training was the exten t to which social work students felt prepared to deal wi th violent or potentially violent clients. Pearsons correlations were done, using extent prep ared as a continuous variable. Lower scores on the extent of preparedness represent feel ings of more preparedness. Thus negative correlations indicate that as perception o f preparedness is stronger, violent incidents increase. A negative correlation that app roached significance at a .01 probability level was found between preparedness an d direct physical assault ( r = -.106, p = .012). Other negative correlations that trended towards significance at the .05 probability level were found between preparedness a nd threats of physical harm ( r = -.083, p = .050), threat of lawsuit ( r = -.100, p = .021), and sum of direct client violence ( r = -.09, p = .041). Hypothesis 1-10 is rejected for all types of violen ce Findings Related to Fear as Mediator Hypotheses The second section of hypothesis testing addresses establishing fear of future violence as a mediator between direct and indirect client violence and occupational
168 commitment and career withdrawal intentions. This s ection begins with a discussion of the conditions that are required to proceed with es tablishing mediation. This includes the section two hypotheses. It was hypothesized that fear of future violence wo uld mediate the relationship between client violence and occupational commitment and also that it would mediate the relationship between client violence and career wit hdrawal intentions. Testing this overall hypothesis required a series of hypotheses related to the steps involved with establishing mediation. The four steps required for mediation we re delineated by Baron and Kenny (1986) and further expounded upon by Kenny (2008). For each step, preconditions must be met in order to proceed with establishing mediat ion. Client Violence Relationship with Occupational Comm itment and Career Withdrawal Intentions Baron and Kenny (1988) state that the first step in establishing mediation is to show that the initial variable, in this case, clien t violence, is correlated with the outcome variable. In this case there are two outcome variab les, occupational commitment and career turnover intention, so they will be discusse d separately. The first condition for mediation was not written in the proposed hypothese s but it had to be addressed prior to proceeding to the proposed hypotheses. To test this part of the mediation path, Pearsons correlations were calculated between the six direct violence variables and six i ndirect exposure variables to client violence variables and the three types of occupatio nal commitment: affective commitment; normative commitment; and continuance c ommitment. It should be noted
169 that the occupational commitment scale is construct ed of these three subscales. Each of the three subscales represents distinctly different constructs, so it is not appropriate to sum the three scales for a total occupational viole nce score. Results of the Pearsons correlations between types of direct violence and o ccupational commitment subscales are presented in table form in Appendix A, Table 41, wh ere-as the Pearsons correlations between indirect violence and occupational commitme nt subscales are presented in Table 42. Client violence and affective commitment. Analyses with Pearsons correlations showed no statistically significant correlations be tween any type of direct violence and affective commitment. Neither do the analyses indic ate any significant correlations between any type of indirect client violence and af fective commitment. The finding indicates that there is no relationship between dir ect or indirect violence and the type of commitment reflecting an emotional attachment and i dentification with the profession of social work and a desire to remain in the professio n. Client violence and continuance commitment. A second series of Pearsons correlations was completed between direct and indir ect violence variables and continuance commitment. Again, there were no statis tically significant correlations between any type of violence, direct or indirect, a nd continuance commitment. Otherwise stated, there was no relationship between exposure to client violence and the type of commitment that reflects an obligation to stay beca use too much has already been invested to leave the profession now and perceived costs of leaving are too high.
170 Client violence and normative commitment. The last type of commitment to be tested with client violence variables was normative commitment. Normative commitment reflects a perceived responsibility to the professi on. Pearsons correlations were completed and indicated that there was no significa nt relationship between indirect exposure to client violence and normative commitmen t. Per the Bonferroni adjustments, analyses further showed that there was no statistic ally significant relationship between any types of direct client violence, though physica l assault trended towards significance with a probability level of .01 ( r = .105, p = .011). Interpreted, this would mean that when a student is physically assaulted, he/she may tend to be more strongly committed to the profession of social work because of a sense of obl igation and responsibility to the profession. Client violence and career withdrawal intentions. Pearsons correlations were calculated between both direct and indirect type of violence and career withdrawal intentions. There were no statistically significant relationships between any type of violence and career withdrawal intentions. The resu lts of the analyses can be seen in Table 41 and 42 (Appendix A). Fear of Future Violence as Mediator Since the first condition for mediation was not me t for any of the variables, it was not appropriate to proceed with the other three ste ps to establish that fear of future violence was a possible mediator between client vio lence and occupational commitment and career withdrawal intentions. However, since th e study specified that the hypotheses for these three steps would be explored, Hypotheses 2-1, 2-2, and 2-3 are analyzed here.
171 Hypothesis 2-1 Experience of client violence has a positive and di rect effect on fear of future violence. This hypothesis addressed the second precondition f or establishing fear as mediator. The condition was that the predictor var iable(s), which in this case were 12 direct and indirect violence variables, would be si gnificantly related to the proposed mediator, which in this case is fear of future viol ence. To test this hypothesis all the direct and indirect types of violence were placed in Pears ons correlation analyses with fear of future violence variables. Fear of future violence was represented in two ways. The first was with a summed fear of violence variable that re flected a general fear that any type of violence could occur in the next year. The second w as with specific questions about fear of the five types of violence used throughout the s tudy. Direct client violence and fear of future violence. After Bonferroni adjustments, Pearsons correlations between direct client violen ce and general fear of future violence were positive and statistically significant for thr eatened harm ( r = .178, p .001); verbal abuse ( r = .126, p = .002); threats of lawsuit ( r = .152, p .001); and summed direct violence ( r =.146, p .001). General fear was correlated with direct pro perty damage ( r = .084, p = .043) at the .05 probability level, with this fi nding trending towards significance. All analyses results can be seen in T able 43 (Appendix A). After Bonferroni adjustments, correlations between direct violence and three specific fears of the same type of violence were st atistically significant with even stronger relationships than with general fear of vi olence. Threat of harm was significantly
172 and positively related to fear of future threatened harm ( r = .216, p .001), threatened lawsuit was significantly and positively related to fear of future threatened lawsuit ( r = .194, p .001), and verbal abuse was significantly and posi tively related to fear of future verbal abuse ( r = .131, p = .001). Those students who had experienced physical assault reported that they have fear of physical assault ( r = .084, p = .042), with a .05 probability level that trends towards significance. Those who had experienced property da mage had fear of property damage ( r = .085, p = .040), with this finding approaching significanc e at a .05 probability level. Indirect exposure to client violence and fear of f uture violence. Even after Bonferroni adjustments, indirect exposure to client violence proved to be significantly related to fear of future violence for every type o f violence. Total indirect violence was positively and significantly correlated with genera l fear of future violence within the next year ( r = .236, p .001). When Pearsons correlations were calculated for the types of violence and general fear of future violence, all c orrelations were positive and statistically significant at the .001 level. The s trongest relationship was between threatened lawsuit and fear of future violence ( r = .270, p .001). Threatened physical harm ( r = .210, p .001) and verbal abuse ( r = .206, p .001) followed with similar relationships with fear of future violence. Even th e two types with the lowest strength of relationship with fear of future violence, physical assault( r = .168, p .001) and property damage ( r = .153, p .001), were more strongly related than types of di rect violence with fear of future violence. Analyses results of indire ct types of violence and fears of violence are noted in Table 44 (Appendix A). In su mmary, every one of the relationships
173 between indirect violence and general fear of viole nce were positive and significant and they were even stronger than the relationships betw een types of direct violence and general fear of future violence. As was the case with specific types of direct viole nce and specific fear of future violence, indirect exposures to all specific types of violence were positively and significantly related to specific types of fear of the same type of violence. Every relationship was significant at the .001 level. The strongest relationship was between indirect exposure to threats of lawsuits and fear o f future threat of lawsuit ( r = .293, p .001). Indirect threatened harm was related to fe ar of future threatened harm ( r = .232, p .001). There followed three other significant rela tionships: indirect verbal abuse to fear of future verbal abuse ( r = .212, p .001), indirect physical assault with fear of future physical assault ( r = .170, p .001), and indirect property damage with fear of future property damage ( r = .135, p = .001). Hypothesis 2-1 is accepted for every type of violen ce. Hypothesis 2-2 Fear of future violence has a negative and direct e ffect on occupational commitment. This hypothesis was tested by computing Pearsons c orrelations to assess the relationship between fear of future violence and th e three subscales of occupational commitment, which are affective commitment, normati ve commitment, and continuance commitment. Fear of future violence was measured by using the five fear variables that addressed specific types of violence and a variable for summed or general fear of
174 violence. Results of the correlation analyses betwe en fear of violence and the occupational commitment subscales are presented in tabular form in Table 45 (Appendix A). Fear of future violence and affective commitment. The first set of correlations was between the fear variables and affective commit ment. All of the correlations with affective commitment were negative as predicted. F ear of physical assault was negatively related to affective commitment ( r = -.102, p = .014), at a .01 probability level that approached significance. Fear of future threat ened harm was also related to affective commitment ( r = -.083, p = .046) at the .05 probability level that trended towards significance. However, after Bonferroni adjustments none of the relationships between fear of future violence and affective commitment we re significant. Fear of future violence and normative commitment. The next Pearsons correlations were conducted to assess relationship between fear variables and normative commitment. Even after Bonferroni adjustments, ever y correlation was positive and statistically significant at the .004 level. Overal l fear of future abuse was positively and significantly related to normative commitment ( r = .168, p .001). Of the specific types of fear, fear of future verbal abuse had the strong est relationship with normative commitment ( r = .144, p .001). The remaining relationships between specifi c types of fear and normative commitment were as follows, in d escending order: fear of physical assault ( r = .144, p = .001); fear of threats of physical harm ( r = .130, p = .002); fear of threat of lawsuit ( r = .126, p = .002); and fear of property damage ( r = .124, p = .003).
175 Fear of future violence and continuance commitment. The final set of Pearsons correlations assessed the relationships between fea r variables and continuance commitment. Again, after Bonferroni adjustments, th ere were significant, positive relationships between every type of fear variable a nd continuance commitment. Overall fear of future violence was positively and signific antly related to continuance commitment ( r = .169, p .001). Fear of verbal abuse again had the stronges t relationship with continuance commitment ( r = .174, p .001). The next strongest relationship was between fear of threat of lawsuit and continuance commitment, with a Pearsons correlation of .149 ( p .001). Other positive significant relationships cl ustered around the same number: fear of physical assault ( r = .122, p = .003); fear of threat of harm ( r = .121, p = .004); and fear of property damage ( r = .121, p = .004). Hypothesis 2-2 is rejected for every type of violen ce with affective, normative, and continuance commitment. Hypothesis 2-3 Fear of future client violence has a negative and direct effect on career withdrawal intentions. This hypothesis is the third precondition for estab lishing fear as mediator between client violence and career withdrawal conditions. It is necessary to show that the proposed mediator, fear of future violence, signifi cantly affects career turnover intentions.
176 This hypothesis was tested by calculating Pearsons correlations between the six fear of violence variables and the career withdrawa l intentions scale. There were no significant correlations for these analyses. Hypothesis 2-3 is rejected for every type of fear o f future violence. Predictor of Violence Analyses Predicting Client Violence Based on the Effect of I ndependent Variables In order to determine the best predictors of direc t client violence among social work students multiple regression analyses were con ducted to analyze the separate effects of a set of independent factors on experiencing cli ent violence. The independent variables chosen for each analysis were those that had statis tical significance with any type of violence in the previous analyses. Since many of th e initial hypotheses findings were based on categorical variables, dummy variables wer e created to represent the groups of people who had been found to experience significant ly more violence or who had higher means of client violence. The independent variables included in the analyses were: males, MSW students, students aged 25-30, students who mad e 11 or more home visits, those who worked 51-75% of their practicum during evening hours, Native Americans, Asians, students with mixed ethnic heritage, unmatched min ority supervision dyads, and students who did their practicum hours in criminal justice s ettings, school social work, children and youth settings, alcohol/substance abuse setting s, mental health settings, or developmental disability settings. Quantitative var iables were used for years of experience, total training, field seminar training, field agency training, and training in other places.
177 Methods of analysis. The direct violence variables all had extremely hig h skewedness and kurtosis due to the nature of the va riables. For example, the most frequently experienced type of violence was verbal abuse and it was experienced by 37.9% of the sample. This meant that 62.1% had expe rience no verbal assault. Thus 366 people had zero incidents of verbal abuse. Efforts were made to curb the skewedness and kurtosis through various means including log transf ormations, inverse transformations, and truncating the variables at 4 standard deviatio ns, truncating the incidents at 7 incidents, and eliminating cases over 4 standard di ssertations. None of these methods resulted in skew and kurtosis values that were acce ptable. In approaching multiple regression analysis, it was necessary to have varia bles and variable combinations that approximated a normal curve. Because of this a deci sion was made to calculate two multiple regression analyses for total client viole nce. The first calculation utilized as the dependent variable a dichotomous variable to analyz e what combinations of characteristics/ groups of students predicted wheth er or not client violence had occurred. This allowed for an analysis with all students incl uded, regardless of whether or not they had experienced client violence. After this initial analysis a second analysis was conducted using only the students who had experienc ed violence. This had the effect of decreasing the skew and kurtosis by taking students with zero incidents of violence out of the calculations. This method has been illustrated with economic costs projects where many subjects spend nothing and a few spend large a mounts (Jones et al., 2006). After taking out students with no experience of violence, there remained unacceptable values of
178 skew and kurtosis for the sum of direct violence. T his variable was transformed using a log transformation. This achieved acceptable skew a nd kurtosis values under two. Data screening. Data screening was done for each of the multivariat e combinations of all the independent variables and t he dependent variables as described above. Using data from Mahalanobis distance, the su m of direct violence continued to have multivariate outliers. With this variable, 11 cases had to be dropped as they had chi square critical values over 48.26 (as determined by chi square critical values chart for 20 degrees of freedom and a significance of p .001). Data screening showed that total training had an u nacceptable tolerance rate of .000 due to using the other variables that were use d to compute the total rates. A decision was made to delete the total training variable from all multivariate analyses and to keep the training totals from each of the four training venues. Screening for multicolinearity. Correlations between the other 18 independent factors were established through the us e of Pearsons correlations. The MSW group was significantly correlated with 10 of t he other factors, though the most significant correlation (with years of experience) was .368 ( p = .000). The remainders of the MSW correlations were under .232. Being male co rrelated with 4 factors, of which the strongest relationship was with field agency tr aining ( r = .130, p =.003). There were 8 significant relationships among the 6 practice sett ings that were used for independent variables, with the strongest correlation between t he school setting and children and youth setting ( r = -.135, p = .001). There were two significant correlations b etween making 11 or more
179 home visits and practice settings, with the stronge st correlation with children and youth settings ( r = .181, p = .000). The unmatched minority supervision dyads overlapped with two of the minority groups, with the highest correl ation with students of mixed heritage ( r =.180, p = .000). There were 2 significant correlations bet ween years of experience and practice settings, with the highest correlation wit h students in mental health settings ( r = .106, p = .017). The variables with the strongest relation ships were the training variables. There was a positive significant relatio nship between total training received in social work classes and field seminar ( r = .473, p = .000). Additionally there were significant correlations between total field agency training and total social work class training( r = .259, p = .000) and total field seminar training ( r = .370, p = .000). In summary, there were several statistically signi ficant correlations among the twenty variables. However, there is a relatively sm all shared variance in all of these significant correlations, except for training in so cial work class and training in field seminar, which would be considered a moderate co-va riance. Therefore, the correlations create limited multicollinearity and do not affect the multiple regression models. Assumptions of multiple regression analyses. Though every effort was made to achieve normality, the violence variables continued to show a significant deviance from normality, per KolmogorovSmirnov tests. However, Tate (as cited in Mertler & Vannatta, 2005) has stated that moderate violations of the normality assumption may be ignored, particularly with a large sample size, sin ce there are no adverse effects on the analysis. Tabachnick and Fidell (2007) agree that w ith large samples the significance level of skewedness and kurtosis is not as importan t as the size of the skew and kurtosis
180 and they further state the underestimates of varian ce with positive kurtosis disappear with samples of 100 or more cases. Visual screening of the residual scatterplots revea led some degree of non-linearity as well as possible heteroscedasticity. However, Me rtler and Vanatta (2005) state that moderate violations weaken the regression analysis, but do not invalidate it. Multiple Regression Analyses Results The result of the multiple regression analyses are presented here beginning with the sum of violence and proceeding through the five different types of violence. As stated above, two multiple regression models are presented for the sum of client violence. The first model addresses the prediction of the rate of direct client violence and the second model predicts the frequency of direct client viole nce. For each specific type of violence only the first multiple regression could be calcula ted as there were not sufficient numbers of people who had experienced each type of violence With the exception of verbal abuse, the amount of people who experienced each type of v iolence did not meet the minimum number required to conduct multiple regression with the number of predictor variables that were being used (Green, 1991). Predicting sum of client violence. A stepwise multiple regression was conducted to determine which of the independent var iables predict occurrence of client violence. Since the dependent variable was dichotom ous there were no cases that had to be eliminated. Regression results indicate that the overall model significantly predicts occurrence of client violence, R2=.063, R2 adj = .056, F (4/504) = 8.47, p .001. The model accounts for 5.6% of variance in occurrence o f client violence. A summary of
181 regression coefficients is presented in Table 46 (A ppendix A) and indicates that only four of the seventeen variables, male gender, working in a mental health/psychiatric setting, making 11 or more home visits, and working in an al cohol/substance abuse setting, significantly contributed to the model. A second stepwise multiple regression was conduct ed to determine the accuracy of the 18 independent variables predicting the freq uency of client violence. Note that this analysis was conducted using only those students wh o had experienced client violence and that the sum of violence variable, already tran sformed to reduce cases over 4 standard deviations, was transformed again via a lo g transformation. Even so, eleven cases had to be excluded due to a chi square over t he critical value. After all combinations of variables, there remained 194 stude nts in this analysis. Native Americans, Asians, and unmatched minority supervisi on dyads were deleted from the computation because they had missing correlations. The results indicate the model significantly predicts frequency of client violence R2 = .034, R2 adj = .029, F (1/192) = 6.68, p = .011. The model accounts for 2.9% of the varianc e in frequency of client violence. Regression coefficients are presen ted in Table 47 (Appendix A). Only being male significantly contributes to the model. Predicting of physical assault. A stepwise multiple regression was computed between the dependent variable (dichotomou s variableno physical assault vs. at least one physical assault) and the 18 independe nt variables. Because there was no variability in the dependent variable, there were n o cases to delete and residual scatterplot had two distinct lines. The model significantly pre dicted occurrence of physical assault, F
182 (2/508) = 5.36, p = .005. R2 for the model was .021 and the adjusted R2 was .018. Table 48 (Appendix A) displays unstandardized coefficient s and standardized coefficients for the two variables in the model. The model predicts that the highest rate of physical assault occurs to students who are working in a dev elopmental disability setting and students who have more field agency training. Thoug h the prediction is significant, these two factors contribute only 1.7% in shared variabil ity. Predicting of threatened physical harm. Using the dichotomous variable representing no experience with threats of harm ver sus experience with threats of harm, a stepwise multiple regression analysis was conducted to ascertain which, if any, of the 18 independent variables might predict the occurrence of threat of physical harm. Data screening did not result in the deletion of any cas es. Because there was no variability in the dependent variable, scatterplot showed two even ly distributed lines. As indicated in Table 49 (Appendix A), the analysis revealed the mo del significantly predicted occurrence of threat of physical harm, R2 = .053, R2 adj = .050, F (2/504) = 14.239, p .001. Being male ( t = 4.53, p .001) and interning at a mental health/psychiatric setting ( t = 2.25, p =.025) significantly predicted that threat of viole nce would happen. These two factors accounted for 5% of the variance in threats of physical harm. Predicting verbal abuse. The rate of verbal abuse for this sample was predicted by using a dichotomous variable to determ ine whether or not a student had been verbally abused with 20 independent variables in a stepwise multiple regression. No cases were removed due to the dichotomous nature of the dependent variable. Likewise, residual plots demonstrated two straight lines. Tab le 50 (Appendix A) shows that the
183 regression analysis revealed a model that significa ntly predicted occurrence of verbal abuse, F (4/502) = 6.63, p .001. R2 for the model was .05 and the adjusted R2 was .043. Thus the model indicates that the overall model predicts 4.3% of the variance in the occurrence of verbal abuse. Four variables predicted whether v erbal abuse would occur. The significant predictors were: being male; interning in a mental health/psychiatric setting; making 11 or more home visits; and interning in an alcohol/substance abuse setting. Predicting threat of lawsuit. Prediction of occurrence of threat of lawsuit was measured using a stepwise multiple regression a nalysis to determine which, if any, of 18 independent variables predicted such an occurren ce. Occurrence of lawsuit was measured through a dichotomous variable that report ed whether or not a student had experienced threat of lawsuit. No cases needed to b e dropped since there was no variability in the data. The regression model signi ficantly predicted occurrence of threat of lawsuit, F (4/502) = 8.99, p .001. Table 51 (Appendix A) indicated that R2 for the model was .067 and the adjusted R2 was .059, indicating that 5.9% of the variance in occurrence of threat of lawsuit. Four of the twenty independent v ariables significantly predicted whether threat of lawsuit would occur. Being aged 2 5-30, being of mixed ethnic heritage, making 11 or more home visits, and interning in alc ohol/substance abuse settings predicted occurrence of threat of lawsuit. Predicting damage to personal or professional prop erty. A stepwise multiple regression was computed to determine the a ccuracy of 18 independent variables
184 predicting the occurrence of property damage. Since the dependent variable predicted only whether or not property damage had occurred, t here were no cases dropped and residual diagram reflected two straight lines, as t here is no variability in a dichotomous variable. The rate of property damage was significa ntly predicted by a three factor model, F (3/503) = 8.10, p .001. The R2 was .046 and the adjusted R2 was .040, thus indicating that that the model accounted for 4% of the varianc e in occurrence of property damage. Table 52 (Appendix A) presents the regression coeff icients for the three factors, which were being male, having a mixed minority supervisio n dyad, and interning in a mental health/psychiatric setting.
185 Chapter V Summary and Discussion Summary of Findings Response Rate A total of 1,500 surveys were mailed to randomly se lected social work students from the NASW student membership roster for 2008, m aking this the only social work student client violence study ever done using proba bility sampling. The response rate for the current study was 45.25%, which was substantial ly higher than the 30% that was projected. This rate is higher than all but one res ponse rate in national social work education studies published in two major journals i n the past six years. Additionally, the number of respondents who met the inclusionary crit eria ( N = 595) was by far larger than the 159 recommended by Green (1991) in order to hav e 18 predictor variables, a statistical power of .80, an alpha of .05, and a me dium effect size. Also, this sample size was more than double the largest sample size to dat e ( N = 258) in a social work student client violence study (Knight, 1996). Characteristics of Sample Gender/age. The social work students in this client violence study were largely female (88.2%), similar to every client violence st udy done with social work students to date. The largest age group represented was the 2530 group (27.4%), followed closely by those under 25 (24.4%), thus slightly more than half of the sample were traditional
186 aged BSWs and MSWs. The remaining students (47.9%) were over 30. This is a slightly lower proportion of students under 30 than the 61% of BSW students under 30 in Knights (1999) client violence study, yet a higher proportion than in Elwood and Reys (1996) student study, in which 24.4% of the social work students were under 30. (The Elwood and Rey study did not specify whether the st udents were BSW, MSW, or both.) The mean age in the current study was 33.1, which i s higher than the mean age in other student studies that included only BSW students (Kn ight, 1996; Mama, 2001). Only one other client violence study has queried both MSW an d BSW students and mean age was not included in that study. Experience. The majority (49.4%) of the sample had no paid so cial work experience, with an additional 15.1% of students ha ving less than 3 years experience. This is the least experienced group of social work students in any social work student client violence studies to date. Race/ethnicity. Seventy one percent of the students in this sample were White. Only one student study had a larger percentage of m inority students (Knight, 1999). Only one other student study with a very small sample re ported specific minority ethnicities of students (Schwarzmueller, 1998). Thus the current s tudy is unique in its effort to solicit responses from higher numbers of non-White students and to report and analyze the effects of their specific ethnicities on client vio lence. Degree program. By design, the sample was made up of half BSW stu dents and half MSW students. Of the questionnaires included i n the study 57.1% were those of MSW students, whereas 42.2% were from BSW students. Only one other social work
187 student study included both MSW students and BSW st udents (Tully et al., 1993) and an additional study focused on 37 MSW students (Schwar zmueller, 1998). Together, only 109 MSW students have previously been asked about c lient violence. This study contributes to the literature by including response s of 340 MSW students. Home visits. About half of the students in this study made no h ome visits during the course of practicum. It is implied that many of the respondent completed most of their practicum experience in agency settings. Of those s tudents who did make home visits, 56% made 11 or more home visits. Thus, if students were in a practice setting where they make any home visits, they are prone to see numerou s clients in their homes. Students in child/youth/child protection settings were much mor e likely to make increased amounts of home visits than students in all of the other pr actice settings. Other practice settings where students made increased amounts of home visit s are mental health/psychiatric settings and family service settings. Evening hours. One third of the students said that they did not wo rk any evening hours and an additional 42.7% of students worked be tween 1% and 25% of their time during evening hours. This implied that most of the respondents completed the majority of their practicum experience primarily during dayt ime hours. However, many students in corrections/criminal justice settings, alcohol/subs tance abuse settings, and family services settings worked more than 50% of their practicum ti me during evening hours. Practice settings. The three largest practice settings for this sampl e were child/youth/protective services, psychiatric/mental health, and medical/healthcare services. Higher percentages of MSW students were i n hospital medical settings,
188 alcohol/substance abuse settings, schools, communit y organizations, and psychiatric mental health settings. Higher percentages of BSW s tudents were in developmental disabilities, corrections/criminal justice, childre n and youth/child protection, and family services. The only previous social work student cli ent violence study that mentioned specific internship sites of students reported only that the majority of the 37 students were placed in mental health related placements (Mama, 2 001), so this study is the first student client violence study to report information about m ore specific practice settings. Training. Of 19 possible safety content areas that could hav e been taught in training, a mean of 7.63 were covered in social wor k practice classes. A slightly lower mean number ( M = 6.24) of content areas were received in field ag encies. An even lower mean amount of content was received in field semina r ( M = 3.92). Forty two (7.2%) of the students reported that they received no safety training in social work practice classes, while 12.9% had no safety training at the field age ncy. One third of the students received no safety content in field seminar. Slightly more t han half of the sample (54.4%) received at least some safety training from other sources. The safety content area received by the most stude nts in social work practice classes was where to sit when interacting with a cl ient. In field seminar and at the field agency, the training area most frequently received was keeping the supervisor informed of ones itinerary. Conversely, the area covered l east often in field seminar, social work practice classes and the field agency was physical techniques for self protection. Interestingly, the two content areas received most often from sources other than the social work program was physical techniques for self prote ction and verbal de-escalation.
189 The only student study to previously ask about saf ety content that students had received was Elwood and Rey (1996). The content are as covered in social work education about which they asked cannot be directly compared to the current study as they are more general content areas. From Elwood an d Reys study, four training content areas covered in the field agency can be compared t o content areas received by students in the current study. Both studies questioned conte nt on verbal de-escalation. Elwood and Reys study asked about content on predicting viole nce. Though the current study does not inquire about this same content area, three con tent areas imply prediction of violence: characteristics/life experiences of people more lik ely to commit violent acts; forms of mental illness associated with violent behavior; an d assessing history of violence in clients. Elwood and Reys study reported on trainin g on physical management of violence, which compared to the question in the cur rent study on physical techniques for self defense. Similar percentages of students from both studies reported experiencing these comparable training content areas. However, f ewer students in the Elwood and Rey study had knowledge of agency safety policies than in the current study. The remaining 13 training content areas in the current study have not been reported in any other student studies. Many of the safety content areas reported by stude nts in this study can be compared to the safety content area reported to hav e been covered by BSW field programs in Faria and Kendras (2007) study. For ev ery content area that can be compared the students recall receiving substantiall y less training than the field directors said that they had provided. (See Table 53, Appendi x A). For example, 100% of the field
190 directors in Faria and Kendras study said that the y had provided safety content on characteristics of high-risk situations. In contras t, only 48.1% of the students in the current study recalled having training in this same area in social work practice classes. There were only two areas that approached agreement between the field director perspective and the student perspective. Over half (55.9%) of the student respondents felt that they were either somewhat prepared or not prepared at all to deal with violen t or potentially violent clients. Males feel significantly more prepared than females and M SW students feel significantly more prepared to deal with violence than BSW students. I ncreased work experience also helped students to feel more prepared. Students un der the age of 24 felt the least prepared to deal with violence. Fear of future violence. Students reported that they had more fear of future verbal abuse than any other type of violence. Almost half had at least a moderate amount of fear that verbal abuse would occur to them in the future Students were least fearful about the prospect of damage to their personal or professiona l property. Overall, 36.2% strongly or moderately feared that some type of violence would occur to them in the next year. Occupational commitment. Social work students in this study have extremely high levels of affective commitment, which is the t ype of commitment that demonstrates a pride in and identification with the profession o f social work. They scored a mean score of 23.35 of a possible 25 on the affective commitme nt subscale. On the normative commitment subscale, students scored a mean score o f 11.69 of a possible 20, thus demonstrating that they are moderately committed to the profession due to a sense of
191 obligation or responsibility. Finally, students sco red a mean score of 16.30 of a possible 25, demonstrating that students are moderately comm itted due to feelings that they have invested too much of their lives in the profession to turn back. Only two student studies that included measurement of occupational commitment could be located for comparison to the present stud y (Lazar, Cohen, & Guttman, 1995; Meyer, Allen, & Smith, 1993). The first was the onl y previous social work student study to measure occupational commitment (Lazar et al, 19 95). These researchers created a version of the scale that was used in the current s tudy, but they measured only affective commitment and continuance commitment and the scale s had more items. Given these differences, it can be noted that, like the current study, affective commitment was stronger than continuance commitment. The U.S. stud ents in the current sample had a much stronger identification with and pride in bein g a part of the profession of social work than the social work students in Israel, but t he U.S. students also had stronger feelings of being committed due how much it has cos t them monetarily, time wise, and energy wise to be in the profession of social work. The other comparison student study was a study of nursing students by the creators of the scale used in the current study (Me yer, Allen, & Smith, 1993). Like the current study, affective commitment was higher than normative and continuance commitment. The social work students had higher lev els of all three types of occupational commitment than the nursing students. This seems to indicate that while the social work students in the present study have stronger feeling s of identity with their profession and
192 stronger feelings of responsibility toward the prof ession, they also have stronger feelings of being committed to the profession because it is too costly for them to quit. Career withdrawal intentions. Few social work students in this sample had intentions or thoughts about withdrawing from the p rofession of social work. The mean score on the career withdrawal intentions scale was 4.40 of a possible 15, with 3 indicating the least amount of withdrawal intention s. No studies have previously focused solely on students so there are no comparisons that can be made between this study and other studies. Summary of Prevalence of Direct Exposure to Client Violence Of 595 social work students 41.7% ( n = 248) experienced at least one incident of client violence during their practicum. The most co mmon type of violence experienced was verbal abuse (37.5%). Slightly more than14% of students experienced threat of physical harm, while 9.5% were threatened with a la wsuit and 7.3% had damage to personal or professional property by a client. Twen ty one students (3.5%) were physically assaulted within their practicum experience. In tot al, social work students experienced 1591 incidents of client violence during practicum. Previous studies have indicated that child protect ion/child welfare settings are among the most common settings for social workers t o experience client violence (Beaver, 1998; Jayarante et al., 2004; Newhill, 199 6; Song, 2005). Songs (2005) national study drawn from NASW membership that focused on th e client violence experience of child protection/child welfare workers is felt to b e a good comparison study due to its similar definitions of violence. A direct compariso n cannot be made due to the practicing
193 social workers in Songs study reporting incidents within the past year and social work students in the current study reporting incidents t hat occurred during practicum. However, it should be considered that a large major ity of students in the current study had been in practicum for less than 1000 hours, whi ch is the equivalent of a half year of work at the most. When compared to annual violence rates for practicing child protection social workers in Songs study, the rates of client violence towards the social work students in the current study are higher in every c omparable category. Thus, in less than 1000 hours of practice, social work students in thi s study reported higher rates of every type of violence than social work practitioners rep orted in one year at one of the most dangerous practice settings reported in previous so cial work client violence studies. It is more difficult to compare rates of client vi olence in other social work student studies with the current study as there has been li ttle consistency in the way questions about violence have been asked. Three of the studen t studies listed total client violence rates of 43.5% ( N = 78) (Elwood & Rey, 1996); 54% ( N = 37) (Mama, 2001); and 26% ( N = 121) (Tully et al., 1993). The rate of client vi olence in the current study is 41.7%. Verbal abuse rates have been reported between 22% ( Tully et al., 1993) and 49% (Mama, 2001), whereas the verbal abuse rate in the current study is 37.5%. Physical assault has been reported between 2.1% (Knight, 1996) and 15.8% (Mama, 2001). The current studys physical assault rate fell on the low side of this continuum, at 3.5%. Summary of Prevalence of Indirect Exposure to Clien t Violence The majority (60.2%) of the social work students i n this study had either heard about or witnessed client violence occurring to coworkers and/or fellow social work
194 students. As with direct client violence, the most prevalent type of violence that was indirectly experienced was verbal abuse (54.1%). Al most one quarter (23.4%) of the students had been indirectly exposed to threat of l awsuit and physical assault. Finally, 20.5% of students had been exposed indirectly to pr operty damage. Two previous student studies asked students if they were aware of violence in the workplace. Elwood and Rey (1996) found that 43.6% o f the 78 social work students questioned were aware of violence toward profession als in the workplace. This percentage is less than the total percentage of stu dents who indirectly encountered client violence in the current study. Tully and colleagues (1993) found that 25% of social work students knew of or had seen violence toward other personnel at their practicum site. This is also lower than the indirect exposure to client violence experienced by the students in the current study. The social work students in the current study had s ubstantially higher indirect exposure to client violence during their practicum (60.2%) than child welfare social workers in Songs (2005) study had in one year (37. 2%). In all comparable categories, this sample of social work students heard about or witnessed more of every type of violence during their practicum than social work pr actitioners experienced one years time at one of the most dangerous practice settings Summary of Prevalence of Client Violence Hypotheses Hypotheses 1-1 through 1-5 were related to personal demographic characteristics. Hypothesis 1-1 projected that male social work stud ents would experience more of every type of violence than female social work students. Descriptive data reflected that male
195 social work students experienced every type of viol ence at almost twice the rate of female social work students. Males also experienced a high er mean of client violence incidents than females for every type of violence. Additional ly, males experienced significantly more threats of physical harm. Findings trended tow ard significance with males experiencing more verbal abuse and sum of direct vi olence than females. Thus, the hypothesis was accepted for threats of physical har m. There were no significant differences between males and females in numbers of physical assaults, property damage, verbal abuse, threats of lawsuits, and sum of direc t client violence, thus the hypothesis was rejected for those types of violence. Hypothesis 1-2 projected that younger social work students experience more client violence than older social work students. Th ere was no significant relationship between age and client when age was analyzed as a c ontinuous variable. When age was categorized into groups of traditional aged BSW stu dents (aged 24 and under), traditional aged MSW students (aged 25 to 30), and non-traditio nal aged students (aged 31 and older), it was found that students aged 25-30 exper ienced more threats of lawsuit than students under the age of 25 or over the age of 30, at a level that approached significance. There were no significant findings for any type of violence. Thus, for all types of violence, the hypothesis was rejected. Hypothesis 1-3 projected that less experienced stud ents would experience more of every type of client violence than social work stud ents with more paid social work experience. Strong positive relationships at levels that trended towards significance were found between social work students years of experi ence and verbal abuse and between
196 years of experience and summed client violence. Sim ilarly, students who had 3-5 years experience encountered more verbal abuse and summed client violence than students with no experience. This finding also approached si gnificance, though with more stringent Bonferroni requirements, the relationship was not statistically significant. There were no significant relationships between years of experience and any other type of client violence. Though students with 3-5 years experienc e experienced more verbal abuse and summed violence than students with 1-2 years exper ience or students with 5-10 years of experience, the differences were not statistically significant. Hypothesis 1-3 was rejected for every type of client violence. Hypothesis 1-4 projected that there would be no di fference in exposure to all types of violence by students of various ethnicitie s/racial backgrounds. The highest rates of violence were among students with mixed ethnic h eritage, followed by Asian/Pacific Islanders and Native Americans. It was found that N ative Americans had significantly more verbal abuse than every other ethnic group and had significantly more summed client violence than every other ethnic group excep t Asian/Pacific Islanders. Threat of lawsuit was significantly more likely to occur to s ocial work students with mixed ethnic heritage. Property damage was more likely to occur to student s of mixed ethnic heritage than Whites, Blacks, and Latinos and it was more li kely to occur to Native Americans than Blacks. After Bonferroni adjustments, these fi ndings were near, but did not reach significance. There were no significant differences between ethnic groups for physical assault or threat of physical harm. Therefore, the hypothesis was accepted for physical
197 assault, property damage, and threat of physical as sault. However, the hypothesis was rejected for verbal abuse, threat of lawsuit, and s ummed client violence because particular groups of minority ethnic students suffe red more of those kinds of violence than Latino/Hispanic student, Black students or Whi te students. Hypothesis 1-5 projected that BSW students would e xperience more client violence than MSW students. MSW students experience d higher rates of threats of lawsuit, verbal abuse, threats of physical harm and summed client violence, while BSW students experienced higher rates of physical assau lt and property damage. It was additionally found that MSW students experienced si gnificantly more threats of physical harm than BSW students. MSW students also experienc ed total violence at a higher level than BSW students, as the .03 probability level ind icated, however with Bonferroni adjustments, the findings only trended toward signi ficance. There were no other significant findings. The hypothesis is rejected fo r all types of violence. Hypothesis 1-6 projected that there would be signi ficant differences in numbers of client violence incidents occurring to social work students according to where the violence occurs. The most recent violent incidents of almost every type of client violence took place at highest rates at office settings. Onl y physical assault happened more often in settings other than the office or the clients h ome. Because the questions about place of violence asked only about the most recent incident of violence, it was not possible to inferentially test the hypothesis. Therefore the hy pothesis could not be accepted or rejected. A related question was analyzed to deter mine if social work students who make more home visits experience a higher mean number of violent incidents. It was found that
198 threats of lawsuits and summed client violence were more likely to occur to students who made the most home visits, though the probability l evels only trended toward significance after Bonferroni adjustments. Hypothesis 1-7 projected that there would be a sig nificant difference in numbers of client violence occurring to social work student according to the time of day. For every type of violence, the highest rate of violence occu rred during daytime hours. The highest rate of violence was among those who worked 51-75% of their practicum during evening hours. Additionally, physical assault trended towar ds significance when students worked 51-75% of their practicum during evening hours. The hypothesis could not be directly rejected or accepted. Through a similar analysis co ncerning exposure to client violence per amount of evening hours worked, the hypothesis was also rejected for every type of violence. Hypothesis 8 projected that there would be differe nces in exposure to client violence depending upon the practice settings of th e students. It was found that there were differences in client violence exposure depend ing on practice settings and the findings approached significance. The three practic e settings with the highest rates of violence were mental health/psychiatric settings, d evelopmental disabilities settings, and alcohol and substance abuse settings. Students in t hese settings were more than two times more likely to encounter client violence than in co mmunity planning/community organizing settings, which had the lowest rate of v iolence. Criminal justice settings had the highest mean of client violence, followed by sc hools, developmental disabilities settings, mental health services settings, and alco hol/substance abuse settings. The only
199 type of violence that was substantially related to practice setting was threats of lawsuit, which was more likely to happen in alcohol/substanc e abuse settings than any other setting. This finding approached significance at th e .05 level, but was not significant after Bonferroni adjustments. The hypothesis was rejecte d none of the findings reached a level of significance. Hypothesis 9 projected that there would be a diffe rence in exposure to client violence depending upon whether the students race/ ethnicity was different from their field educator. Students who were in unmatched min ority supervision dyads had the highest rates of every type of violence. This group also had the highest mean of client violence, as opposed to matched supervision dyads, White student/minority supervisor dyads, and minority student/White supervisor dyads. It was additionally found that property damage was experienced by significantly mo re students who were in mixed minority supervision dyads than the other three gro ups. There were no other significant findings for any other type of client violence. The hypothesis was accepted for property damage occurring to students who are in mixed minor ity supervision dyads. The hypothesis was rejected for all other types of clie nt violence. Hypothesis 1-10 projected that students who have m ore safety training offered by their social work program would experience less of every type of violence than social work students who have had less safety training in their social work program. There was no significant relationship between total training in social work classes and exposure to client violence. Total training in field seminar wa s positively and significantly related to verbal abuse, property damage, and summed client vi olence. Total training was also
200 related to threatened lawsuit at the .04 level and threatened physical harm at the .05 level, but these relationships failed to be significant af ter Bonferroni adjustments. Total training at the field agency was positively related to every type of violence. There were statistically significant findings for t he relationships between training at the field agency and threatened physical harm and sum o f client violence. The relationships between field agency safety training and physical a ssault and property damage were near significance with .01 probability levels. Verbal ab use and property damage also trended toward significance, with less than .05 probability levels, but none of the four types of training met the significance criteria after Bonfer roni adjustments. For training in places other than the social work p rogram positive relationships that were nearly significant (at the .05 probabilit y level) were found with verbal abuse, property damage, and summed client violence; howeve r, after Bonferroni adjustments, these relationships did not meet significance requi rements. The positive correlations between training in vari ous venues and types of violence indicate that as students received more tr aining they also experience more violence. Though there were some significant findin gs, they were not in the direction predicted. Therefore, this hypothesis must be rejec ted for every type of violence. Summary of Additional Analysis After all prevalence of client violence hypothese s had been addressed, further analyses were completed to determine the best predi ctors of each type of violence. A set of 18 independent variables were chosen on the basi s of findings in the prevalence of client violence analyses. They were then entered in to a multiple regression analyses for
201 each type of violence as well as overall client vio lence. The analysis predicted the occurrence of the type of violence. For summed clie nt violence a second multiple regression analysis predicted the frequency of viol ence. Based on previous analyses findings, 18 independen t variables were entered into all of the multiple regression analyses to determin e any significant predictors of the types of violence. When the independent variables were e ntered into the first stepwise multiple regression analysis to predict occurrence of summed client violence, it was found that being male, working in a mental health/psychiatric setting, making 11 or more home visits, and working in an alcohol/substance abuse s etting significantly predicted occurrence of client violence. The second stepwise multiple regression found whic h factors predicted frequency of summed client violence. This analysis indicated that being male significantly predicted frequency of client violence. The previous analyses found that the following fac tors strongly affected experience of physical assault in social work stude nts: being aged 25-30, working 51-75% evening hours, and having more field agency training. When 18 variables were entered into a multiple regression analysis, two va riables, working in a developmental disability setting and having more field agency tra ining, significantly predicted occurrence of physical assault. Hypotheses analyses in this study indicated that b eing male, being of mixed heritage, being an MSW student, and having more fie ld agency training are significantly related to increased threat of physical harm. Furth er step-wise analysis to check which
202 factors significantly contributed to occurrence of threat of physical harm found that being male and interning in a mental health practice sett ing were significant predictors. This study has shown that the following factors ar e strongly related to social work students encountering verbal abuse: being male, Nat ive American, MSW student, having 3-5 years experience, and receiving more training in field seminar, the field agency and sources other than the social work program. When 18 factors were entered individually to check for factors that most significantly predict o ccurrence of verbal abuse, four factors were significant: being male, interning in a mental health setting, making 11 or more home visits, and interning in an alcohol/substance abuse setting. Bivariate hypotheses analyses completed previously in this study indicated that making increased home visits, working in an alcohol /substance abuse setting, and having more field agency training contributed to increased threats of lawsuits. When a step-wise multiple regression analysis was completed to deter mine the best predictors of occurrence of threats of lawsuits, it was found that being 2530 years old, being of mixed ethnic heritage, making more home visits, and interning in an alcohol/substance abuse setting all significantly predicted occurrence of threats of la wsuit. Previous bivariate hypotheses analyses indicated t hat being of mixed ethnic heritage, being a student in a mixed minority super vision dyad and having more field agency training were strongly associated with stude nts encountering damage to their personal or professional property. Multiple regress ion analysis showed that being male, having a mixed minority supervision dyad, and inter ning in a mental health/psychiatric setting significantly predicted property damage.
203 Summary of Fear as Mediator Hypotheses The second series of hypotheses were related to an alyzing whether fear of future violence mediated between experience of client viol ence and occupational commitment. A precondition for full mediation to occur was for client violence and occupational commitment to be significantly related. Though ther e was no hypothesis for this step, it was projected that any correlations would be negati ve. For this step, Pearsons correlations were calculated between the six direct client violence variables and the six indirect exposures to client violence variables and the three types of occupational commitment. There were no significant correlations between any type of direct or indirect violence and affective, normative, or cont inuance commitment. A four step mediator analysis had been planned ini tially. However, because the first step of the mediator analysis found no signif icant relationships between all of the types of direct and indirect violence and the three types of occupational commitment, it was not appropriate to proceed with the other three steps of the mediator analysis. In other words, experience with client violence did no t significantly affect the students perceptions concerning their commitment to the prof ession of social work. Since there was no relationship to be mediated, further planned steps were not required. Though fear of future violence was not found to me diate the relationship between client violence and occupational commitment or care er withdrawal intentions and the mediator analyses were thus not completed, the hypo theses were analyzed, as proposed. Hypothesis 2-1 projected that client violence would have a positive effect on fear of future violence. Pearsons correlations between eve ry type of violence and general fear of
204 future client violence were positive and significan t for every type of violence except physical assault. Correlations between every type o f violence and specific fears of the same type of violence were even more strongly posit ive and significant. Indirect exposures to every type of client violence were pos itively and significantly correlated with general fear of future violence for every type of violence. Indirect exposures to client violence were even more strongly and positiv ely related to fears of the specific type of violence to which the student had been exposed. Hypothesis 2-1 was accepted for every type of direct client violence and every type of indirect exposure to client violence. Hypothesis 2-2 projected that fear of future viole nce would have a negative effect on occupational commitment. All correlations betwee n fear of specific types of future violence and affective commitment were negative, bu t only two types of fear approached significance. Fear of physical assault were related to affective commitment at the .01 probability level and fear of threatened harm was r elated at the .05 probability level; however, after Bonferroni adjustments, the correlat ions did not meet the required level of significance. The next series of Pearsons correlations for Hypot hesis 2-2 was done between specific fear of future client violence and normati ve commitment. Fears of all specific types of violence were positively and significantly correlated with normative commitment. The final series of correlations were c alculated between specific fears of future violence and continuance commitment. Fears o f every specific type of future violence were positively and significantly correlat ed with continuance commitment. Since none of the correlations between fear and aff ective commitment reached the
205 required level of significance and since the relati onships between fear and normative commitment and continuance commitment, though signi ficant, were not in the projected direction, Hypothesis 2-2 was rejected in every cas e. The final hypothesis, 2-3, projected that fear of f uture violence would have a negative effect on career withdrawal intentions. Pe arsons correlations were calculated between fears of specific types of client violence and career turnover intentions. There were no significant correlations, thus this hypothe sis was rejected for every type of fear of future violence. In summary, findings indicated that all types of di rect and indirect exposure to client violence significantly increased fear of fut ure violence. Furthermore, fear of future violence was then found to significantly increase s tudents normative and continuance commitment. Discussion Gender and Client Violence This study found that males encountered significan tly more threats of harm and they experienced more verbal abuse and sum of total violence at levels that approached significance, yet they did not encounter significan tly higher levels of physical assault, threat of lawsuit, or property damage. This is part ially supported in the literature. Most previous research points to males generally encount ering more violence, though violence in most studies is not broken down into the differe nt types (Guterman et al., 1996; Jayarante et al., 1995; Jayarante et al., 2004; New hill, 1996; Ringstad, 2005). Carmel and Hunter (1991) speculated that male workers encounte ring more violence in a psychiatric
206 setting might be due to a belief that male staff sh ould be involved in clinical situations with a potential for violence. Additionally, they s tated that for cultural reasons, both patients and staff might react differently when mal e staff was involved. Newhill (1996) speculated that males may have higher incidents of client violence because they have a higher willingness to work in settings where there is more likelihood of more dangerous clients. In the current study it was found that mal es are significantly more likely to feel prepared to handle client violence. It may be that their level of perceived preparedness may lead them to undertake roles in more dangerous settings or with more dangerous clients. The current study also found that a larger percentage of males work in alcohol/substance abuse settings and mental health/ psychiatric settings than females and these settings were among the most dangerous for so cial work students. Age and Client Violence This study found that older traditional aged studen ts, those aged 25-30, have higher rates of overall violence and three specific types of violence. This group is also more likely to be threatened with lawsuits at level s that approach significance. The mixed finding that some younger, aged 25-30, social work students do experience more client violence, yet some even younger, under 25, social w ork students dont experience more client violence is somewhat supported in the litera ture. The largest body of literature on social work practitioners experience with client v iolence has found that younger social workers experience more violence (Beaver, 1998; Gut erman et al., 1996; Jayarante et al., 1995; Jayarante et al. 2004; Song, 2005). However, Newhill (1996) did not find a
207 significant relationship between the age of the soc ial work practitioner and client violence. Only two social work student studies have explored the correlation between age and client violence and they have conflicting findi ngs. Elwood and Rey (1996) found that age did not affect experience with client violence. However, the age categories used for that study were different in that students under 30 were all in one category, unlike the present study where this age group was subdivided. Tully and colleagues (1993) found that younger student clinicians were less likely to experience client violence than professionally seasoned MSWs (p. 197). Though thi s statement is not well explained in the study, it is presumed that the researchers are referencing the ages of students versus the ages of the field educators who were also quest ioned. Tully and colleagues speculated that younger students may be protected from violenc e by their status as students. Since the current study included only students, it cannot be stated that they are sheltered from violence due to their student status. However, it m ay be possible that younger traditional aged students, those under age 24, may be sheltered from some violent encounters due to links with the specific expectations for BSW studen ts versus MSW students. BSW students are trained to be generalist practitioners which may call more on their roles to link, broker, and manage cases. These roles may mea n that there is more mezzo and macro level client contact, which may reduce the am ount of direct contact with potentially violent clients. Knights (1996) study provided some support for this theory, reporting that BSW students in the study had limite d exposure to some practice activities in which other workers in the agencies engaged on a routine basis. It was surmised that
208 the students might be shielded from potentially vio lent practice situations. This will be further discussed under the section Degree Program and Client Violence. On the opposite end of the age spectrum there is a possible explanation for students over 30 experiencing less client violence. Older students may take more precautions to protect themselves from client viole nce. Castellanos (1998) found a significant positive correlation between the age of child welfare workers and the number of current personal safety practices employed, indi cating that as age increased, workers were increasingly likely to make efforts to protect themselves. Another possible explanation for older students experiencing less vi olence may be found in Schwarzmuellers (1998) study of MSW students. This researcher found that there was a negative association between age and the extent to which students felt physically threatened during home visits. This indicates that as age increased, social work students were less likely to feel physically threatened duri ng home visits. When a student feels less physically threatened there may be less feelin gs of vulnerability in the student and thus, a greater ability to share power with clients reducing the possibility that the client will act out violently. Work Experience and Client Violence The finding that trended toward significance that social work students with 3-5 years of paid social work experience were exposed t o more verbal abuse and summed client violence than students with no experience ha s not been supported in the literature. Most articles that report on the effect of worker e xperience on client violence have found that those with less experience encounter more viol ence (Bernstein, 1981; Beaver, 1998;
209 Guterman et al., 1996; Privitera et al., 2005; Tull y et al., 1993). However, others have not found a significant relationship between experi ence and client violence (Jayarante et al., 1995; Newhill, 1996; Ringstad, 2005). In the current study, chi square tests were perform ed to more specifically assess characteristics of students with 3-5 years experien ce to determine if other factors may be influencing this finding. Indeed, many of the chara cteristics of the students with 3-5 years experience have been found elsewhere in this study to contribute to exposure to client violence. Students with this moderate level of experience are significantly more likely to be aged 25-30, MSW students, and in some of the practice settings that have been found to be the most dangerous: schools, alcoh ol/substance abuse settings, and mental health/psychiatric settings. Though not of s tatistical significance, several other factors were found to be elevated in the group of s tudents with 3-5 years experience. They were more likely to be students who were male, Asian, Native American, of mixed ethnic heritage, or Black. They also are likely to make increased home visits and to work more evening hours. Guterman and colleagues (1996) found that although less experienced social workers in the U.S. encountered more client violenc e, this was not the case with social workers in Israel. The researchers surmised that th e difference might be due to differences in the workplace, stating that in Israe l, workers may have less opportunity to advance in an agency; thus, they may have fewer cho ices concerning the clientele with whom they will work. Applying some of this argument to the students in this study, it
210 may be that as students, clients are more carefully chosen for them, so less experienced students may be less likely to be victimized. As with the age hypothesis and as was surmised by T ully and colleagues (1993) in their study of BSW and MSW students, it is possible that those students with no experience are somewhat more sheltered by their fie ld agency from immediately rendering direct service to clients due to their st udent status. Perhaps they may have a tendency to shadow other workers prior to being req uired to function independently. Additionally, they may be given more time to acclim ate to an agency, whereas students with a few more years of experience may be presumed to be already prepared to begin giving direct, independent service to clients. Ethnicity and Client Violence The finding that some ethnic minority groups exper ience more client violence has been only briefly alluded to in the literature (But t, 2000, as cited in Brockman, 2002). In a book chapter about discrimination in social servi ces in England, it was reported that of Black and Asian social service workers who had been verbally abused, approximately half attributed a racial motive to the abuse, compa red with 2% of the White staff (Aye Maung & Mirrles-Black, 1994, as cited in Davey, 199 9). In the current study one Black female wrote in that the verbal abuse she had encou ntered was racially motivated. Only four social work studies have mentioned analy ses of ethnicity and client violence. Three national random social work practit ioner studies found that race was not a significant predictor of client violence (Beaver, 1999; Jayarante et al. 2004; Ringstad, 2005). However, Beaver (1999) found that Native Ame rican social workers had the
211 highest rate of violence (50%), which was almost th e same rate of summed client violence of Native American students in the current study (53.8%). One student study included approximately 108 minority students and fo und that there was no relationship between race and likelihood of being assaulted (Kni ght, 1996). This study did not mention whether the minority ethnic groups were dif ferentiated so it is unknown if they were analyzed by specific group or as an entire gro up. One other study of workplace assaults on minority health and mental health care workers found that race was a weak predictor of assault (Sullivan & Yuan, 1995). Further analyses between the student race variable and other demographic variables indicated Native American students may su ffer more client violence in part because they work more evening hours, they work in more violent settings, and a larger percentage are male. Mixed heritage students may en counter more violence because of being in more dangerous practice settings. It is of interest that larger minority groups, Bla cks and Hispanics, did not encounter violence at significantly higher levels. In fact, Hispanic students in this study encountered the least amount of violence. Rather, t he minority students who were most victimized by client violence were from even smalle r racial/ethnic groups. These groups are less represented in U.S. Perhaps these students were practicing in areas where their differences are more noticed and perhaps where they are more vulnerable to discrimination. It was not possible to discover mor e about this, as there were no questions concerning the geographic location of the students.
212 The findings about ethnicity affecting exposure to client violence should be interpreted cautiously due to the low numbers of et hnic minority students in this study and because there may be other factors contributing to client violence other than ethnicity. However, because most other client viole nce studies have not considered the possibility that non-White people may suffer more c lient violence, the ethnicity findings from the current study should not be easily dismiss ed. Degree Program and Client Violence Contrary to the thought that BSW students would ex perience more client violence, it was found that MSW students actually e xperienced significantly more threats of physical harm and they experienced summed client violence at a level that approached significance. The two social work practitioner clie nt violence studies that addressed educational level did not support the educational l evel findings from the current study (Beaver, 1998; Privitera et al., 2005). However, i t should be noted that those studies included social workers who were older and had more years of experience and in Beavers study, the largest group of practitioners were in private, for profit settings. Additionally in Beavers study only 20 social work practitioners had a BSW degree and in Privitera and colleagues study, only 13 were BS Ws. The only student study to include both MSW and BSW students found no significant differences between MSW and BSW student s in encounters with violence (Tully et al., 1993). Thus the current study is th e only student study that has indicated that MSW students had higher amounts of certain typ es of client violence. It is possible that MSW students encounter more client violence du e to being in practicum longer.
213 However, it may be that MSW students are placed in agencies where they are more often expected to make independent clinical decisions. Th e MSW students in this study were more than two times as likely to be in mental healt h/psychiatric settings as BSW students and being in a mental health/psychiatric setting wa s one of the four significant predictors that client violence would occur. As mentioned in a n earlier hypothesis discussion, the largest group of MSW students in this study had 3-5 years of experience. It is possible that field agencies feel comfortable with placing M SW students more immediately into direct clinical practice where they may be more lik ely to encounter potentially violent clients. Finally, there are differences between th e expectations for BSW students compared to MSW students. BSW students may be in ro les that involve less direct clinical work, as discussed in previous results sec tions. Place of Client Violence It was not possible to inferentially examine the possibility that violence occurs more often in one setting than another. However, it was found that the most recent occurrence of every type of violence except physica l assault was in the office. This finding resonates with the findings of the two stud ent client violence studies (Mama, 2001; Tully et al., 1993). It also somewhat confirm s Newhill and Wexlers (1997) finding that different types of violence occur in different places. In both Newhill and Wexlers study and the current study, property damage occurr ed most often in the office and physical assault happened in various places. The cu rrent study showed that the most recent incident of threatened physical harm occurre d most often in the office, replicating Rapp-Pagliccis (2004) finding concerning probation officers exposure to violence;
214 however, this finding refuted Newhill and Wexlers (1997) finding that threats of physical harm occurred most often in the homes of the child and youth workers clients. Of students who made home visits, those who made in creased visits experienced total client violence and threats of lawsuits at ra tes that trended toward significance than students who made fewer home visits. This finding r esonates with other studies that have found that large percentages of home visiting socia l workers have experienced client violence while making home visits (Castellanos, 199 8; Vergara, 1998). Barling, Rogers, and Kelloway (2001) stated that when workers see cl ients in their clients home, access to the protection that might be offered in a tradition al office setting may be delayed or limited. They further stated that in-home workers m ay be forced to rely on their own resources to avoid or lessen the impact of violence that may be incurred upon them in the home. Isolation of social workers on home visits is seen as a problem that may lead to greater exposure to client violence. While many strategies can be employed to reduce ris k of client violence on home visits, social work students may not have had ample time to practice these techniques and may lack confidence to implement them. Knight (1999 ) found in a pretest and posttest about anxieties related to risk of violence in the social work practicum that more than one third of the 78 BSW students continued to be anxiou s about making home visits, even at the end of the field experience. No other studies have found that social workers exp erience more threats of lawsuits when they make increased home visits. Per haps the increased threats of lawsuit could be associated with the practice settings of w orkers making increased home visits.
215 The highest percentage of social work students maki ng 11 or more home visits interned in child/youth/child protection settings. Though th e current study did not find this setting to be statistically associated with threats of laws uit, it might be inferred that while making increased visits in the clients home settin g, students may be more vulnerable to threats of lawsuit. Parents may feel that in their own homes, they can threaten legal recourse without compromising the possibility of ha ving their children returned to them on a permanent basis. Time of Client Violence Though time of client violence could not be evalua ted inferentially, it was possible to find that, at levels that approached si gnificance, students who worked increased amounts of evening hours experienced more physical assault than students working less evening hours. One epidemiology study of violent injury in the workplace similarly found that workers in many occupations ex perience more homicide and nonfatal assault in the afternoon and evening hours (Kraus, 1996). No social work studies could be located indicating that the time of day significant ly influences experience of client violence. In the current study it is possible that increased client violence during evening hours may be influenced by the practice settings of the students who worked more evening hours. The practice settings where student s were likely to work increased evening hours were criminal justice settings, which had the highest mean of overall violent incidents, and alcohol/substance abuse sett ings, which was one of the settings with highest rate of overall client violence.
216 Practice Setting and Client Violence Though there were no significant findings after th e Bonferroni adjustments regarding practice settings that might be more dang erous, there were findings that approached significance that indicated that three s ettings were the most dangerous for the social work students: mental health/psychiatric, al cohol/substance abuse, and developmental disabilities. The finding that there were higher rates of violence in mental health settings has been supported in the literatur e. Jayarante and colleagues (1995) found that social work practitioners in institution al mental health settings had the largest number of physical threats. He and other colleagues again found in 2004 that mental health settings were among the most dangerous. Ring stads (2005) study concurred with this finding. No previous social work student studi es have asked about specific types of practice settings. Many studies have not included alcohol/substance abuse settings in their list of possible practice settings, thus the finding that s ocial work students encountered more violence in alcohol/substance abuse settings has on ly been echoed in one other study (Newhill, 1996). The finding that social work stude nts experience significantly more threats of lawsuit in alcohol/substance abuse setti ngs has been similarly found by Jayarante and colleagues (1995). They found that su bstance abuse social workers experienced the second highest rate of threats of l awsuit. Since many substance abuse clients are court ordered, it may be that clients f eel they can risk threatening lawsuit against a social worker whereas if they were to be more blatant in their threats or actions, they might fear incurring more severe legal sanctio ns.
217 No American studies could be found that analyzed e xperience of client violence in social workers from developmentally disability s ettings. The only study that included developmental disability workers had only 13 worker s in this category, so the category was excluded from further analysis (Jayarante et al ., 2004). A study of Canadian developmental disability workers indicated that in Canada, the rate of injury in the developmental services sector was higher than the r ate of injury in other social service settings (Baines, 2004). Though there is very limit ed evidence that indicates that social workers in developmental disability settings experi ence higher rates of violence, the current study finds otherwise. Clients who are deve lopmentally disabled may have less internal controls, thus more propensity for impulsi vity which could lead to more violent incidents. The current study did not find child welfare setti ngs to be among the most dangerous for social work students. This conflicts with numerous other studies that have shown child welfare types of settings to be most da ngerous (Jayarante et al., 2004; Newhill, 1996; Schultz, 1987). It is possible that the lack of findings in this area may be related to more BSW students in the current study b eing in child welfare settings and BSW students were found in the present study to exp erience less violence. The lack of findings might also be attributed partially to the fact the students in child/youth/child protection setting had the second highest amount of safety training in their field agencies. It is possible that previous research into child we lfare settings and anecdotal evidence of workers being severely harmed in these settings may have contributed to workers having increased effective training in this setting.
218 Field Educator/Student Race Supervision Dyads and C lient Violence No studies of any kind have explored a possible re lationship between client violence and the racial match of supervision dyads. Thus, the finding that students in mixed minority racial/ethnic supervision dyads expe rienced significantly more property damage and higher rates of every type of client vio lence is a new finding. The ethnicities of the students in the mixed minority supervision d yads were generally not the ethnicities of students found to experience higher rates of vio lence (Native Americans and mixed ethnic heritage). In fact, the group of minority st udents who had field educators of a differing ethnicity was a small, but mixed group. T hree fourths of these students were not part of ethnic groups that were found in the curren t study to experience more client violence. This is significant because it indicates that the students in the mixed minority dyads may not experience less violence solely becau se they are in a minority group that might be victimized at higher levels. Though other factors could explain why these students experienced more violence, it is plausible to consider that some aspect of the supervision relationship could have contributed to having more experience with violence. The findings concerning mixed minority supervision dyads should be interpreted with caution, since so few students are represented in t hese dyads. However, since there is such a paucity of information about cross cultural supervision, the finding should not be disregarded, as it may contribute some baseline dat a. Training and Client Violence The finding that students who have more training i n the field agency, field seminar and places other than the social work progr am have increased amounts types of
219 client violence was not anticipated. Though no soci al work practitioner studies have looked at the relationship between safety training and client violence, virtually every client violence study recommends training. Only one social work student study compared the experience of training and client violence in s ocial work students versus medical students (Elwood & Rey, 1996). Only descriptive inf ormation was reported and it indicated that social work students experienced les s violence than medical students and social work students had more training in violence prevention than the medical students. Possible explanations of the finding that more tra ining is related to more violence may be found in analyses completed in the study. Fo r example, some of the practice settings offering the most training also had the hi ghest amounts of client violence. Males also were more likely to receive more training and they also encountered more client violence. It may be that the finding of more traini ng being related to more client violence occurred in part because practice settings that are known to be more dangerous practice settings actually are providing more training to be prepared for this potential violence and regardless of the training, client violence persist s. If this is the case, we do not know how much violence may have occurred if no training was offered. Training variable analyses indicated that three of the content areas with the strongest correlations with client violence were ph ysical techniques for self protection, physical signs that violence is imminent, and verba l de-escalation. Actually, two of these content areas, physical techniques and recognizing physical signs of violence, were the least likely content areas to be covered by the soc ial work programs. In other words, very small percentages of social work students received training in these areas; however, these
220 content areas appeared to have strong influence ove r the finding that more training relates to more violence. These three content areas are oft en found in aggression control courses. It may be that field agencies where potential for c lient violence is higher may offer more intense training on self defense, thus, the finding that more client violence occurs when there is more training could be a result of the par ticular type of training that is emphasized. Again, we have no knowledge of how much violence would have occurred nor how many injuries may have occurred had there b een no training with these content areas. Finally, younger, less experienced BSW students we re more likely to receive safety training in the academic part of the social work program (social work practice classes and field seminar), whereas MSW students wi th 3-5 years experience and aged 25-30 had less training within the academic parts o f the social work program. The latter group is the most likely group to encounter client violence. For MSW students with 3-5 years experience and aged 25-30, it appears that h aving less training relates to encountering increased violence. This finding may r elate to assumptions that educators and agencies have about advanced level students wit h a few years experience. It may also relate to the nature of field seminar for BSW students versus MSW students. If discussions in field seminar strictly revolve aroun d case presentations, perhaps there is not enough opportunity to discuss more practical, y et necessary issues such as safety and violence.
221 Client Violence and Occupational Commitment When direct and indirect client violence variables were correlated with the three subscales of occupational commitment, only one of 3 6 possible correlations approached significance. Physical assault was positively corre lated with normative commitment. No previous studies have directly examined the relatio nship between client violence and occupational commitment. One study of nurses indica ted that job tensions are negatively related to career (occupational) commitment (Cohen, A. 1999) and a meta-analysis of 76 studies on occupational commitment indicated that j ob stress is negatively related to occupational commitment (Lee et al., 2000), indicat ing that stressful events may be related to reduced occupational commitment. Convers ely, affective commitment and normative occupational commitment have been found t o be positively related with positive training experiences in student nurses (Me yer et al., 1993). The current study indicates that occupational commitment is largely u naffected by the stressful event of client violence. However, it does indicate when stu dents are physically assaulted they have increased commitment possibly due to a sense o f obligation or responsibility to the profession of social work. The finding that even when students directly or ind irectly encounter client violence, it largely does not affect their occupati onal commitment is somewhat surprising. Only one other study has reported occup ational commitment with students in training and this study found that student nurses affective and normative commitment actually decreased over the course of their trainin g (Meyer et al., 1993). In the current study, the affective commitment for all the social work students was near the maximum
222 affective commitment that could be scored. Because of the timing of the questionnaire and the amount of practicum hours that both the BSW and MSW students noted on the questionnaire, it is presumed that the majority of these students were near the end of their degree program. Thus it could be surmised that as m ost of this group of social work students are ready to graduate, they are extremely proud of their identity as professional social workers. This may speak to the professional socialization of social work students in social work education programs. Something very p ositive must be occurring in social work programs and in field agencies for students to have such a high level of affective commitment, even in the midst of client violence. Client Violence and Career Withdrawal Intentions There were no significant correlations between dir ect or indirect violence and career withdrawal intentions. Two previous social w ork studies have indicated that there may be a relationship between client violence and c areer withdrawal intentions. A study from England looked at career patterns of social wo rk graduates from 4 different years spanning 12 years and found that for the people who had left the profession, work related factors played a major role (Lyons et al., 1995). S pecifically, they found that the 12% of respondents who were still in social work positions and who had experienced violence and abuse had considered leaving the profession as a result and 25% of those who had left the profession stated that experiencing violen ce was an important factor in their decision to leave. A likely reason for the differen ce in findings is that Lyons and colleagues followed social workers several years af ter graduation and the present study questioned them prior to graduation. Lyons and coll eagues pointed out that most of their
223 sample had entered social work after graduation. Th ey stated, people who have just invested a great deal of time and effort to become qualified social workers are very unlikely not to enter the profession after qualifyi ng even if only for a relatively short period of time (p. 179). One student study in the U.S. found that of studen ts who had experienced client violence almost 8% said that they were re-thinking their career in social work as a result of the client violence incident (Knight, 1996). Tho ugh the measures used in this study were different than the ones used in the present st udy, it is possible to draw some comparisons. In the current study of those students who were physically assaulted, none had medium to high levels of career withdrawal inte ntions. Of students who were threatened with physical harm, 9.6% had moderate to high levels of turnover intentions. Approximately 14% of the students who experienced t he other types of client violence had moderate to high levels of career turnover inte ntions. Thus the turnover intentions in this study were somewhat comparable to the turnover cognitions mentioned in Knights (1996) study and the turnover cognitions mentioned in the study by Lyons and colleagues (1995). Client Violence and Fear of Future Violence In the present study, every type of direct client violence was significantly correlated with fear of future violence and indirec t client violence was even more strongly correlated with fear of future violence. T hese findings have been heard and seen in numerous studies. Guy and Brady (1998) reported that fear of future victimization was the most frequently reported emotional consequence of psychotherapists who were
224 victims of physical attack. Song (2005) found that past victimization by clients significantly contributed to fear of future victimi zation. Three qualitative articles all found between 50-90% of victims of client violence who were child welfare workers or mental health workers were fearful of further viole nce at work (Atkinson, 1991; Littlechild, 2005; Snow, 1994). Waters and Morgan ( 2002) stated that with workplace violence there is a tendency to fear returning to t he site of victimization. Newhill and Wexler (1997) found that more than a third of those who had been attacked were scared or felt fear. Similar to the current study, they re ported that twice this amount of workers were scared or fearful, even if they had not been d irectly attacked. Horwitz (2006) found that vicariously experienced events were more stron gly associated with traumatic effects than events directly experienced. The research stat ed that perhaps more dramatic verbal abuse threatened by a client was more promptly and thoroughly processed, which might give the worker more of a sense of control over whe ther the event might reoccur, whereas vicarious events reflect the social workers lack o f control. One social work student client violence study repo rted on several factors related to fear (Knight, 1996). It was found that 90% of th e students who had encountered client violence felt that they would be more cautious in t heir work, 40% felt that they would avoid certain practice situations and 12.9% said th at they would be more fearful in their work. In the current study students who had encount ered client violence were significantly more likely to report more fear of th at specific type of violence. Those who experienced threats of physical harm, threat of law suit, and property damage were twice as likely as those who didnt experience those type s of violence to agree that they were
225 afraid of the same type of violence in the future. Depending on the type of violence, between 24% and 60% of the students who had experie nced a type of violence were afraid of that type of violence occurring again. Th is is substantially higher than the percentage of client violence victims that Knight r eported as having fear in their work. At least part of the reason for the major differences in proportions of students who were afraid of future violence could be because students in the current study were asked about specific fears. Fear of Future Violence and Occupational Commitment No previous studies have explored the possible rel ationship between fear of future violence and occupational commitment. The present s tudy indicates that fear of physical assault and fear of threats of physical harm were n egatively related to affective commitment at probability levels that approach sign ificance, yet fear of every type of client violence was positively and significantly re lated to normative and continuance commitment. Similar to the current study, fear of f uture client violence has been found to be negatively related to organizational affective c ommitment, implying that fear of violence is related to a lessening of pride in and identification with the job (Barling et al., 2001; Rogers & Kelloway, 1997). Though organization al affective commitment and occupational commitment are different constructs, t hey have been found to be positively correlated in 10 articles measuring both in the Nor th America (Meyer, Stanley, Herscovitch, & Topolnytsky, 2003), so while the fin ding above (concerning organizational commitment) may not directly corresp ond with the finding in the current study, it may marginally support the current findin g.
226 In reference to the finding that normative commitm ent was found to be positive while affective commitment was found to be negative one of the authors of the occupational commitment scale used in this study st ates in a meta-analysis of organizational commitment that affective commitment and normative commitment have a strong natural link and typically would be correlat ed positively (Meyer et al., 2003). However, it is acknowledged that there is a possibi lity that an employee can experience an obligation to pursue a course of action (normati ve commitment) in the absence of the desire to do so (affective commitment). The same m eta-analysis indicates that continuance commitment is typically negatively rela ted to affective commitment and normative commitment. The three types of commitment need to be examined in relationship with each other to understand more abo ut the overall occupational commitment. In the current study, it appears that students who have fear of future physical assault and fear of threat of physical harm, whethe r or not they have actually experienced these events, have less feelings of strong identity with the profession of social work. At the same time, they have more normative commitment, or commitment as a result of obligation or responsibility. They additionally hav e more continuance commitment or commitment because they have already invested so mu ch into their degree program and the profession of social work. The overall picture of this contains a cautionary note. When students experience violence, they may not hav e less occupational commitment. However, when they have greater fear of future clie nt violence, especially if they have never actually experienced client violence, their c ommitment to the profession of social
227 work may be coming from a sense of obligation (norm ative commitment), perhaps instilled through professional socialization, and a sense that they have already invested too much time, money and effort to turn back (conti nuance commitment). These latter two types of commitment are unlikely to sustain the m in the profession of social work over the long run. Fear of Future Violence and Career Withdrawal Inten tions No studies have reported specifically on fear of fu ture violence and how it may affect career turnover intentions. Workplace violen ce studies have largely reported on how fear of future violence may affect organization al turnover intentions. One social work study predicted that fear of future violence w ould have a positive direct effect on organizational turnover intentions, however, there was a nonsignificant path between the two (Song, 2005). As with the discussion of organi zational versus occupational commitment, there may be some similarities between organizational turnover intentions and career turnover intentions, but they are not th e same construct. However, it could be noted that the findings of Songs study were simila r to the current study in that fear of future violence did not significantly affect organi zational withdrawal intentions. Fear of Future Violence as Mediator No studies have previously explored the possibility of fear of violence mediating the relationship between client violence and occupa tional commitment. In the current study fear of future physical assault was not found to mediate the relationship between client violence and occupational commitment, becaus e no significant relationship was found between the client violence and occupational commitment, thus there was no
228 relationship to be mediated. Some previous studies have shown that fear of future violence mediates between workplace violence or cli ent violence and decreased organizational commitment (Barling et al., 2001; So ng, 2005). However, other studies have not found fear of future violence to mediate b etween violence and organizational commitment (Rogers & Kelloway, 1997). Previous stud ies have additionally found that fear of future violence mediated the relationship b etween workplace violence and organizational turnover intentions (Barling et al., 2001; Rogers & Kelloway, 1997). In the present study all of the relationships betw een client violence and occupational commitment or career withdrawal intent ions were nonsignificant. Thus there were no existing relationships to be mediated However, other analyses in this study showed that direct and indirect experiences w ith client violence consistently were related to increased fear. Fear was then significan tly related to occupational commitment, though it was not related to career turnover intent ions. to summarize, it cannot be substantiated that fear mediates the relationship b etween client violence and occupational commitment. However, fears of future violence do ha ve significant relationships with occupational commitment in this sample of social wo rk students. Implications for Social Work Education Micro Level Planning and Intervention Field placement and practice considerations. The personal demographic factors in this study that may affect increased exposure to client violence need to be acknowledged by both field directors and field educ ators both during the process of placing students and during the completion of the f ield practicum. Students need to be
229 aware of factors that may tend to increase exposure to client violence. Greater awareness could cue students to use greater caution with the client, especially in situations where there is a felt disequilibrium of power between the client and the social work student. When practicum settings are discussed with student s, field directors need to be able to share what efforts the practicum site has m ade to ensure safety of its employees and practicum students. This of course will require that field agencies inform field directors of these safety measures and policies. Field agencies that require frequent home visits n eed to be responsible to train students specifically on safety precautions to take when visiting a client in their home. When possible, students should accompany more seaso ned social workers when beginning the practice of home visiting. Field directors and instructors should additionall y show due concern when allowing a student to complete a placement predomin antly during evening hours. Students should be aware of possible increased risk of violence, particularly when the practicum setting is a setting more prone to have v iolent encounters during evening hours. Field faculty consultation to students. It may be helpful for field faculty to offer individual debriefing and support to a student who has encountered client violence (Digiulio, 2000). This can supplement any debriefin g that may be done by the agency. Because of the possible negative effect of indirect exposure to violence, it may also be necessary for the field faculty to debrief other st udents who fear becoming victims of violence.
230 Use of field seminar to support students. Discussion about violence should be encouraged in field seminar (Dunkel, Ageson, & Ralp h, 2000). Social work students in the current study had significant fears about both violence that occurred and violence that they had heard about or witnessed and these fears w ere beginning to impact their feelings about being a social worker. Snows (1994) sample o f home visiting social workers noted that having a person available to talk with helped them to process when violence had occurred to them. They recommended the use of suppo rt groups and peer supervision. Knight (1999) found that over 90% of the social wor k students she questioned had talked with other students about their anxieties related t o field. This demonstrates that there is value in the ongoing support that students can rece ive when they meet and talk openly about their fears of victimization. The small amount of students in this study having s afety training in field seminar begs the question of what types of issues are routi nely discussed in field seminar. Additionally, it appears that field seminar may be used for different purposes for MSW students than BSW students, as BSW students have mo re opportunities for safety training in field seminar than MSW students. Perhaps the foc us in MSW field seminars may be more on case specific clinical issues, leaving litt le time for discussion of various dynamics of the field placement. Few places could b e healthier settings to discuss fears of violence and actual violent incidents that have occ urred in the practicum. Field seminar facilitators in both MSW and BSW programs are encou raged to cautiously open the field seminar agenda to include processing of issues rela ted to client violence.
231 Because exposure to graphic violence material may l eave listening students vulnerable to feelings of fear and anxiety, caution should be exercised when allowing this material to be discussed in field seminar. Cunningh am (2004) suggests introducing students to the concept of vicarious traumatization to help them understand possible reactions to trauma materials. She also discusses s everal strategies to balance the need to expose students to traumatic case material while re ducing the risk for other students to be traumatized by listening. One such strategy is for the instructor to elicit students feelings and responses to material presented, inviting react ions and modeling responses to the material. This gives them an opportunity to process any feelings they may be having. Mezzo Level Planning and Intervention Field agency selection. Section 2.1.3 of CSWE curriculum standards notes th at it is the responsibility of the social work field educ ation program to specify policies, criteria and procedures for selecting and maintaining field educators and to evaluate agency effectiveness in providing field instruction (CSWE, 2003). A university may be liable for the selection of field sites and ensuring that the field site offers proper protection and encourages safety precautions (Digiulio, 2001). The social work program should be careful when screening new potential field sites. R eeser and Wertkin (2001) suggest that social work schools place students only in field pl acements that have previously paid attention to the safety of clients, staff and stude nts. It is essential that agencies have policies and procedures for social work safety and a mechanism to orient new workers to these procedures (Rey, 1996).
232 Agency policies and procedures. Adamson found that the agencys response to a crisis determined a more negative impact than the i ncident itself (2006). He further stated that we individualize the impact of stress and tra uma at our peril (p.58). It is essential that the field practice organizations intentionally address ways to both prevent workplace violence and to address the effects when it occurs. It is important that they have environmental safeguards, with a zero tolerance for violence (Reeser & Wertkin, 2001). Additionally, agencies should have a clearly deline ated protocol for handling potentially violent and dangerous situations (Jayarante et al., 2004). When client violence has occurred, agency procedure s must include a formal process for reporting violence (Jayarante et al., 2 004). Policies may also include a consideration of temporary relief of duty or decrea sing the workload (Snow, 1994). The student should be encouraged to seek medical attent ion if necessary. Agency procedures should include practices t hat could help to decrease the likelihood of exposure to client violence. Suggesti ons have included having all students to carry a cell phone with them during the time the y are in the practicum and doing home visits during regular business hours (Castellanos, 1998). Agency procedures should recognize the importance of creating a safe office space and the establishment of offices where there is easy access to help and escape (Jaya rante et al., 2004). Ongoing safety training within the agency. Many field educators have stated that they did not receive adequate training or preparati on on dealing with personal safety issues prior to entering the field. This has also b een found to be true with social work
233 practitioners in general. This being the case, scho ols of social work can develop continuing education opportunities related to safet y for practitioners (Tully et al., 1993). It is important that agencies provide formal staff meetings and/or trainings to discuss client violence on a regular basis. It is r ecommended that safety training be updated at least annually (Mama, 2001; Newhill, 199 6; Vergaras, 2006). It has been found that frequent training can increase the possi bility of avoiding assaults (Carmel & Hunter, 1990). In their study on client violence to ward child welfare workers, Horejsic and Garthwait (1994) stated that training once a ye ar would constantly remind staff of potential dangerous situations and procedures for h andling them. Social work educations curriculum on violence. Some researchers have suggested requiring a specific course to address vi olence and safety issues (Reeser & Wertkin, 2001). Perhaps realizing the difficulty of adding another course to an already full curriculum in many schools, some have suggeste d the alternative of infusing this content across the social work curriculum (Reeser & Wertkin, 2001; OKeefe & Mennan, 1998; Tully et al., 1993). Maidment (2003) proposed a field education work safety curriculum that could be implemented across two yea rs of field placement. This plan suggests specific work safety content and assignmen ts that can be completed beginning in pre-placement and continuing at various intervals t hroughout the field placement. Ellison (1996) suggested this subject could be discussed in practice courses, human behavior courses and policy courses, believing that exposure in the classroom would better prepare students to face the realities of practice. This w ould require program faculty to
234 coordinate their efforts in this area. They should know how safety issues are taught in the field practicum, as well as in other courses (Faria & Kendra, 2007). Content of safety training curriculum. This study surprisingly found that increased training in the various parts of the soci al work program was related to increased client violence. It appears that this finding may b e partially related to practicum settings. Those practicum settings that are more dangerous ma y be the same agencies that provide more training, especially in content areas related to self defense or aggression control. Sarkisian and Portwood (2002) point out that train ing methods suggested to date may serve to strengthen the workers defenses and begin to resolve the discrepancy between reported and actual incidents against workers, but such measures in no way reduce the likelihood that workers will become the target of v iolence (p. 48). They go on to suggest that helping social workers develop a more thorough understanding of environmental and systemic influences that may heig hten the chances of client violence may give a more realistic picture of client violenc e. They recommend that social work practitioners and students need to understand more about systemic barriers to client empowerment and where opportunities may exist for c lient empowerment as a major mechanism for reducing violence. This raises challenging choices concerning safety training for social work students. Do we continue to design or support train ings that are oriented toward self defense in agencies that we know to be more dangero us, thus placing the bulk of responsibility for protecting self in the hands of the workers themselves? Or do we design more training that is designed to teach practitione rs and students how to empower clients,
235 thus equalizing the power and authority in worker-c lient relationships, perhaps reducing the possibility of client violence? It seems that we must do both. Social work students need to continue to be taught how to disarm threate ning situations by using empowerment techniques. However, the reality of dan ger cannot be overlooked. In situations where students may encounter more violen ce, students may need to be exposed to self defense strategies that place greater empha sis on verbal de-escalation. Numerous safety training content areas have been su ggested by previous researchers. Many of the suggestions for training c ontent could serve the purposes of both teaching practitioners how to empower clients to re duce violence and as necessary, equipping practitioners to protect themselves in th e event of impending violence. Training suggestions below are arranged in a contin uum beginning with client empowerment and proceeding towards ways to respond to violence when it has occurred. Assessing and understanding of potential lack of cl ient empowerment and opportunities for client empowerment Skills of assessmentPresent experiences and histo ry of violence in clients (Newhill, 1996; OKeefe& Mennan, 1998) Developing an understanding of the precursors of vi olencefactors that might lead a client to act violently (Digiulio, 2001; Jay arante, Croxton, & Mattison, 2004; Reeser & Wertkin, 2001; Rey, 1996; Ringstad, 2005) Knowing community resources (OKeefe & Mennan, 1998 ) Advocating on behalf of clients (OKeefe & Mennan, 1998) Advocating effectively for practice conditions favo rable for violence prevention (Newhill, 1996)
236 Understanding how violence may affect victims (Reac tions to potential violence) (OKeefe & Mennan, 1998) Learning how to cope actively with stresses; Stress management procedures (Jayarante, Croxton, & Mattison, 2004; Jayarante, Davis-Sacks, Chess, 1991) Managing feelings that can arise when working with victims and perpetrators of violence (Rey, 1996). OKeefe and Mennan state t hat the student must be able to deal with the immense emotionality of issu es around violence (p. 95, 1998). Examining ones own value system around violenceW hat are the students assumptions regarding victims and perpetrators? (O Keefe & Mennan, 1998) Understand how ones own culture impacts clients e xposure to violence (OKeefe & Mennan, 1998; Adamson, 2006; Ringstad, 2 005) Gaining ability to manage ones own anger (Reeser & Wertkin, 2001) Understanding of students rights ( for example, t he right to refuse to make a home visit) (Faria & Kendra, 2007) Understanding the dynamics of violence Recognizing theories of violence and having the ab ility to apply those theories (OKeefe and Mennan, 1998; Rey, 1996) Understanding the prevalence of different forms and types of violence (Digiulio, 2001; OKeefe & Mennan 1998; Reeser and Wertkin, 2001; Rey, 1996) Understanding the cycle of violence (OKeefe & Men nan, 1998) Understanding the use and misuse of power and the d ynamics to various types of violence (OKeefe & Mennan, 1998; Ringstad, 2005 ) Direct violence prevention tools Be able to use techniques on verbal intervention an d de-escalation through use of non threatening communication skills; Effective strategies for working with angry, hostile clients (Digiulio, 2001; Jayarante, Croxton, & Mattison, 2004; Knight, 1999; Newhill, 1996; Reeser & Wertkin, 2001 )
237 Have knowledge of office safetylearning to arrang e work space to maximize safety (Digiulio, 2001; Rey, 1996) Learn protocols on home visit safety (Digiulio, 200 1; Newhill, 1995; Reeser & Wertkin, 2001; Rey, 1996) If desired, physical techniques for self protectio n; Self defense; How to defend ones self in potentially dangerous situatio ns (Castellanos, 1998; Digiulio, 2001; Knight, 1999) Addressing fear of violence or actual violence Be able to use supervision and consultation to add ress issues of client violence (Ringstad, 2005) Recognize the value of the peer support group in c reating an atmosphere of acceptance and support (Newhill, 1995) What to do in the aftermath of a client violence in cidentDebriefing and support after an incident (Digiulio, 2001; Faria & Kendra, 2007; Reeser & Wertkin, 2001; Rey, 1996) Developing resourcefulness and resilience through s elf advocacy and negotiation skills (Adamson, 2006) Learn procedures for formal recording incidents of violence (Ellison, 1996; Reeser & Wertkin, 2001) Methods of training on safety and violence. The current study indicates that more training may be associated with increased violence. A possible explanation for this is that the training is simply not effective. A synthesis o f literature on implementing evidence based practices indicates that practitioners at imp lementation sites need to know when, where, how, and with whom to best use new skills (F ixen, Naoom, Blas, Friedman, & Wallace, 2005). Joyce and Showers (as cited in Fixe n el al., 2005) found in reviewing numerous studies on training effectiveness that eff ective training appears to consist of presenting knowledge about a subject, providing a d emonstration, and assuring
238 opportunities to practice the skills in the trainin g setting. The nursing profession has produced several articles on aggression management training effectiveness. One particular model of aggression management training was a three day training that was conducted with first year nursing students (Beech & Leather, 2003). It included knowledge on the subject as well as opportunity to practice the skills. When evaluated longitudinally, the training was found to be effect ive in sustaining prevention and good management of aggression issues with patients. Short of using a specific training seminar to addr ess and practice safety skills, other pedagogical methods may be attempted in socia l work classrooms to reinforce learning about safety and violence issues. Research ers who have written about how to address client violence in the classroom suggest a range of experiential activities. Methods may include: Simulations and role play (Guy & Brady, 1998; Horej sic & Garthwait, 1994; OKeefe & Mennan, 1998) Dramatic re-enactment ( Guy & Brady, 1998) Detailed discussion about cases (Guy & Brady, 1998; OKeefe & Mennan, 1998) Small group exercises (OKeefe & Mennan, 1998) Discussion about field experiences (OKeefe & Menna n, 1998) Use of excerpts from popular dramas to illustrate c auses of violence and emotions surrounding violence (Rey, 1996) Use of actors, both live and on video, to perform c lient violence situations in social work situation (Leadbetter & Phillips, 19 90) Bring agency-based social workers into the classroo m to explore the realities of violence in social work (Adamson, 200 6)
239 Provide students with reality bytes of context, s ocial work issues, and tensions (Adamson, 2006) Training for field educators. It is the responsibility of social work programs to provide orientation, field instruction training, a nd continuing dialog with agencies and field instructors (CSWE, 2.1.5, 2003, p. 38). Init ial field educator orientation should ideally take place before a student is placed at an agency. This orientation should have a component on client violence and safety precautions (Reeser & Wertkin, 2001). Field educators should be encouraged to talk with student s about their anxieties when they enter the placement (Knight, 1999). Additionally, t hey should be encouraged to develop ways to help the student balance safety against pro fessional responsibilities and obligations (Knight, 1996). Suggestions for trainin g of field educators include: Ability to validate reactions and feelings when stu dents have been exposed to client violence (Snow, 1994) Being aware of students individual experience and assessing student readiness for specific field activities (Dunkel et al, 2000; Vergara, 2006) The need to consider case transfers, as necessary ( Vergara, 2006) Specific preparation for home visiting (Dunkel et a l, 2001) to include: o Knowing the history of the family o Knowing who is likely to be home o Deciding whether two or more people should make the home visit (Allen & Tracy, 2004) Awareness of potential violent situations and clien ts and sharing this information with students (Guy & Brady, 1998)
240 Establishment of a trusting/open relationship with the student, thus creating the possibility of offering advise, suppor t and protection through the mentoring process (Guy & Brady, 1998; Reeser & Wertkin, 2001) Being familiar with all of students cases and assi gning cases appropriately (Dunkel et al, 2001; Reeser & Wertkin 2001). Being aware of students whereabouts (Reeser & Wert kin, 2001) Consider providing worker shadowing during the init ial phase of the internship (Vergara, 2006) Gaining ability to help students cope with fear of future violence as an ongoing long-term process when client violence has occurred Being aware of the possibility of the effects of in direct exposure to client violence, particularly when more than one intern is in an agency and/or when group supervision is offered. Field educators need to be made aware that students may have some discomfort and concern regarding the racial mix of the supervi sion dyad that could impact all of the functions of the supervisor. This could be discusse d openly in field educator orientation to increase awareness. Cultural competency of both the field educator/supervisor and the student/employee could also be the subject of conti nuing education training for the many social work practitioners who are in the role of su pervisor and/or field educator. In situations where a non-White supervisor may be supe rvising a non-White student, field directors can lend added support to the supervision process. Macro Level Planning and Interventions Social work program policies. It is essential that social work programs have written policies regarding the risks and benefits o f field learning (Gelman, 1988; Faria & Kendra, 2007). Zakutansky and Sirles (1993) state t hat field agencies need to clarify with students if the student will have personal injury i nsurance. They further noted that
241 students who are entering high risk situations need to be apprised of who will accept financial responsibilities should any damages be in curred. Gelman additionally stated that students informed consent should be sought regardi ng possible harmful situations in the field. Gelman stated that learning can only be enh anced if students become informed participants (p. 77). School policies should contain a clear definition of client violence (Faria & Kendra, 2007). This is a necessary part of informin g students of foreseeable dangerous situations (Zakutansky & Sirles, 1993). A formal reporting system should be designated in o rder to insure that all encounters with client violence are reported (Ellis on, 1996; Faria & Kendra, 2007; Ringstad, 2005). Horejsic and Garthwait (1994) reco mmended specifically that policies include a stipulation that all threats against a wo rker should be reported to the supervisor. In the field placement, this should include reporti ng to both the field educator and the field faculty. Furthermore, protocols should be dev eloped as to when threats should be reported to law enforcement. Policies should also require schools of social wor k to maintain records of client assaults to social work students (Mama, 2001; Rings tad, 1995). This would help social work educators to better know the prevalence of the problem. CSWE policies. CSWEs educational policy requires that social work education contain a coherent, integrated professional founda tion in social work practice (p.37). It has been suggested that personal safety content ought to be mandated for inclusion in the required content areas by CSWE (Tully et al., 1 993). To date, such content is not
242 mandated. This study illustrates that left to using their own discretion on implementing effective safety training for students, many social work programs either do not train their students at all, provide small amounts of training, or the training is so insignificantly discussed that students may have no recall of the c onversation. Social work program administrators and field directors need to continue to encourage CSWE to mandate some training in the area of safety and client violence. Until such a time that such content is mandated, it remains responsibility of individual s ocial work programs to include and appropriately emphasize information about personal safety in their curriculum. Federal policy on social work safety. In the United Kingdom, there is a National Action Plan on Violence against Social Care Staff, which was recognized to be a step toward a central policy on violence (Brockmann, 200 2). In the United States, at least two states have implemented policies to insure the prot ection of social workers. The Michigan Social Welfare Act of 2001 was passed in response t o the brutal murder of a county child protection worker (Sarkisian & Portwood, 2003). It included laws that mandated home visit training, use of buddy systems where high ris k has been predetermined, and criminal penalties for clients who exhibit violence towards social workers. More recently, West Virginia passed a state law that includes protectiv e services workers and health care workers in a class of workers protected though incr eased penalties for those committing felony or misdemeanor assault against the workers w hen they carrying out work duties (Pace, 2008). Unfortunately, this new legislation w as also in direct response to yet another social worker who was viciously murdered wh ile making a home visit.
243 The United States does not have a national policy c oncerning violence to social workers and other human service workers. In 2005, U nited States Congressman Dennis Moore announced a resolution to raise awareness abo ut potential job-related violence against social workers and case workers. The resolu tion encouraged state and local agencies to improve the safety of social service wo rkers. The resolution honored the memory of Teri Zenner, a social worker who was stab bed to death by her 17 year old client while making a home visit on August 17, 2004 (Sedensky, 2004). Teri Zenner was a 26 year old graduate student at the School of Soc ial Welfare at the University of Kansas at the time of her death (University of Kansas, 200 4). Representative Moore followed this resolution with a proposed bill that would award gr ants to states to provide safety measures such as GPS tracking devices, facilities s afety improvements, and safety training for social workers and other helping profe ssionals who work with potentially violent clients (Teri Zenner Social Work Safety Act 2007). This bill was not passed by the 110th Congress. Social work schools are in a unique position to pro vide information about social work safety to Congressional representatives. Polic y classes could provide a good forum to encourage advocacy for a federal and state polic y in this area. Surely we need not wait for another brutal murder of a social worker to urg e state and federal policies for social work protection. Coordination of safety training throughout the soci al work program. Though this study found that increased training was related to increased client violence, this is not license to discontinue or decrease the amount of tr aining provided to social work
244 students. Social work students in the current study recalled having a very small amount of safety training in their social work program. They had more training in their social work practice classes than in field seminar, but even in social work classes, many students had little to no training on safety for social workers. It appears that various parts of the social work program may not be aware of what the other par ts are doing to train students regarding safety and violence, as evidenced by the low percentages of students receiving training in every venue. This gives the impression that perhaps each training venue has trusted the other to provide training, thus excusin g them from the responsibility of safety training. The end result is that students are not b eing adequately trained anywhere in the social work program. It has been found that pre-ser vice and in-service training may be efficient ways to provide knowledge of a new skill, but most new skills are best learned with the help of a consultant/coach (Fixen et al., 2005). This lends support to the idea of doing training in social work classes, which can be followed by discussions in field seminar, when students may have actually directly o r indirectly encountered client violence. Additionally, the consultant/coach in the case of social work practicum is the field educator. When the student is on the job in t he internship, the field educator can reinforce previous training in the academic sector of the social work program and at the field agency with ongoing discussions and support a bout social worker safety. Efforts must be made to coordinate safety training that is infused across the curriculum, in standalone trainings or conferences within or outside of the program, in field seminar, and at the field agency practicum sites.
245 Risk management concerns. Social work programs and field agencies have a responsibility to students to protect their well be ing to the extent that this is possible. Both must also be aware of potential financial liab ilities that may exist when students safety needs have not been adequately anticipated. Nuehring and Houston (1992) note that risk management seeks to protect the provider s financial assets, along with human and intangible resources. They further stated that risk management means prevention, early detection, and immediate intervention in inju rious situations in the organization, in order to eliminate or minimize the risk of human ha rm or loss of resources (p. 58). The first level of administrative response in risk mana gement plans requires that administrators need to create measures that focus o n both staff safety and meeting client needs. Recommended strategies include safety traini ng. Though safety training is not a panacea through which all students will miraculousl y avoid violent encounters, social work programs and field agencies that do not adequa tely train their students fail to do so at their own peril. A precedent has been set for legal litigation towa rds programs and field agencies that may not adequately attempt to protect students A 1995 Florida case was brought by a 23 year old female doctoral student placed at a f ield agency that she knew to be in a dangerous neighborhood (Nova Southeastern vs. Betha ny Jill Gross, S.C. Case No. 94,079). Upon beginning at the agency she was told that the agency director had recently been robbed at gunpoint in the agency parking lot. She was given a manual which included safety precautions and it was suggested th at she use a buddy system when leaving the building in the evening. Six months aft er her arrival she was robbed and raped
246 in the parking lot of the agency. She filed suit ag ainst the agency and the university. The field agency settled out of court for $900,000. The university continued to fight the suit through several appeals that lasted for at least 4 years. At one point the court found the university liable for her injuries because they had a duty beyond the duty to warn. They were judged to have a duty to take reasonable cauti ons to protect the student in her internship. Regardless of the legal outcome, any la wsuit is costly in terms of the time, energy and financial costs to an organization. Partially in response to the university case above Yeomans (2004), an attorney, suggests that service learning program directors an d faculty take all possible steps to prevent injuries from occurring. She stated that i f injuries do occur liability can be avoided by showing that reasonable actions were tak en. Some practices to consider are: Careful selection of volunteer sites and activities allowing students to choose from a list of options Research the potential risks and dangers involved w ith volunteer sites and activities. (Create a questionnaire for community p artners to complete, having them disclose possible risks and safety procedures, insurance coverage, etc. Warn students in writing of known or potential dang ers involved (as part of liability waiver) Keep records to show youve done your due diligence Define volunteer roles and make sure students know the extent of their volunteer assignments. Develop safety practices/procedures and train volun teers (or be sure volunteer site does this.) Include descriptions of course required volunteer a ctivities in course description and syllabus.
247 For certain tasks, you may need to obtain volunteer references and/or background checks. Inspect community partners premises for safety con cerns. Communicate with community partners about potential hazards and safety procedures. Keep records to show how student volunteers are sup ervised at the site. Allow students a choice in volunteer assignments. Execute liability waivers and informed consent cont racts. (p.7) As large as financial responsibility could be for a university or a field agency, social work education programs primary responsibil ity must continue to be for protection of human life, both clients and social work students. Study Limitations Though the 45% response rate was better than most other national social work education studies and better than the 30% projected response rate, it still may be difficult to generalize the findings to the larger population of social worker students. Social work students who chose not to participate may not have possessed the same characteristics of the participants. There may have been unidentified intervening variables/factors that influenced the participants response or nonrespons e. The reliability of self-report information may have been an issue, as it sometimes is in survey research. Recall of events may have be en inaccurate and may have biased the research. However, because students were asked to r ecall incidents during their practicum and most were still in their practicum, the time be tween the recalled incidents and the questionnaire was minimized.
248 Data reliability may have been threatened by soci al desirability bias. It is possible that the students might have attempted to answer qu estions in a way that they thought was desirable for a student who is training to be a professional social worker. Studies indicate that social workers may be reluctant to re port client violence for fear of what their fellow workers and supervisors may think of t hem. Additionally, they may be reluctant to acknowledge if they are thinking about leaving the profession. This study had numerous significant findings, but the effect sizes on most of the analyses were small and in a few cases, moderate si zed. This indicates that though knowledge of most of the independent variables can be generalized to the population, it is of little help in predicting client violence. Some of the findings are likely attributed to the large sample size. The cross sectional design of this study prohibits making any causal statements about the data in this study. Interpretations of th e data are limited to descriptive information. Suggestions for Future Research Since most social work student client violence stu dies have not included MSW students, more research is needed to document MSW s tudents experiences with violence. Additionally, more large scale random stu dies are encouraged as these studies have more potential for generalizing to the larger population of social work students. Studies should attempt to replicate the questions a bout personal demographics in this student study as this study demonstrates some diffe rent findings than social work practitioner client violence studies.
249 It may be beneficial for social work client violen ce studies that draw a random sample from the NASW membership roster to include s ocial work students in the sample. Most practitioner studies have deliberately exclude d student social workers. Including students might give a more comprehensive view of so cial workers in every stage of career development. It would be of interest to complete a client viole nce study with social workers who are just beginning their career post graduation This type of study would have increased strength and reliability if a longitudina l study could be completed, following students through their academic and field preparati on and into the early years of their career. This study indicates that social work stude nts have very high affective commitment to the profession of social work. More i nformation is needed to know if this level of commitment can be sustained in the absence of the protected environment that the social work program may provide. Social work client violence studies need to make d eliberate effort to oversample social workers and/or social work students from eth nic minorities. Since the current study indicated that Native American students have higher rates of client violence, it may be beneficial to deliberately gather information from Native American students. A qualitative study might allow for more specific foc us on Native Americans experiences with client violence. Similarly, no American client violence studies hav e been directed toward social workers in developmental disability settings and fe w studies have included a practice category for alcohol/substance abuse. Since this st udy indicated that workers in these
250 settings experience high rates of violence, it may be beneficial to make a deliberate effort to sample workers from these practice settings. Aga in, qualitative inquiry with workers in these practice settings may help to better understa nd the essence of their experience with client violence. Certainly, more study is needed in general about t ransracial supervision, especially as it relates to social work. The litera ture is woefully scant on this subject. Since social work students are expected to spend on e hour a week in supervision and students are expected to effectively use supervisio n, further questions need to be asked about the strengths and issues associated with tran sracial supervision. In future studies questions need to be asked about place and time of client violence in a way that will permit inferential anal ysis. It is important for field directors and field educators to know if particular places or times lend to the possibility of students experiencing client violence. Much more information is needed about safety train ing in social work education programs to more accurately address how training mi ght be able to help reduce incidents of client violence. Since there is a dearth of soci al work literature on safety training effectiveness, there is a need to explore other occ upational literature to discover what factors increase training effectiveness. The follow ing questions would be beneficial to explore: How is training offered in the social work curricul um? Is it offered in a specific safety unit, is it infused across the social work c urriculum, or both? What pedagogical techniques are being used to teach about safety and client violence?
251 What efforts are being made in social work programs to coordinate the training received across the curriculum and in field seminar ? There is additionally a need to continue to seek in formation concerning the safety and violence training provided by field agencies. Only one study has addressed this directly with the field agencies, yet in other social work p rogram studies, many of the field directors have no knowledge of whether field agenci es have safety and violence training. Questions might include the following: What safety content areas are covered in agency tra ining? How do students receive training in the field agenc y? (Do students have the training via a classroom setting, on video, or thou gh policy/procedural manuals?) Is there any opportunity to interact with other sea soned practitioners during the safety training? Do field agencies provide or require aggression con trol training? More study is needed on the development and sustain ing of occupational commitment to the profession of social work. Ideally, longitud inal studies need to be done on commitment at the beginning of the educational proc ess and commitment near the end of the educational process. Study on occupational comm itment may also contribute to knowledge about professional socialization of socia l workers. Though one study has been done on fear mediating th e relationship between client violence and organizational commitment, further stu dies are needed to confirm what psychological reactions to client violence may infl uence organizational commitment, occupational commitment, and career turnover intent ions. As Song (2005) recommended, social support could be investigated as a possible moderator between client violence and emotional or psychological responses to the
252 violence. Specifically, as this relates to social w ork education, more information is needed on what types of support are helpful to stud ents to reduce fear and anxiety that may be experienced in reaction to client violence. Conclusion As Adamson (2006) concluded, heightened awareness o f risk and response may not control the occurrence of client violence to so cial work students, but the awareness may bring increased competence appraisal, knowledge of systems that affect the student, and a strengthened level of resilience. We must rem ain committed to the expansion of knowledge and safety skills of social work students Perhaps in doing so, we will help to create more competent practitioners who will have t he ability to sustain their commitment to the profession of social work. Certainly much remains to be done in terms of resea rch, policy and practice concerning the issue of client violence towards soc ial workers and social work students. We must not grow calloused concerning workers exper iencing client violence. If some of these incidents occurred to the clients who are ser ved in social work, there would be outrage, demand for appropriate rehabilitation and/ or penalties for the perpetrators, and never failing dedication to the amelioration of the problems that caused the violence. May we do no less for our fellow social workers, partic ularly those who about to become the newest members of the profession.
253 References Acker, G. (2004). The effect of organizational cond itions (role conflict, role ambiguity, professional development, and social support) on j ob satisfaction and intention to leave among social workers to leave among social w orkers in mental health care. Community Mental Health Journal, 40 65-73. Adamson, C. (2006). Stress, trauma, and critical in cidents: the challenge for social work education. Social Work Review, 5058. Allen, N., & Meyer, J. (1990). The measurement and antecedents or affective, continuance, and normative commitment to the organ ization. Journal of Organizational Psychology, 63, 1-18. Allen, S., & Tracy, E. (2004). Revitalizing the rol e of home visiting by school social workers. Children and Schools, 26 197-208. American Federation of State, County, and Municipal Employees (n.d.). Violence in the workplace. Retrieved on January 8, 2007 from www:afscme.org/publications/2747.cfm Arnold, J. (1990). Predictors of career commitment; A test of three theoretical models. Journal of Vocational Behavior, 37 285-302. Aryee, S., Chay, Y., & Chew, J. (1994). An investig ation of the predictors and outcomes of career commitment in three career stages. Journal of Vocational Behavior, 44, 1-16. Astor, R., Behre, W., Wallace, J., & Fravil, K. (19 98). School social workers and school violence: Personal safety, training, and violence programs. Social Work, 43 223233. Retrieved January 25, 2005 from Ebscohost data base. Atkinson, J. (1991). Worker reaction to client assa ult. Smith College Studies in Social Work, 62, 34-42. Baines, D. (2004). Losing the eyes in the back of our heads: Social service skills, lean caring, and violence. Journal of Sociology and Social Welfare, 31 31-50.
254 Balloch, S., Pahl, J., & McLean, J. (1998). Working in the social services: Job satisfaction, stress, and violence. British Journal of Social Work, 28 329-350. Baretti, M. (2004). What do we know about the profe ssional socialization of our students? Journal of Social Work Education, 40 255-283. Barling, J. (1996). The prediction, experience, and consequences of workplace violence. In Bulatao, E. & VandenBos, G. (Eds.) Violence in the workplace American Psychological Association: Washington D.C. pp. 2949. Barling, J., Rogers, G., & Kelloway, K. (2001). Beh ind closed doors: In-home workers experience of sexual harassment and workplace viol ence. Journal of Occupational Health Psychology, 6, 255-269. Baron, R., & Kenny, D. (1986). The moderator, media tor-mediator variable distinction in social psychological research: Conceptual, stra tegic, and statistical considerations. Journal of Personality and Social Psychology, 51 1173-1182. Beaver, H. (1999). Client violence against professi onal social workers: Frequency, worker characteristics, and impact on worker job s atisfaction, burnout, and health. (Doctoral dissertation, University of Arka nsas, 1999). Retrieved from Proquest Dissertations, UMI Number: 9932738. Bedeian, A., Kemery, K., & Pizzolatto, A. (1991). C areer commitment and expected utility of present job as predictors of turnover i ntentions and turnover behavior. Journal of Vocational Behavior, 39 331-343. Beech, B., & Leather, P. (2003). Evaluating a manag ement of aggression unit for student nurses. Journal of Advanced Nursing, 44 603-612. Bernstein, H. (1981). Survey of threats and assault s toward psychotherapists. American Journal of Psychotherapy, 35 542-549. Bibby, P. (1994). Personal safety for social workers Aldershot, Arena. Blankertz, L. & Robinson, S. (1997). Turnover inten tions of community mental health workers in psychosocial rehabilitation services. Community Mental Health Journal, 33, 517-529. Blau, G. (1985). The measurement and prediction of career commitment. Journal of Occupational Psychology, 58, 277-288. Blau, G. (1988). On the aggregation of individual w ithdrawal behaviors into larger multi-item constructs. Journal of Organizational Behavior, 19 437-451.
255 Blau, G., Paul, A., & St. John, N. (1993). On deve loping a general index of work commitment. Journal of Vocational Behavior, 42 298-314 Blau, G., Tatum, D., & Ward-Cook, K. (2003). Correl ates of professional versus organizational withdrawal cognitions. Journal of Vocational Behavior, 63 72-85. Breakwell, G. & Rowett, C. (1988). Violence and social work In K. Brown (Ed.), Human aggression: Naturalistic approaches. (pp. 230-258). London: Routledge. Brierly, J. (1996). The measurement of organization al commitment and professional commitment. The Journal of Social Psychology, 136 265-267. Brockmann, M. (2002). New perspectives on violence in social care: Journal of Social Work, 2, 29-44. Brown, R., Bute, S., & Ford, P. (1986). Social workers at risk: the prevention and management of violence. Basingstoke, Macmillan. Budd, J., Arvey, R., & Lawless, P. (1996). Correlat es and consequences of workplace violence. Journal of Occupational Health Psychology, 1 197-210. Carmel, H., & Hunter, M. (1989). Staff injuries fro m inpatient violence. Hospital and Community Psychiatry, 40, 41-46. Carmel, H., & Hunter, M. (1991). Psychiatrist injur ed by patient attack. Bulletin of the American Academy of Psychiatry and the Law, 19, 309-316. Caselman, T., and Brandt, M. (2007). School social workers intent to stay. School Social Work Journal, 31, 33-48 Castellanos, D. (1998). Personal safety concerns: S ocial workers perceptions about working in the field. (Masters Thesis, California State University, Long Beach). Retrieved from Proquest Dissertations UMI Microform 1391600. Cohen, A. (1999). Relationship among five forms of commitment: An empirical assessment. Journal of Organizational Behavior, 20 285-308. Cohen, B. (1999). The willingness to seek help: Its role in social workers professional commitment. Journal of Sociology and Social Welfare, 26 21-35. Cook, D., & Helms, J. (1988). Visible racial ethnic group supervisees satisfaction with cross-cultural supervision as predicted by relatio nship characteristics. Journal of Counseling Psychology, 35, 268-274.
256 Council on Social Work Education. (2006). Statistic s on social work education in the United States: A summary. Council on Social Work E ducation: Alexandria,VA. Council on Social Work Education Commission on Accr editation. (2003). Handbook of accreditation standards and procedures. Council on Social Work Education: Alexandria, VA. Cunningham, M. (2004). Teaching social workers abou t trauma: Reducing the risks of vicarious traumatization in the classroom. Journal of Social Work Education, 40, 305-317. Davey, B. (1999). Discrimination at work, in Balloc h, S., McLean, J., & Fisher, M. (eds.), Social services: Working under pressure. The Policy Press: Great Britain. Dillman, D. (2000). Mail and internet surveys: the tailored design meth od John Wiley & Sons: New York. DiGiulio, J. (2001). The power of collaboration: De veloping a safety training program for student interns. Journal of Baccalaureate Social Work, 7 70-77. Dunkel, J., Ageson, A., & Ralph, C. (2000). Encount ering violence in fieldwork; A risk reduction model. Journal of Teaching in Social Work. 20 5-18. Ellison, M. (1996). Field can be hazardous to your well-being: Fact or fiction? Journal of Baccalaureate Social Work, 2, 79-89. Elwood, A., & Rey, L. (1996). Awareness and fear of violence among medical and social work students. Family Medicine, 28 488-492. Faria, G., & Kendra, M. (2007). Safety education: A study of undergraduate social work programs. Journal of Baccalaureate Social Work, 12 141-153. Fishbein, M., & Ajzer, I. (1975). Beliefs, attitude s, intentions, and behavior. Reading, MA: Addison-Wesley. Fixen, D., Naoom, S., Blase, K.,Friedman, K., & Wal lace, F. Implementation of the research: A synthesis of the literature University of South Florida: Tampa, FL Fong, M., & Lease, S. (1997). Cross-cultural superv ision: Issues for the white supervisor. In D. Pope-Davis & H. Coleman (Eds.) M ulticultural counseling competencies assessment, education, training and s upervision. Thousand Oaks, CA: Sage Publications.
257 Freund, A. (2005). Commitment and job satisfaction as predictors of turnover intentions among welfare workers. Administration in Social Work, 29 5-21. Gelman, S., & Wardell, P. (1987). Whos responsible ? The field liability dilemma. Journal of Social Work Education, 7 0-78. Gifford, E. (2003). An examination of organizationa l and professional commitment among public, not-for-profit, and proprietary soci al service employees. Administration in Social Work, 27 5-23. GovTrack.us.H.R. 2165110th Congress. (2007). Terri Zenner Social Work Safety Act. Retrieved from GovTrack.us at http://govtrack/congress/?bill=h110-2165 on October 15, 2008. Green, R, Gregory, R., & Mason, R. (2003). Its no picnic: Personal and family safety for rural social workers. Australian Social Work, 56, 94-106. Green, S. (1991). How many subjects does it take to do a regression analysis? Multivariate Behavioral Research, 26 499-510. Guterman, N., Jayarante, S., & Bargal, D. (1996). W orkplace violence and victimization experienced by social workers: A cross-national st udy of Americans and Israelis, In VandenBlos, G. & Bulatao, E. (Eds.) Violence in the workplace American Psychological Association: Washington D.C. pp. 175 -188. Guy, J., & Brady, J. (1998). The stress of violent behavior for the clinician. In P. Kleepsie (Ed.) Emergencies in mental health practice: Evaluation a nd management (pp. 398-417). Guy, J., Brown, C., & Poelstra, P. (1990). Who gets attacked?: A national survey of patient violence directed at psychologists in clin ical practice. Professional Psychology: Research and Practice, 21, 493-495. Hall, D. (1971). A theoretical model of career subi dentity development in organizational settings. Organizational Behavior and Human Performance, 6 50-76. Harvey, M. (1996). An ecological view of psychologi cal trauma and trauma recovery. Journal of Traumatic Stress, 9, 3-23. Heppner, P., & Heppner, M. (2004). Writing and publishing your thesis, dissertation, & research: A guide for students in the helping pr ofessions. United States: Thomson.
258 Hilton, D., Russell, R., & Salmi, S. (1995). The ef fects of supervisors race and level of Support on perceptions of supervision. Journal of Counseling & Development, 73, 559-563. Horejsic, C., & Garthwait, C. (1994). A survey of t hreats and violence directed against child protection workers. Child Welfare, 73, 173-180. Retrieved 2/16/05 from Ebscohost database. Horwitz, M. (2006). Work-related trauma effects in child protection social workers. Journal of Social Services, 32 1-18. Jayaratne, S., Croxton, T., & Mattison, D. (2004). A national survey of violence in the practice of social work. Families in Society, 85 445-454. Jayarante, S., Davis-Sacks, M., & Chess, W. (1991). Private practice may be good for your health and well being. Social Work, 36 224-229. Jayarante, S., Vinokur-Kaplan, D., Nagda, B., & Che ss, W. (1995). A national study on violence and harassment of social workers by client s The Journal of Applied Social Sciences, 20, 1-13. Jones, K., Chen, H., Jordan, N., Boothroyd, R., Ram oni-Perazzi, J., Shern, D. (2006). Examination of the effects of financial risk on th e formal treatment costs for a Medicaid population with psychiatric disabilities. Lippencott Williams & Wilkins. Kadushin, A. (1976). Supervision in social work New York: Columbia University Press. Kenny, D. (2008). Mediation. Retrieved July 14, 2008 from http://davidkenny.net/cm/mediate.htm Knight, C. (1996). The implications of BSW students experiences with danger in the field practicum. Journal of Baccalaureate Social Work, 2, 133-149. Knight, C. (1999). A study of BSW students percept ions of and experiences with risks to their personal safety in the field practicum. Journal of Baccalaureate Social Work, 4, 91-108. Kraus, J., & McArthur, D. (1996). Epidemiology of v iolent injury in the workplace. Occupational Medicine, 11 201-217. Landsman, M. (2001). Commitment in public child we lfare. Social Service Review, 75, 386-419.
259 Lazar, A., Cohen, B., & Guttman, D. (1995). Profess ional commitment among graduating BSW students in Israel. International Social Work, 38, 341353. Lazarus, R., & Folkman, S. (1984). Stress, appraisal, and coping New York: Springer. Leadbetter, D. (1993). Trends in assaults in social work staff: the experience of one Scottish department. British Journal of Social Work, 23 613-628. Leblanc, M. & Kelloway, K. (2002). Predictors and o utcomes of workplace violence and aggression. Journal of Applied Psychology, 87 444-453. Lee, K., Carswell, J., & Allen, N. (2000). A meta-a nalytic review of occupational commitment: Relations with personand work-relate d variables. Journal of Applied Psychology, 85, 799-811. Lenz, R. (October 18, 2006). Foster baby missing af ter social worker is killed. ABC News. Retrieved on January 8, 2007 from http://abcnews.go.com/US/print?id=2581844 Leong, F., & Wagner, N. (1994). Cross-cultural coun seling supervision: What do we know? What do we need to know? Counselor Education and Supervision, 34 117-132. Lewis, R., & Gingerich, W. (1980). Leadership chara cteristics: Views of Indian and non-Indian students. Social Casework, 61, 494-497. Littlechild, B. (2005). The stressors arising from violence, threats, and aggression in child protection social workers. Journal of Social Work, 5, 61-82. London, M. (1983). Toward a theory of career motiva tion. Academy of Management Review, 8, 620-630. Lyter, S., & Abbott, A. (2007). Home visits in a vi olent world. The Clinical Supervisor, 26, 17-33. Lyons, K., LaValle, I., & Grimwood, C. (1995). Care er patterns of qualified social workers: discussion of a recent survey. British Journal of Social Work, 25, 173190. Macdonald, G., & Sirotich, F. (2001). Reporting cli ent violence. Social Work, 46, 107115.
260 Macdonald, G., & Sirotich, F. (2005). Violence in t he social work workplace: The Canadian experience. International Social Work, 48 772-781. Maidment, J. (2003) Problems experienced by student s on field placement: Using research findings to inform curriculum design and c ontent. Australian Social Work, 56, 50-60. Mama, R. (2001). Violence in the field: Experiences of students and supervisors. Journal of Baccalaureate Social Work, 7, 17-26. Mayer, J. & Rosenblatt, A. (1979). Encounters with danger: Social workers in the ghetto. Sociology of Work and Occupations, 2, 227-245. McRoy, R., Freeman, E., Logan, S., & Blackmon, B. ( 1986). Cross-cultural field Supervision: Implications for social work educatio n. Journal of Social Work Education, 22, 50-56. Mertler, C., & Vanatta, R. (2005). Advanced and mul tivariate statistical methods: Practical application and interpretation. Glendale CA: Pyrczak Publishing. Meyer, J., & Allen, N. (1991). A threecomponent c onceptualization or organizational commitment. Journal of Applied Psychology, 69, 372-378. Meyer, J., Allen, N., & Smith, C. (1993). Commitmen t to organizations and occupations: extension and test of a three-component conceptual ization. Journal of Applied Psychology, 78, 538-551. Meyer, J., Stanley, D., Herscovitch, L., & Topolnyt sky, L. (2003). Affective, continuance, and normative commitment to the organ ization: A meta-analysis of antecedents, correlates, and consequences. Jour nal of Vocational Behavior, 61, 20-52. Mobley, W., Griffeth, H., Hand, H., & Meglino, B. ( 1979). Review and conceptual analysis of the employee turnover process. Psychological Bulletin, 86 493522. Montcalm, D. & Royse, D. (2002). Data analysis for social workers Boston: Allyn and Bacon. National Association of Social Workers and NASW Fou ndation (NASW). (2006a). Assuring the sufficiency of the frontline workforce : A national study of licensed social workers, executive summary NASW Center for Workforce Studies. Retrieved http://workforce.socialworkers.org/studies/nasw_06_ execsummary.pdf
261 NASW Center for Workforce Studies & Center for Heal th Workforce Studies (NASW). (2006b). Licensed social workers in the United Stat es, 2004. NASW & University of Albany, School of Public Health: Wash ington D.C. & Rensselaer, NY. Retrieved from http://workforce.socialworkers.o rg/studies/fullStudy0806.pdf Nakagawa, S., & Foster, M. (2004). The case against retrospective statistical power analysis with an introduction to power analysis. Acta Ethologica, 7 103-108. Newhill, C. (1996). Prevalence and risk factors for client violence toward social workers. Families in Society, 77 488-496. Retrieved January 14, 2005 from Ebscohos t database. Newhill, C. (2003). Client violence in social work practice: Prevention intervention, and research. New York: the Guilford Press. Newhill, C., & Wexler, S. (1997). Client violence t oward children and youth services social workers. Children and Youth Services Review, 19, 195-212. Norris, D., with Kedward, C. (1990). Violence against social workers: the implications for practice. London: Jessica Kingsley Publishers. Nova Southeastern University vs. Bethany Jill Gross (1994). S.C. Case No. 94,079; DCA Case No. 97-01335. Retrieved from www.fsu.ed/library/flsup/sc94079/94079a.pdf Nuehring, E., & Houston, M. (1992). Administrative responses to client violence: A risk management perspective. In Healy, L., Pine, B., & Keys, P. (eds.) Managers choices: Compelling issues in the new decision env ironment. Occupational Safety and Health Administration. (200 4). Preventing workplace violence for health care and social service workers. Retrie ved March 1, 2005 from http:// www.osha-slc.gov/SLTC/workplaceviolence/healthcare OKeefe, M. & Mennan, F. (1998). Integrating conten t on violence into a social work practice curriculum. Journal of Teaching in Social Work, 17 81-100. Pace, P. (2008). Slaying renews safety concerns. NASW News, 53 NASW: Washington D.C. Page, M. (2003). Race, culture, and supervisory rel ationship: A review of the literature and a call to action. Journal of Curriculum and Supervision, 18, 161-174.
262 Privitera, M., Weisman, R., Cerulli, C., Tu, X., & Groman, A. (2005). Violence toward mental health staff and safety in the work environ ment. Occupational Medicine, 55, 480-486. Ramirez, N. (2003). Views toward organizational arr angements for ethnic-sensitive supervision in clinical settings serving Latino pe rsons. Journal of Ethnic and Cultural Diversity in Social Work, 12, 1-18. Rapp-Paglicci, L. (2004). Juvenile probation office rs: Safe and sound? Stress, Trauma, and Crisis, 7, 17-28. Reeser, L., & Wertkin, R. (2001). Safety training i n social work education: A national survey. Journal of Teaching in Social Work, 21 95-113. Rey, L. (1996). What social workers need to know ab out client violence. Families in Society, 77, 33-40. Retrieved January 14, 2005 from Proquest da tabase. Ringstad, R. (2005). Conflict in the workplace: Soc ial workers in the workplace as victims and perpetrators. Social Work, 50, 305-313. Rogers, K., & Kelloway, K. (1997). Violence at work : Personal and organizational outcomes. Journal of Occupational Health Psychology, 2 63-71. Rowett, C. (1986). Violence in social work: a research study of violen ce in the context of local authority social work, Institute of Criminology, University of Cambridge, Occupational Papers 14. Sarkisian, G., & Portwood, S. (2003). Client violen ce against social workers: From increased worker responsibility and administrative mishmash to effective prevention policy. Administration in Social Work, 27 41-59. Scalera, N. (1995). The critical need for specializ ed health and safety measures for child welfare workers. Child Welfare, 74, 337-350. Schat, A,. & Kelloway, K. (2000). Effects of percei ved control on the outcomes of workplace aggression and violence. Journal of Occupational Health Psychology, 5, 386-402. Schat, A., & Kelloway (2003). Reducing the adverse consequences of workplace aggression and violence: the buffering effects of organizational support. Journal of Occupational Health Psychology, 8, 110-122. Schultz, L. (1987). The social worker as a victim o f violence. Social Casework: The Journal of Contemporary Social Work, 68 240-244.
263 Schultz, L. (1989). The victimization of social wor kers. Journal of Independent Social Work 3, 51, 51-63. Schwarzmeueller, M. (1998). MSW students perceptio n of personal safety in field placement. (Masters thesis, California State Univ ersity, Long Beach). Retrieved from Proquest Dissertations .UMI Microform 1392068 Sedensky, M. (2004). Job protection sought for soci al workers. WTOP News. Associated Press. Kansas City, MO. www.terizenner.com/all_articles/2.html Seeck, S. (1998). Violence in the workplace: A stud y of violence by clients directed toward psychologists and social workers in Los Ang eles. (Masters thesis, California State University, Long Beach) Retrieved from Proquest Dissertations. UMI Microform1391678. Smith, M., & Nursten, J. (1998). Social workers ex perience of distress: Moving towards change? British Journal of Social Work, 28 351-368. Snow, K. (1994). Aggression: Part of the job? the p sychological impact of aggression on child and youth workers. Journal of Child and Youth Care, 9, 11-29. Song, K. (2005). Prevalence of client violence towa rd child and family social workers and its effects on burnout, organizational commitm ent, and turnover intention: A structural equation modeling approach. (Doctoral d issertation, Columbia University, Retrieved from Proquest Dissertations, UMI # 3188797. Spencer, P., & Munch, S. (2003). Client violence to ward social workers: the role of management in community mental health programs. Social Work, 48 532-545. Retrieved January 25, 2005 from Ebscohost database. Star, B. (1984). Patient violence/ Therapist safety Social Work, 29 225-230. Sullivan, C., & Yuan, C. (1995). Workplace assaults on minority health and mental health care workers in Los Angeles. American Journal of Public Health, 85, 1001-1014. Super, D. (1957). The psychology of careers Harper and Brothers: New York City Tabachnik,B., & Fidell, L. (2007). Using multivariate statistics. Boston: Pearson/Allyn and Bacon.
264 Tully, C., Kropf, N., and Price, J. (1993). Is fiel d a hard hat area? A study of violence in field placements. Journal of Social Work Education, 29, 191-200. Retrieved January 14, 2005 from Ebscohost database. Vandenburg, R., & Scarpello, V. (1994). A longitudi nal measure of the determinant relationship between employee commitments to the o ccupation and the organization. Journal of Organizational Behavior, 15 535-547. Vander Kolk (1974). Relationship of personality, va lues, and race to anticipation of the supervisory relationship, Rehabilitation Counseling Bulletin, 18 41-46. Vanmaanen, J., & Schein, E. (1977). Career develop ment. In R. Hackinson & L. Suttle (Eds.) Improving life at work (pp. 30-94) Santa Monica, CA: Goodyear. Vergara, E. (2006). Social workers perceptions of personal safety and level of satisfaction with safety policies and procedures. (Masters thesis: California State University, Long Beach) Retrieved from Proqu est Dissertations, UMI Number: 1434602. University of Kansas. (2004). University community mourns murder of graduate student In Overland Park. Retrieved from www.terizenner.com/all_articles/4.html Waters, J., Lynn, R., & Morgan, K. (2002). Workpla ce violence. In L. Rapp-Paglicci, A. Roberts, & J. Wodarski (Eds.). The violence handbook pp. 378-413. New York: John Wiley. Wayne, J., & Raskin, M. (1996). Attitudes and pract ices of field education directors toward student safety concerns. Arete, 21 1-11. Whitman, R., Armao, B., & Dent, O. (1976). Assaults on therapists. American Journal of Psychiatry, 133 426-431. Wrenn, L. (2005). The relationship between personal trauma exposure and secondary trauma exposure and secondary traumatic stress for social workers. (Doctoral Dissertation, Loyola University, Chicago). Retriev ed from Proquest dissertations, UMI Number: 3174273. Yeomans, D. (2004). Negligence liability issues for Floridas service-learning programs. Florida Campus Compact. Retrieved from www.floridacompact.org/resources/NegligenceLiabilit y.pdf Zakutansky, T. & Sirles, E. (1993). Ethical and leg al issues in field education: Shared responsibility and risk. Journal of Social Work Education, 29 338-346.
266 Experience of Client Violence: Direct/Indirect Physical Threat of physical harm Verbal abuse Threat of lawsuit Property damage Fear of Future Victimization (Mediator) Career Withdrawal Intentions Personal Demographics: Age Experience Gender Ethnicity Education Occupational Commitment: Affective Normative Continuance Organizational Demographics: Place of violence Time of violence Training Practice Setting Field educator/ student race/ethnicity match Appendix A: Tables and Figures Figure 1 Research Model
267 Appendix A (Continued) Table 1 Summary of Hypotheses, Variable, Instruments, and S tatistical Methods No. Hypothesis/Questions Dependent Variable Independent Variable Statistical Test 1-1. Male social work students will experience more client violence than female social work students Client violence Gender t-test for independent samples 1-2. Younger social work students will experience more client violence than older social work students Client violence Age Age categories Pearsons r One way ANOVA 1-3. Less experienced social work students will experience more client violence than more experienced social work students Client violence Years of Experience Experience categories Pearsons r One way ANOVA 1-4. There will be no difference in exposure to client violence by students of various ethnicities/racial backgrounds. Client violence Ethnicity/racial background One way ANOVA 1-5. BSW social work students will experience more client violence than MSW students Client violence Education level ttest for independent samples 1-6. There will be a significant difference in numbers of client violence incidents according to the place of social work practice (office, home visit, other). Client violence Location of violence One way ANOVA 1-7. There will be a significant difference in the numbers of client violence incidents according to the time of day. Client violence Time of client violence One way ANOVA 1-8. There will be a significant difference in the numbers of client violence incidents according to the practice setting. Client violence Social Work Practice Settings One way ANOVA
268 Appendix A (Continued) Table 1 (Continued) Summary of Hypotheses, Variable, Instruments, and S tatistical Methods No. Hypothesis/Question Dependent Variable Independent Variable Statistical Test 1-9. There will be a significant difference in the numbers of client violence incidents according to the race/ethnicity match of the students and their field educators. 1-10. Social work students who have more safety training offered by their social work program will experience less of every type of client violence than social work students who have had less safety training in their social work program. Client violence Client violence Ethnic match of field educator and student Total training received One way ANOVA Pearsons r Best predictors of client violence among social work students Client violence Gender Age Experience Practice setting Location of Violence Time of violence Ethnicity/racial background BSW/MSW Total training content areas Site of training Multiple regression 2-1. Experience of client violence has a positive and direct effect on fear of future violence Fear of Future Violence Client violence For each set of variables: Pearsons r; then Simple linear regression; Then hierarchical multiple regression 2-2. Fear of future violence has a negative and direct effect on occupational commitment. Occupational commitment Fear of future violence 2-3. Fear of future client violence has a negative and direct effect on career withdrawal intention. Career withdrawal intention Fear of future violence
269 Appendix A (Continued) Table 2 Personal Demographic Characteristics of Respondent s Characteristics BSW MSW Combined MSW/ BSW % n % n % n Gender ( N = 586) Female 91.9 227 86.4 293 88.5 525 Male 8.1 20 13.6 46 11.4 68 Missing .3 2 Total 100 247 100 339 100.0 595 Age (N = 558) 24-under 44.5 102 13.0 42 24.4 145 25-30 17.0 39 38.2 123 27.4 163 31-40 18.3 42 21.4 69 18.8 112 41-50 12.7 29 17.4 56 14.5 86 51-up 7.4 17 9.9 32 8.7 52 Missing 6.3 37 Total 100.0 229 100.0 322 100.0 595 Experience (N = 566) None 77.7 181 33.6 112 49.4 294 1-2 years 9.9 23 20.1 67 15.1 90 3-5 years 7.7 18 18.6 62 13.8 82 6-10 years 2.1 5 18.3 61 11.1 66 11-20 years 2.1 5 7.8 26 5.4 32 21-up years .4 1 1.5 5 1.0 6 Missing 4.2 25 Total 100.0 233 100 333 100.0 595 Note: Total sample is reported in percentages that reflect the percentage of missing cases. BSW/MSW c olumns reflect valid percentages.
270 Appendix A (Continued) Table 2 (Continued) Personal Demographic Characteristics of Respondent s Characteristics BSW MSW Combined MSW/ BSW % n % n % n Ethnicity ( N = 581) Latino/Hispanic 4.1 10 4.5 15 4.3 25 Native American 7.9 7 1.8 6 2.2 13 White 71.4 175 73.8 248 71.4 425 Asian/Pacific Island 0 0 1.5 5 .8 5 Black 18.4 45 14.6 49 16.6 99 Mixed Race 3.3 8 .3 1 3.6 21 Missing 1.2 7 Total 100.0 245 100.0 336 100.0 595 Degree Sought ( N = 588) MSW/BSW 42.2 248 57.1 340 98.8 588 Missing 1.2 7 Total 42.2 248 57.8 340 100.0 595 Note: Total sample is reported in percentages that reflect the percentage of missing cases. BSW/MSW c olumns reflect valid percentages.
271 Appendix A (Continued) Table 3 Organizational/ Professional Demographics Characteristics BSW MSW Combined MSW/ BSW % n % n % n Number of home visits (N = 585) 0 visits 42.3 105 49.6 167 46.4 274 1-5 visits 14.9 37 13.6 46 14.2 84 6-10visits 11.3 28 8.0 27 9.3 55 11 or more visits 31.5 78 28.8 97 30.0 177 Missing .8 5 Total 100.0 248 100.0 337 100.0 595 % of evening hours worked (N = 587) 0 evening hours 36.8 91 31.2 106 33.3 198 1-25% evening hours 42.1 104 43.5 148 42.7 254 26-50 evening hours 11.7 29 10.9 37 11.1 66 51-75% evening hours 6.9 17 7.4 25 7.2 43 76-100% evening hours 2.4 6 7.1 24 5.0 30 Missing .7 4 Total 100.0 247 100.0 340 100.0 595 Ethnic mix of supervision dyad (N = 501) Matched ethnicity 63.8 136 70.5 203 57.5 342 Field educator White/Student non-White 18.3 39 12.5 36 12.6 75 Field educator non-White/Student White 15.0 32 14.6 42 12.9 77 Field educator/student mixed non-White 2.8 6 2.4 7 2.2 13 Missing 14.8 88 Total 100.0 213 100.0 288 100.0 595
272 Appendix A (Continued) Table 3(Continued) Organizational/ Professional Demographics Characteristics BSW MSW Combined MSW/ BSW % n % n % n Practice Setting (N = 580) Medical/health care 8.2 20 14.3 48 11.9 71 Alcohol/drug/substance abuse 5.7 14 8.0 27 7.1 42 Developmental disabilities 3.3 8 2.1 7 2.5 15 Corrections/criminal justice 8.2 20 5.4 18 6.4 38 Children & youth/Child protection 23.0 56 14.0 47 17.5 104 Community organization/planning 5.7 14 7.1 24 6.6 39 Family services 11.5 28 6.8 23 8.7 52 Group services .4 1 .9 3 .7 4 School 7.8 19 8.3 28 7.9 47 Service to the aged 8.2 20 6.2 21 6.9 41 Occupational/vocational 1.6 4 .6 2 1.0 6 Psychiatric/mental health 10.7 26 22.3 75 17.0 101 Public assistance/welfare 0 0 .6 2 .3 2 Other 7.8 19 5.4 18 4.4 25 Total 100.0 244 100.0 336 100.0 595 Note: Total sample is reported in percentages that reflect the percentage of missing cases. BSW/MSW c olumns reflect valid percentages.
273 Appendix A (Continued) Table 4 Total Training in the Social Work Program per Conte nt Area Training Content Received training in s.w. class, field seminar and field agency Received training in field seminar and field agency Received training in social work practice class and field agency Assessing history of violence in clients 6.6% ( n = 34) 6.5% ( n = 37) 15.0% ( n =86) Characteristics/life experiences of people more likely to commit violent acts 4.5% ( n = 26) 5.1% ( n = 29) 11.9% ( n = 68) Forms of mental illness associated with violent behavior 5.9% ( n = 34) 6.3% ( n = 36) 18.2% ( n = 104) Characteristics of high risk situations (i.e. non-public, isolated places) 10.7% ( n = 61) 13.1% ( n = 75) 19.9% ( n = 114) Identifying and managing feelings that can arise when working with victims and perpetrators of violence 10.0% ( n = 63) 11.7% ( n = 67) 20.5% ( n = 117) Maintaining a confident, secure demeanor 17.5% ( n = 100) 18.7% ( n = 107) 24.7% ( n = 141) Recognizing verbal acts of violence 8.7% ( n = 50) 9.8% ( n = 56) 16.6% ( n = 95) Physical signs that an attack is imminent 5.2% ( n = 30) 6.1% ( n = 35) 10.1% ( n = 58) Understanding of students rights (For example, the right to refuse to make a home visit) 10.7% ( n = 61) 13.8% ( n = 79) 16.3% ( n = 93) Verbal de-escalation techniques ( How to behave with an angry client) 11.5% ( n = 66) 12.4% ( n = 71) 20.3% ( n = 116) Knowledge of office safety (arranging work space to maximize safety) 9.8% ( n = 56) 11.7% ( n = 67) 16.1% ( n = 92) Where to sit when interacting with a client 11.2% ( n =64) 13.5% ( n = 77) 21.5% ( n = 123) Home visit safety 10.1% ( n = 58) 13.6% ( n = 78) 15.9% ( n = 91) Keeping supervisor informed of ones itinerary 16.4% ( n = 94) 21.5% ( n = 128) 23.8% ( n = 136) Physical techniques for self protection 3.5% ( n = 20) 4.0% ( n = 23) 6.1% ( n = 35) Debriefing and support after an incident (Reporting the incident) 11.9% ( n = 68) 15.9% ( n = 91) 19.6% ( n = 112) Recording incidents of violence 9.6% ( n = 55) 11.5% ( n = 66) 17.3 % ( n = 99) Note: Missing =23 in every cell
274 Appendix A (Continued) Table 5 Percentages of Students Who Had Safety Training Con tent per Training Venue Training Content Social Work Practice Classes % ( N= 572) Social Work Field Seminar % ( N= 572) Field Agency % ( N= 572) Other % ( N= 572) Assessing history of violence in clients 33.9 ( n= 202) 8.6 ( n= 51) 27.1 ( n= 161) 14.8 ( n= 88) Characteristics/life experiences of people more likely to commit violent acts 42.2 ( n= 251) 8.9 ( n= 53) 20.2 ( n= 120) 16.3 ( n= 97) Forms of mental illness associated with violent behavior 55.5 ( n= 330) 11.4 ( n= 68) 24.5 ( n= 146) 23.0 ( n= 137) Characteristics of high risk situations (i.e. non-public, isolated places) 48.1 ( n= 286) 24.4 ( n= 145) 31.1 ( n= 185) 23.9 ( n= 142) Identifying and managing feelings that can arise when working with victims and perpetrators of violence 51.1 ( n= 301) 20.8 ( n= 124) 28.6 ( n= 170) 18.7 ( n= 111) Maintaining a confident, secure demeanor 52.8 ( n= 214) 26.7 ( n= 159) 33.3 ( n= 198) 22.5 ( n= 134) Recognizing verbal acts of violence 41.8 ( n= 249) 16.1 ( n= 96) 27.2 ( n= 162) 23.5 ( n= 140) Physical signs that an attack is imminent 24.9 ( n= 148) 12.1 ( n= 72) 23.0 ( n= 137) 21.8 ( n= 130) Understanding of students rights (For example, the right to refuse to make a home visit) 38.0 ( n= 226) 33.6 ( n= 200) 22.9 ( n= 136) 8.2 ( n= 49) Verbal de-escalation techniques( How to behave with an angry client) 44.0 ( n= 262) 19.5 ( n= 116) 31.3 ( n= 186) 25.2 ( n= 150) Knowledge of office safety (arranging work space to maximize safety) 38.8 ( n= 231) 20.3 ( n= 121) 32.6 ( n= 194) 19.0 ( n= 113) Where to sit when interacting with a client 58.0 ( n= 345) 26.6 ( n= 158) 32.6 ( n= 194) 19.0 ( n= 113) Home visit safety 35.5 ( n= 211) 25.2 ( n= 150) 29.9 ( n= 178) 17.3 ( n= 103) Keeping supervisor informed of ones itinerary 35.6 ( n= 212) 30.1 ( n= 179) 49.1 ( n= 292) 17.6 ( n= 105) Physical techniques for self protection 14.5 ( n= 86) 8.7 ( n= 52) 17.5 ( n= 104) 25.9 ( n= 154) Debriefing and support after an incident (Reporting the incident) 31.3 ( n= 186) 19.8 ( n= 118) 38.7 ( n= 230) 19.8 ( n= 118) Recording incidents of violence 27.6 ( n= 164) 15.3 ( n= 91) 37.8 ( n= 225) 19.3 ( n= 115) Knowledge of social work programs safety policies 40.5 ( n= 241) 30.9 ( n= 184) 25.0 ( n= 149) 10.4 ( n= 62) Knowledge of field agencys safety policies 16.5 ( n= 98) 15.6 ( n= 93) 67.3 ( n= 385) 6.5 ( n= 6.2) Note: Missing = 23 in every cell
275 Appendix A (Continued) Figure 2 0 10 20 30 40 50 60 70 80 90 100 Training Content Areas Percent of Social Work Students Receiving Client Violence Training Content Areas by Venue Social Work Classes Field Seminar
276 Appendix A (Continued) Table 6 Percentages of Students Who Did Not Have Safety Tra ining Content per Training Venue Training Content Social Work Practice Classes % ( N= 572) Social Work Field Seminar % ( N= 572) Field Agency % ( N= 572) Other % ( N= 572) Assessing history of violence in clients 62.2 ( n= 370) 87.6 ( n= 521) 69.1 ( n= 411) 81.3 ( n= 484) Characteristics/life experiences of people more likely to commit violent acts 53.9 ( n= 321) 87.2 ( n= 519) 76.0 ( n= 452) 79.8 ( n= 475) Forms of mental illness associated with violent behavior 40.7 ( n= 242) 84.7 ( n= 504) 71.2 ( n= 426) 73.1 ( n= 435) Characteristics of high risk situations (i.e. non-public, isolated places) 48.1 ( n= 286) 71.8 ( n= 427) 65.0 ( n= 387) 72.3 ( n= 430) Identifying and managing feelings that can arise when working with victims and perpetrators of violence 46.9 ( n= 268) 75.3 ( n= 448) 67.6 ( n= 402) 77.5 ( n= 461) Maintaining a confident, secure demeanor 43.4 ( n= 258) 69.4 ( n= 413) 62.9 ( n= 374) 73.6 ( n= 438) Recognizing verbal acts of violence 54.3 ( n= 323) 80.0 ( n= 476) 68.9 ( n= 410) 72.6 ( n= 432) Physical signs that an attack is imminent 71.3 ( n= 424) 84.0 ( n= 500) 73.1 ( n= 435) 74.3 ( n= 442) Understanding of students rights (For example, the right to refuse to make a home visit) 58.2 ( n= 346) 62.5 ( n= 372) 73.3 ( n= 436) 87.9 ( n= 523) Verbal de-escalation techniques ( How to behave with an angry client) 52.1 ( n= 310) 76.6 ( n= 456) 64.9 ( n= 386) 70.9 ( n= 422) Knowledge of office safety (arranging work space to maximize safety) 57.3 ( n= 341) 75.8 ( n= 451) 63.5 ( n= 378) 77.1 ( n= 459) Where to sit when interacting with a client 38.2 ( n= 227) 69.6 ( n= 414) 63.5 ( n= 378) 77.1 ( n= 459) Home visit safety 60.7 ( n= 361) 70.9 ( n= 422) 66.2 ( n= 394) 78.8 ( n= 469) Keeping supervisor informed of ones itinerary 60.5 ( n= 360) 66.1 ( n= 393) 47.1 ( n= 280) 78.5 ( n= 467) Physical techniques for self protection 81.7 ( n= 486) 87.4 ( n= 520) 78.7 ( n= 468) 70.3 ( n= 418) Debriefing and support after an incident (Reporting the incident) 64.9 ( n= 386) 76.3 ( n= 454) 57.5 ( n= 342) 76.3 ( n= 454) Recording incidents of violence 68.6 ( n= 408) 80.3 ( n= 481) 58.3 ( n= 347) 76.8 ( n= 457) Knowledge of social work programs safety policies 55.6 ( n= 331) 65.2 ( n= 388) 71.1 ( n= 423) 85.7 ( n= 510) Knowledge of field agencys safety policies 79.3 ( n= 472) 80.3 ( n= 478) 31.4 ( n= 187) 89.9 ( n= 535) Note: Missing = 23 in every cell
277 Appendix A (Continued) Table 7 Prevalence Rate of Direct Client Violence during P racticum ( N = 589) n % Experienced direct violence 248 41.7 No experience of violence 347 57.8 Missing data 3 .5
278 Appendix A (Continued) Table 8 Rate of Direct Violence by Type of Violence ( N = 589) Type of Direct Violence n % Verbal abuse 223 37.5 Threat of physical harm 84 14.1 Threat of lawsuit 56 9.4 Damage to personal or professional property 43 7.2 Physical assault 21 3.5 Missing 6 1.0
279 Appendix A (Continued) Table 9 Total Incidents of Direct Client Violence during P racticum ( N = 592) Total Incidents M SD Range Skew Kurtosis 1591 2.69 21.17 0 506 66.04 543.47 1133 (After transforming the extreme cases in all direct violence variables) 1.86 4.66 0 62 6.78 22.87
280 Appendix A (Continued) Table 10 Quantities of Incidents of Direct Client Violence ( N = 592) Type of Client Violence of incidents % of total sum M SD Range Verbal abuse 1208 75.92 2.04 20.75 0-500 Threat of physical harm 167 10.50 .28 .90 0-10 Damage to personal or professional property 87 5.47 .15 .72 0-10 Threat of lawsuit 83 5.22 .14 .55 0-7 Physical assault 46 2.89 .08 .86 0-20 Total 1591 100.01 After reducing cases > 4 SD Verbal abuse 759 68.75 1.28 3.81 0-51 Threat of physical harm 161 14.58 .27 .82 0-5 Threat of lawsuit 79 7.15 .13 .49 0-4 Damage to personal or professional property 77 6.97 .13 .56 0-4 Physical assault 28 2.53 .05 .28 0-3 Total 1104 99.98
281 Appendix A (Continued) Figure 3 Types of Direct Client Violence Incidents against S ocial Work Students
282 Appendix A (Continued) Table 11 Prevalence of Indirect Exposure to Client Violence during Practicum ( N = 587) Exposed to Indirect Violence? n % Exposed indirectly Violence 361 60.2 No indirect exposure to violence 226 38.5 Missing 8 1.3
283 Appendix A (Continued) Table 12 Prevalence of Indirect Exposure to Client Violence by Type of Client Violence ( N = 587) Type of Client Violence n % Verbal abuse 322 54.1 Threat of physical harm 219 36.8 Physical assault 139 23.4 Threat of lawsuit 139 23.4 Damage to personal or professional property 122 20. 5
284 Appendix A (Continued) Table 13 Total Incidents of Indirect Exposure to Client Vio lence Total Incidents of Indirect Exposure ( ) M SD Range Skew Kurtosis 4110 7.05 19.97 0-151 9.42 128.49 3603 ( After reducing extreme cases) 6.18 12.64 0-9 7 4.02 19.38
285 Appendix A (Continued) Table 14 Incidents of Indirect Exposure to Client Violence b y Type of Violence No cases reduced ( N = 587) Type of Violence % of total sum M SD Range Verbal abuse 2060 50.63 3.56 11.15 0-200 Threat of physical harm 857 20.85 1.46 5.31 0-100 Physical assault 425 10.34 .72 2.21 0-24 Damage to personal or professional property 399 9.70 .68 2.75 0-40 Threat of lawsuit 369 8.98 .63 2.07 0-20 Total 4110 100.5 After reducing cases over 4 SD Verbal abuse 1852 51.40 3.17 6.82 0-46 Threat of physical harm 751 20.70 1.27 3.17 0-21 Physical assault 388 10.46 .66 1.74 0-11 Damage to personal or professional property 324 8.91 .55 1.61 0-11 Threat of lawsuit 309 8.52 .52 1.34 0-8 Total 3603 99.9
286 Appendix A (Continued) Figure 4 Types of Indirect Exposure to Client Violence by So cial Work Students
287 Appendix A (Continued) Table 15 Descriptive Statistics of Fear of Future Violence Variables Mean SD Range Skew Kurtosis Fear of future physical assault 2.19 1.20 1-5 .625 -.751 Fear of future threat of harm 2.52 1.33 1-5 .273 -1.24 Fear of future verbal abuse 3.13 1.37 1-5 -.316 -1.174 Fear of Future threat of lawsuit 2.45 1.28 1-5 .299 -1.106 Fear of future property damage 2.28 1.17 1-5 .496 -.755 General fear of future client violence 2.77 1.38 15 .052 -1.304
288 Appendix A (Continued) Table 16 Results of Exploratory Factor Analysis of Occupatio nal Commitment (Revised per Initial Factor Analysis ) Occupational Commitment Items Factor 1 Factor 2 Factor 3 Communality Reliability Continuance Commitment .81 CC1 .60 .55 CC 2 .68 .55 CC 3 .81 .70 CC 4 .85 .63 CC 6 .78 .57 Affective Commitment .91 AC 2 -.85 .70 AC 3 -.87 .74 AC 4 -.87 .72 AC 5 -.86 .74 AC 6 -.85 .72 Normative Commitment .82 NC 2 -.87 .69 NC 3 -.81 .65 NC4 -.76 .56 NC 5 -.72 .51 Eigenvalue 4.47 3.19 1.69 Variance 31.96 22.76 12.06 Cumulative Variance 31.96 54.71 66.77 Note: AC 2=I am satisfied with my choice to enter t he profession of social work; AC3=I am proud to be in the social work profession; AC4= I like being a social worker; AC5=I identify with the social work profession; AC 6=I am enthusiastic about social work; NC 2=I feel obligat ed to remain in the social work profession; NC3=I f eel a responsibility to the social work profession to con tinue; NC4=I would feel guilty if I left the socia l work profession; NC 5=Even it if were to my advantage, I do not feel it would be right to leave social work now; CC1=I have put too much into the social work profession to consider ch anging now; CC2=Changing professions would be diffi cult for me to do now; CC3=Too much of my life would be disrupt ed if I were to change my profession now; CC4=It wo uld be costly for me to change my profession now; CC6=Chan ging professions now would require personal sacrifi ce
289 Appendix A (Continued) Table 17 Descriptive Statistics of Occupational Commitment ( After factor analysis indicated adjustments) Occupational Commitment Variables Mean SD Range Skew Kurtosis Affective Commitment 23.35 2.93 5-25 -2.54 8.17 Normative Commitment 11.68 4.18 4-20 -.026 -.649 Continuance Commitment 16.30 5.10 5-25 -.324 -.567
290 Appendix A (Continued) Table 18 Descriptive Statistics of Career Withdrawal Intenti ons Career Withdrawal Intention Variables Mean SD Range Skew Kurtosis Thinking about leaving sw* profession 1.47 .94 1-5 2.06 3.33 Intend to look for new profession 1.46 .88 1-5 2.00 3.41 Intend to stay in sw* profession for some time (R) 1.47 .87 1-5 2.19 4.64 Total Career Withdrawal Intentions 4.39 2.49 3-15 1 .98 3.22 Note: sw = social work
291 Appendix A (Continued) Table 19 Results of Chi Square Tests for Practicum Exposure Rate to Client Violence by Gender ( N = 590) Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence Gender % n % n % n % n % n % n Female 3.1 16 11.9 62 35.5 184 8.7 45 6.4 33 39.3 205 Male 7.4 5 32.4 22 55.9 38 16.2 11 14.7 10 61.6 42 X 2 = 3.18 X 2 = 20.41 X 2 = 10.67 X 2 = 3.93 X 2 = 6.17 X 2 = 12.50 p = .075 p = .000*** p = .001*** p = .048* p = .013** p = .000*** Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
292 Appendix A (Continued) Figure 5
293 Appendix A (Continued) Table 20 Results of t -tests for Incidents of Client Violence by Gender Types of Client Violence Gender Mean t -value p Physical assault Female .04 -1.36 .179 Male .12 Threat of physical harm Female .21 -3.24 .002*** Male .69 Verbal abuse Female 1.05 -2.20 .031* Male 3.09 Threat of lawsuit Female .13 -1.01 .314 Male .19 Property damage Female .12 -1.21 .230 Male .22 Total client violence Female 1.59 -2.45 .017* Male 4.32 Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
294 Appendix A (Continued) Table 21 Results of Chi Square Tests for Practicum Exposur e Rate to Client Violence per Age Category ( N = 592) Age Categories Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence % n % n % n % n % n % n Under 25 4.8 7 15.2 22 36.6 53 6.2 9 4.8 7 37.9 55 Age 25-30 2.5 4 16.7 27 42.0 68 14.8 24 7.4 12 45.7 74 Over 30 3.5 10 12.4 35 36.2 102 8.2 23 8.5 24 41.8 119 X 2 = 1.24 X 2 = 1.65 X 2 = 1.614 X 2 = 7.74 X 2 = 1.92 X 2 = 1.89 p = .538 p = .437 p = .446 p = .021* p = .382 p = .388 Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
295 Appendix A (Continued) Figure 6
296 Appendix A (Continued) Table 22 Results of ANOVAs with Means and Standard Deviation s of Client Violence among Age Categories ( N = 592) Age Categories Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence M SD M SD M SD M SD M SD M SD Under age 25 ( n = 145 ) .69 .35 .23 .65 .84 1.84 .08 .31 .06 .29 1.27 2.60 Age 25-30 ( n = 162) .03 .26 .39 1.05 1.57 4.53 .22 .64 .13 .52 2.35 5.5 5 Over age 30 ( n = 285) .05 .25 .23 .75 1.34 4.11 .11 .45 .16 .67 1.89 4.90 F = .710 F = 2.26 F = 1.48 F = 4.01 F = 1.61 F = 2.03 p = .491 p = .106 p = .229 p = .019* p = .201 p = .132 Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
297 Appendix A (Continued) Table 23 Results of Chi Square Tests for Practicum Exposure Rate to Client Violence per Experience Level ( N = 557) ExperienceCategories Physical Assault Threatened Physical Harm Verbal Abuse Threatened Lawsuit Property Damage Total Client Violence % n % n % n % n % n % n 0 years 2.7 8 10.3 30 36.1 105 7.9 23 6.2 18 40.4 118 1-2 years 4.5 4 19.1 17 24.8 31 10.1 9 9.0 8 40.4 36 3-5 years 4.9 4 18.5 15 37.0 30 7.4 6 11.1 9 41.5 34 6-10years 3.0 2 13,6 9 40.9 27 12.1 8 4.5 3 43.9 29 X 2 = 1.31 X 2 = 6.69 X 2 = .69 X 2 = 1.59 X 2 = 3.42 X 2 = .30 p = .728 p = .083 p = .876 p = .661 p = .331 p = .961
298 Appendix A (Continued) Table 24 Results of Pearsons Correlations for Incidents of Client Violence by Years of Experience Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence r p r p r p r p r p r p Years of experience .01 .796 .06 .177 .09 .040* .08 .068 .00 .915 .09 033* Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
299 Appendix A (Continued) Figure 7
300 Appendix A (Continued) Table 25 Results of ANOVAs with Means and Standard Deviation s of Client Violence among Experience Categories ( N = 529) Experience Categories Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence M SD M SD M SD M SD M SD M SD No experience ( n = 292) .03 .22 .19 .71 .88 1.87 .09 .30 .10 .47 1.29 2.66 1-2 years experience ( n = 89) .04 .21 .33 .94 1.27 3.19 .17 .63 .18 .70 2.00 3.97 3-5 years experience ( n = 82) .10 .46 .37 .94 2.07 6.21 .15 .61 .17 .58 2.85 7.24 6-10 years Experience ( n = 66) .03 .17 .37 1.06 1.21 2.26 .20 .59 .09 .52 1.86 3.5 8 F = 1.38 F = 1.65 F = 3.01 F = 1.53 F = .75 F = 3.45 p = .247 p = .177 p = .030* p = .205 p = .520 p = .016* Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
301 Appendix A (Continued) Table 26 Results of Chi Square Tests for Practicum Exposure Rate to Client Violence by Race/Ethnicity ( N = 585) Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence Race/ Ethnicity % n % n % n % n % n % n Latino/ Hispanic 0 0 12.5 3 41.7 10 12.5 3 4.2 1 36.0 9 Native American 7.7 1 30.8 4 61.5 8 7.7 1 15.9 2 53.8 7 White 3.5 15 13.9 59 36.6 155 9.0 38 8.0 34 40.9 174 Asian 0 0 20.0 1 40.0 2 20.0 1 20.0 1 60.0 3 Black 3.1 3 12.2 12 35.7 35 7.1 7 2.0 2 40.8 40 Mixed Heritage 10.5 2 26.3 5 52.6 10 31.6 6 15.8 3 63.0 12 X 2 = 4.41 X 2 = 5.64 X 2 = 5.48 X 2 = 12.37 X 2 = 9.06 X 2 = 3.86 p = .492 p = .343 p = .360 p = .031* p = .107 p = .355 Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
302 Appendix A (Continued) Figure 8
303 Appendix A (Continued) Table 27 Results of ANOVAs with Means and Standard Deviation s of Client Violence among Race/Ethnic Groups ( N = 585) Practice Setting Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence M SD M SD M SD M SD M SD M SD Latino/ Hispanic ( n = 25) .00 .00 .20 .65 .80 1.32 .20 .58 .04 .20 1.24 2.06 Native American ( n = 13) .15 .55 .46 .77 6.61 14.80 .08 .28 .38 1.12 7.92 18 .01 White ( n = 425) .04 .28 .27 .83 1.15 3.30 .12 .46 .13 .54 1.80 4.42 Asian ( n = 5) .00 .00 .60 1.34 1.20 2.16 .20 .48 .20 .45 3.00 3.7 4 Black ( n = 98) .04 .24 .25 .86 1.25 2.80 .09 .35 .04 .31 1.70 3.43 Mixed Heritage ( n = 19) .16 .50 .37 .68 1.57 2.46 .58 1.12 .47 1.26 2.52 3. 30 F = 1.14 F = .38 F = 5.43 F = 3.54 F = 2.59 F = 4.26 p = .339 p = .858 p = .000*** p = .004*** p = .025* p = .001*** Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
304 Appendix A (Continued) Table 28 Results of Chi Square Tests for Practicum Exposure Rate to Client Violence by Degree Program ( N = 585) Degree Program Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence % n % n % n % n % n % n MSW 3.0 10 16.4 55 39.4 132 11.6 39 6.6 22 24.6 144 BSW 4.5 11 11.3 28 35.6 88 6.5 16 8.5 21 17.4 102 X 2 = .88 X 2 = 3.00 X 2 = .86 X 2 = 4.43 X 2 = .78 X 2 = .10 p = .348 p = .083 p = .353 p = .035* p = .378 p = .752 Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
305 Appendix A (Continued) Figure 9
306 Appendix A (Continued) Table 29 Results of t -tests for Incidents of Client Violence by Degree P rogram Types of Client Violence Gender Mean t -value p Physical assault MSW .03 1.42 .156 BSW .07 Threat of physical harm MSW .35 -2.77 .004*** BSW .17 Verbal abuse MSW 1.52 -1.94 .053* BSW .96 Threat of lawsuit MSW .16 -1.77 .077 BSW .09 Property damage MSW .13 .22 .826 BSW .14 Total client violence MSW 2.24 -1.96 .031* BSW 1.50 Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
307 Appendix A (Continued) Table 30 Practicum Exposure Rate to Client Violence per Plac e of Violence ( N = 274) Place of Violence Physical Assault Threatened Physical Harm Verbal Abuse Threatened Lawsuit Property Damage Total Client Violence % n % n % n % n % n % n Home visit 11.8 2 24.7 20 26.9 57 15.7 8 8.1 3 23.4 148 Office 41.2 7 54.3 44 51.4 109 58.8 30 70.3 26 54.4 64 Other 47.1 8 21.0 17 21.7 46 25.5 13 21.6 8 22.6 62 Note: SPSS could not calculate 2 due to the construction of the variables (See p.14 8)
308 Appendix A (Continued) Table 31 Results of ANOVAs with Means and Standard Deviation s of Client Violence of Students Making Home Visits per Number of Home Visits Made During Practi cum ( N = 316) Number of Home Visits Made Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence M SD M SD M SD M SD M SD M SD 1-5 home visits ( n = 84) .02 .15 .15 .53 .85 1.73 .06 .24 .13 .65 1.21 2.27 6-10 home visits ( n = 55) .07 .33 .16 .60 .73 1.46 .07 .36 .07 .42 1.10 2.31 11 or more home visits ( n = 177) .08 .38 .40 1.09 1.35 2.94 .25 .70 .16 .66 2.24 4.1 7 F = .86 F = 2.71 F = 2.00 F = 4.57 F = .46 F = 3.71 p = .422 p = .068 p = .137 p = .011** p = .632 p = .027* Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
309 Appendix A (Continued) Table 32 Practicum Exposure Rate to Client Violence per Time of Day Time of Violence Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence % n % n % n % n % n % n Daytime 76.5 13 71.8 56 71.6 151 82.0 42 71.4 25 69.69 184 Evening 23.5 5 24.4 19 26.5 56 13.7 7 25.7 9 26.13 69 Early morning 0 0 3.8 3 1.9 4 3.9 2 2.9 1 4.16 11 Note: SPSS could not calculate 2 due to the construction of the variables (See p. 1 51)
310 Appendix A (Continued) Table 33 Results of Chi Square Tests for Practicum Exposure Rates per Percent of Evening Hours Worked ( N = 588) Percent of Evening Hours Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence % n % n % n % n % n % n 0% 3.1 6 12.2 24 33.7 66 7.7 15 5.6 11 38.1 75 1-25% 4.0 10 13.5 34 39.7 100 11.1 28 8.7 22 44.3 29 26-50% 1.5 1 20.0 13 40.0 26 7.7 5 6.2 4 44.6 29 51-75% 9.3 4 18.6 8 41.9 18 9.3 4 9.3 4 46.5 20 76-100% 0 0 13.8 4 37.9 11 13.8 4 6.9 2 36.7 11 X 2 = 6.19 X 2 = 3.20 X 2 = 2.24 X 2 = 2.39 X 2 = 1.96 X 2 = 2.68 p = .192 p = .521 p = .687 p = .674 p = .737 p = .653
311 Appendix A (Continued) Table 34 Results of ANOVAs with Means and Standard Deviation s of Client Violence per Percentage of Evening Hours Worked ( N = 588) % of Evening Hours Worked Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence M SD M SD M SD M SD M SD M SD No evening hours ( n = 197) .04 .23 .24 .80 1.38 5.40 .12 .48 .12 .59 1.89 6.24 1-25% evening hours ( n = 253) .05 .28 .23 .74 1.08 2.38 .16 .54 .14 .59 1.68 3.29 26-50% evening hours ( n = 65) .02 .12 .32 .83 1.07 2.03 .09 .34 .09 .38 1.60 2.77 51-75% evening hours ( n = 43) .16 .57 .53 1.32 2.14 4.45 .09 .36 .20 .80 3.02 5.8 6 76-100% evening hours ( n = 30) .00 .00 .23 .68 1.47 3.08 .20 .55 .20 .80 2.10 3.84 F = 2.34 F = 1.35 F = 81 F = .54 F = .30 F = .83 p = .054* p = .250 p = .520 p = .704 p = .881 p = .501 Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
312 Appendix A (Continued) Table 35 Results of Chi Square Tests for Practicum Exposure Rate to Client Violence by Practice Setting ( N = 585) Practice Setting Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence % n % n % n % n % n % n Medical/ Health 2.8 2 9.9 7 39.4 28 11.3 8 1.4 1 43.7 31 Alcohol/ Substance abuse 7.1 3 14.3 6 50.0 21 16.7 7 14.3 6 52.5 22 Developmental disabilities 13.3 2 20.0 3 46.7 7 0 0 6.7 1 53.3 8 Corrections/ Criminal Justice 2.6 1 13.2 5 36.8 14 10.5 4 7.9 3 42.1 16 Community Organization 0 0 5.3 2 18.4 7 5.3 2 5.3 2 20.5 8 Child & Family/Child Protection 2.9 3 9.6 10 30.8 32 14.4 15 5.8 6 38.5 40 Family services 2.0 1 15.7 8 39.2 20 9.8 5 5.9 3 37.3 19 Schools 4.3 2 19.6 9 37.0 17 6.5 3 8.7 4 40.4 19 Services to the Aging 7.3 3 14.6 6 34.1 14 7.3 3 2.4 1 34.1 14 Mental health/ Psychiatric 3.0 3 23.0 23 47.0 47 9.0 9 14.0 14 54.5 15 X 2 = 9.67 X 2 = 13.60 X 2 = 15.45 X 2 = 12.47 X 2 = 15.62 X 2 = 19.17 p = .470 p = .192 p = .117 p = .254 p = .111 p = .038* Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
313 Appendix A (Continued) Figure 10
314 Appendix A (Continued) Table 36 Results of ANOVAs with Means and Standard Deviation s of Client Violence among Practice Settings ( N = 585) Practice Setting Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence M SD M SD M SD M SD M SD M SD Medical/ Health (n = 71) .04 .26 .17 .61 .85 1.37 .15 .55 .01 .12 1.27 2.04 Alcohol/ Substance abuse (n = 42) .14 .52 .23 .66 1.60 4.25 .38 .96 .21 .68 2.45 5.34 Developmental disabilities (n = 15) .27 .80 .47 1.30 1.73 3.90 .00 .00 .13 .52 4.26 6.8 0 Corrections/ Criminal Justice (n = 38) .00 .00 .24 .75 2.60 8.41 .13 .41 .24 .91 3.37 10.3 2 Community Organization (n = 39) .00 .00 .05 .22 .71 1.95 .05 .22 .13 .66 1.07 2.58 Child & Family/Child Protection (n = 104) .03 .17 .21 .77 .84 1.67 .19 .58 .09 .47 1.36 2.46 Family services (n = 51) .02 .14 .41 1.15 1.08 2.44 .12 .38 .14 .63 1.52 3.2 0 Schools (n = 47) .04 .20 .36 .99 2.10 7.64 .06 .24 .09 .28 2.85 9.44 Services to the Aging (n = 41) .07 .26 .22 .61 .78 1.45 .12 .51 .02 .16 1.21 2.43 Mental health/ Psychiatric (n = 101) .05 .32 .44 1.00 1.73 3.37 .11 .37 .26 .77 2.59 4.3 3 F = 1.77 F = 1.22 F = 1.37 F = 1.88 F = 1.28 F = 1.74 p = .063 p = .274 p = .192 p = .040* p = .238 p = .068 Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
315 Appendix A (Continued) Table 37 Results of Chi Square Tests for Practicum Exposure Rates per Race Match of Student/Field Educator ( N= 588) Race Match Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence % n % n % n % n % n % n Matched race 2.6 9 14.7 50 39.4 134 9.7 33 7.6 26 42.8 146 Student White/ Educator minority 5.3 4 10.7 8 28.0 21 5.3 4 8.0 6 37.3 28 Student minority/ Educator White 2.6 2 13.2 10 39.5 30 11.8 9 2.6 2 43.4 33 Student minority/ Educator minority 8.3 1 25.0 3 50.0 6 16.7 2 25.0 3 53.4 8 X 2 = 2.55 X 2 = 2.06 X 2 = 4.28 X 2 = 2.72 X 2 = 8.06 X 2 = 1.54 p = .466 p = .559 p = .233 p = .437 p = .045* p = .773 Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
316 Appendix A (Continued) Figure 11
317 Appendix A (Continued) Table 38 Results of ANOVAs with Means and Standard Deviation s of Client Violence per Racial/Ethnic Match of Student and Field Educator ( N = 505) Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence Practice Setting M SD M SD M SD M SD M SD M SD Matched student/field educator race ( n = 341) .32 .23 .28 .83 1.26 3.57 .13 .45 .13 .55 1.83 4.41 Student White/Field educator minority ( n = 75) .53 .22 .19 .69 .73 1.86 .09 .50 .13 .55 1.20 2.43 Student minority/Field educator White ( n = 76) .03 .16 .21 .62 .89 1.66 .16 .49 .05 .36 1.34 2.28 Student/field educator of differing minorities ( n = 13) .15 .55 .46 .97 1.00 1.53 .46 1.20 .69 1.49 2.76 3. 96 F = 1.27 F = .65 F = .77 F = 2.10 F = 4.69 F = 1.08 p = .284 p = .585 p = .511 p = .098 p = .003*** p = .357 Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
318 Appendix A (Continued) Table 39 Results of Pearsons Correlations between Incidents of Client Violence and Total Training in Training Venues ( N = 567) Training Venues Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence r p r p r p r p r p r p Social work classes .02 .717 -.01 .848 .07 .098 -.01 .947 .03 .434 .06 .151 Field Seminar .02 .611 .04 .402 .10 .013* .08 .058 .07 .122 .11 009** Field Agency .11 .009** .14 .001*** .09 .029* .08 .044* .11 .010 ** .13 .002*** Other places -.01 .888 .04 .363 .08 .046* .06 .154 .10 .022* .09 .027* Total Training .06 .179 .04 .049* .14 .001*** .09 .039* .12 .004** .15 .000*** Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
319 Appendix A (Continued) Table 40 Results of Pearsons Correlations between Incidents of Client Violence and Total Training per Content Area ( N = 569) Content Areas Physical Assault Threat of Physical H arm Verbal Abuse r p r p r p Assess history .07 .087 .05 .245 .05 .245 Violence characteristics .05 .399 .09 .032* .12 .005** Mental illness signs of violence .05 .274 .05 .234 .11 .009** High risk situations .03 .447 .04 .360 .10 .023* Managing feelings in self .06 .184 .01 .842 .06 .181 Confident demeanor .08 .062 .06 .188 -.01 .753 Verbal signs of violence .03 .513 .04 .368 .08 .069 Physical signs of violence .11 .009** .11 .008** .10 .019* Student rights .05 .213 .10 .017* .06 .177 Verbal de-escalation .09 .022* .10 .022* .11 .012** Office safety -.02 .658 .03 .493 .12 .006** Where to sit -.00 .957 -.03 .403 -.08 .050* Home visit safety .02 .608 .01 .855 .03 .484 Inform supervisor of whereabouts -.01 .789 -.04 .370 .09 .031* Physical techniques for self defense .10 .016* .08 .064 .14 .001*** Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
320 Appendix A (Continued) Table 40 (Continued) Results of Pearsons Correlations between Incidents of Client Violence and Total Training per Content Area ( N = 569) Content Areas Threat of lawsuit Property Damage Tot al Client Violence r p r p r p Assess history .05 .196 .03 .445 .06 .137 Violence characteristics .03 .530 .12 .005** .13 .002*** Mental illness signs of violence .01 .861 .15 .000*** .12 .004*** High risk situations .11 .008** .08 .067 .11 .010** Managing feelings in self .03 .461 .06 .129 .06 .141 Confident demeanor .08 .056 .06 .193 .02 .665 Verbal signs of violence .04 .322 .10 .020* .09 .039* Physical signs of violence .01 .758 .11 .011** .12 .004*** Student rights -.00 .935 .07 .109 .08 .075 Verbal de-escalation .07 .104 .10 .022* .13 .002*** Office safety .06 .911 .01 .864 .10 .017* Where to sit .06 .168 .06 .152 .09 .038* Home visit safety .07 .087 -.01 .764 .03 .433 Inform supervisor of whereabouts .09 .031* .03 .430 .09 .023* Physical techniques for self defense .04 .321 .14 .001*** .15 .000*** Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
321 Appendix A (Continued) Table 41 Results of Pearsons Correlations between Direct In cidents of Client Violence and Occupational Commitment/Career Withdrawal Intentions ( N = 586) (Direct Experience) Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence r p r p r p r p r p r p Affective Commitment .04 .359 .03 .488 .01 .910 .00 .998 .02 .679 .02 .6 82 Normative Commitment .11 .011** .05 .271 .01 .733 .02 .572 .05 .192 .03 .474 Continuance Commitment .06 .143 .04 .384 .02 .686 -.07 .091 -.01 .776 .01 .837 Career Withdrawal Intentions -.06 .162 .00 .839 .780 .35 .00 .944 .05 .516 .01 858 Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
322 Appendix A (Continued) Table 42 Results of Pearsons Correlations between Indirect Exposure to Client Violence and Occupational Commitment/Career Withdrawal Intentions ( N = 589) (Indirect Exposure) Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence r p r p r p r p r p r p Affective Commitment -.02 .580 .00 .926 .04 .301 -.01 .881 -.01 .886 .02 .573 Normative Commitment .00 .954 -.05 .280 -.02 .672 -.03 .464 .01 .852 -.0 2 .642 Continuance Commitment .03 .417 .01 .830 -.01 .871 .02 .686 -.01 .765 .01 .898 Career Withdrawal Intentions -.05 .165 .01 .811 .00 .654 .01 .919 .05 .6091 .01 .677
323 Appendix A (Continued) Table 43 Results of Pearsons Correlations between Direct Ex perience of Specific Types of Incidents of Violence and Fears of Same Type of Violence and General Fear of Violence ( N = 582) (Direct Experience) Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence r p r p r p r p r p r p Fear of Physical Assault .08 .042* ----------Fear of Threat of Harm --.22 .000*** --------Fear of Verbal Abuse ----.13 .000*** ------Fear of Threat of Lawsuit ------.19 .000*** ----Fear of Property Damage --------.09 .040* --Fear of Any Type of Client Violence .02 .564 .18 .000*** .13 .002*** .15 .000*** .08 .0 43*. .15 .000*** Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
324 Appendix A (Continued) Table 44 Results of Pearsons Correlations between Indirect Exposure to Specific Types of Incidents of Violence and Fears of Same Type of Violence and General Fear of Violence ( N = 582) (Direct Experience) Physical Assault Threat of Physical Harm Verbal Abuse Threat of lawsuit Property Damage Total Client Violence r p r p r p r p r p r p Fear of Physical Assault .17 .000*** ----------Fear of Threat of Harm --.23 .000*** --------Fear of Verbal Abuse ----.21 .000*** ------Fear of Threat of Lawsuit ------.29 .000*** ----Fear of Property Damage --------.14 .001*** --Fear of Any Type of Client Violence .17 .000*** .21 .000*** .21 .000*** .27 .000*** .15 .000*** .27 .000*** Note: p .05, ** p .01, *** p .004 (Significant after Bonferroni adjustment)
325 Appendix A (Continued) Table 45 Results of Pearsons Correlations between Fear of V iolence and Occupational Commitment ( N = 589) Fear of Physical Assault Fear of Threat of Physical Harm Fear of Verbal Abuse Fear of Threat of lawsuit Fear of Property Damage Fear of Any Client Violence r p r p r p r p r p r p AC -.10 .014** -.08 .046* -.01 .820 -.04 .285 -.07 .08 9 -.07 .102 NC .14 .001*** .13 .002*** .18 .000*** .13 .002*** .12 .003*** .17 .000*** C C .17 .003*** .12 .004*** .17 .000*** .15 .000*** .12 .004*** .17 .000*** Note: p .05, ** p .01,*** p .004 (Significant after Bonferroni adjustment) Note: AC =Affective Commitment; NC = Normative Comm itment; CC = Continuance Commitment
326 Appendix A (Continued) Table 46 Model Summary for Predictors of Occurrence of Overa ll Client Violence Step Factors R R 2 R 2 adj R 2 Fchg p df1 df2 1 Male .173 .030 .028 .030 15.66 .000 1 507 2 Mental health setting .208 .043 .039 .013 6.98 .008 1 506 3 11 or more home visits .231 .053 .048 .010 5.48 .020 1 505 4 Alcohol/substance abuse setting .251 .063 .056 .0 10 5.14 .024 1 504 Coefficient for Final Model of Occurrence of Overal l Client Violence Factor B t p Bivariate r Partial r Male .201 .125 2.84 .005 .151 .126 Mental health setting .135 .108 2.43 .015 .109 .108 11 or more home visits .115 .109 2.48 .014 .098 .110 Alcohol/substance abuse setting .206 .106 3.37 .018 .085 .105
327 Appendix A (Continued) Table 47 Model Summary for Predictors of Frequency of Overal l Client Violence Step Factors R R 2 R 2 adj R 2 F chg p df 1 df 2 1 Male .183 .034 .029 .034 6.68 .011 1 192 Coefficient for Final Model of Overall Occurrence o f Client Violence Factor B t p Bivariate r Partial r Male .140 .183 2.58 .011 .183 .183
328 Appendix A (Continued) Table 48 Model Summary for Predictors of Occurrence of Physi cal Assault Step Factors R R 2 R 2 adj R 2 F chg p df 1 df 2 1 Developmental disabilities setting .103 .011 .009 .010 5.511 .021 1 509 2 Field agency training .144 .021 .017 .010 5.271 022 1 508 Coefficient for Final Model of Overall Occurrence o f Physical Assault Factor B t p Bivariate r Partial r Developmental disabilities setting .117 .106 2.42 .016 .103 .107 Field agency training .004 .101 2.40 .022 .097 .101
329 Appendix A (Continued) Table 49 Model Summary for Predictors of Occurrence of Threa t of Physical Harm Step Factors R R 2 R 2 adj R 2 Fchg p df1 df2 1 Male .210 .044 .042 .044 23.22 .000 1 505 2 Mental health setting .231 .053 .050 .010 5.073 025 1 504 Coefficient for Final Model of Occurrence of Threat of Physical Harm Factor B t p Bivariate r Partial r Male .230 .198 4.53 .000 .210 .198 Mental health setting .089 .098 2.25 .025 .122 .100
330 Appendix A (Continued) Table 50 Model Summary for Predictors of Occurrence of Verba l Abuse Step Factors R R 2 R 2 adj R 2 Fchg p df1 df2 1 Male .151 .023 .021 .023 11.76 .001 1 505 2 Mental health setting .176 .031 .027 .008 4.31 .038 1 504 3 11 or more home visits .199 .040 .034 .009 4.46 .035 1 503 4 Alcohol/substance abuse setting .224 .050 .043 .0 11 5.60 .018 1 502 Coefficient for Final Model of Occurrence of Verbal Abuse Factor B t p Bivariate r Partial r Male .201 .125 2.84 .005 .151 .126 Mental health setting .135 .108 2.43 .015 .109 .108 11 or more home visits .115 .109 2.48 .014 .098 .110 Alcohol/substance abuse setting .206 .106 2.37 .018 .085 .105
331 Appendix A (Continued) Table 51 Model Summary for Predictors of Occurrence of Threa t of Lawsuit Step Factors R R 2 R 2 adj R 2 Fchg p df1 df2 1 Age 25 to 30 .153 .023 .021 .02 12.11 .001 1 505 2 Mixed ethnic heritage .210 .044 .040 .02 10.96 .001 1 504 3 11 or more home visits .239 .057 .052 .01 6.81 .009 1 503 4 Alcohol/substance setting .259 .067 .059 .01 5.29 .022 1 502 Coefficient for Final Model of Occurrence of Threat of Lawsuit Factor B t p Bivariate r Partial r Age 25 to 30 .098 .162 3.76 .000 .153 .165 Mixed ethnic heritage .201 .130 2.99 .003 .140 .132 11 or more home visits .079 .120 2.95 .003 .113 .131 Alcohol/substance abuse setting .114 .101 2.30 .022 .092 .102
332 Appendix A (Continued) Table 52 Model Summary for Predictors of Occurrence of Prope rty Damage Step Factors R R 2 R 2 adj R 2 F chg p df 1 df 2 1 Male .136 .019 .017 .02 9.53 .003 1 505 2 Mixed minority supervision dyad .182 .033 .029 .02 7.66 .006 1 504 3 Mental health setting .215 .046 .040 .01 6.79 .00 9 1 503 Coefficient for Final Model of Occurrence of Proper ty Damage Factor B t p Bivariate r Partial r Male .112 .128 2.92 .004 .136 .127 Mixed minority supervision dyad .222 .125 2.86 .004 .114 .126 Mental health setting .078 .114 2.61 .009 .126 .115
333 Appendix A (Continued) Table 53 Percentages of Field Directors Reporting Giving Tra ining Content versus Students Reporting Receiving Training Content Safety content areas Faria and Kendra (% of field directors/programs reporting giving training in social work classes) N = 13 Criss (This study) (% of students reporting receiving training in social work classes) N = 595 Characteristics of high risk situations-non-public, isolated places 100% ( n = 13) 48.1% ( n = 286) Creating safe office space 92% ( n = 12) 38.8% ( n =231) Maintaining a confident, secure demeanor 92% ( n = 12) 52.8% ( n = 214) Verbal de-escalation 92% ( n = 12) 44.0% ( n = 262) Where to sit when interacting with a client 92% ( n = 12) 58.0% ( n = 345) Characteristics/life experiences of people more likely to commit violent acts 85% ( n = 11) 42.2% ( n = 251) Keeping supervisor informed of ones itinerary 85% ( n = 11) 35.6% ( n = 212) Recognizing verbal acts of violence 85% ( n = 11) 41.8% ( n = 249) Self awareness of feelings 85% ( n = 11) 51.1% ( n = 301) Physical signs that attack is imminent 77% ( n = 10) 24.9% ( n = 148) Forms of mental illness associated with violent behavior 69% ( n = 9) 55.5% ( n = 330) What to do if one is a victim of violence( filing reports, dealing with physical and emotional aspects) 38% ( n = 5) 31.3% ( n = 186) Training in non-violent defense 31% ( n = 4) 27.6% ( n = 164)
334 Appendix B: IRB Letter
335 Appendix B (Continued)
336 Appendix C: in Focus NASW Information
337 Appendix C (Continued)
338 Appendix D: Client Violence Questionnaire
339 Appendix D (Continued)
340 Appendix D (Continued)
341 Appendix D (Continued)
342 Appendix D (Continued)
343 Appendix E: First Mailing Cover Letter March 18, 2008 Dear Social Work Student: Social work students exposure to client violence d eserves our attention. As a part of a select random sample from national NASW student members, y ou are being asked to participate in an important study for the field of social work. This is the first nationwide study to ask social work students opinions about client violence and how th is may impact commitment to the social work profession. As a social work student, your knowledg e and experience is vital for the success of this research. By completing this questionnaire, yo u will make a valuable contribution to the advancement of knowledge in this area. A genuine attempt has been made to make the items i n the survey instrument straightforward and clear. Almost all of the items can be answered by s imply checking a circle. The survey has a total of 80 questions. You should be able to complete the instrument in about 12-15 minutes. Though your participation is valuable, it is entire ly voluntary. You may choose to skip any questions that you do not wish to answer. Return of the questionnaire will imply your consent to participate. All returned questionnaires are secure d in a locked cabinet, accessible only to the primary researcher. Data regarding individual participants will be tot ally confidential. Your name will never be placed on the questionnaire itself. The questionnai re has an identification number for mailing purposes only. This is so that your number can be t aken off the mailing list when your questionnaire is returned. The information being requested in this survey may be of a sensitive and personal nature. Potential risks of participation include possible n egative feelings related to the recall of incidents of client violence. If you have any concerns that y ou wish to discuss you may contact the primary researcher, Pam Criss, at (863) 667-5153. Please consider participation in this important stu dy. Benefits to the field of social work are an increased understanding of incidents of client viol ence in the internship and the short term and long term effects on students. With this knowledge, we can understand more about how to prevent and appropriately address client violence, if it should occur during the practicum.
344 Appendix E (Continued) Upon completion of the questionnaire, return it in the enclosed self-addressed, stamped envelope. Please return the questionnaire by April 4, 2008. I f you have questions regarding this study, please call me at (863) 6675153 or email me at email@example.com You may also contact Dr. Lisa Rapp-Paglicci at (863) 974-1809. Your contribution to the success of this study is g reatly appreciated. Sincerely, Pam Criss, MSW, LCSW Doctoral Candidate School of Social Work University of South Florida Tampa, Florida
345 Appendix F: Second Mailing Cover Letter University of South Florida College of Arts and Sciences School of Social Work Tampa, Florida April 8, 2008 Dear Social Work Student: I am writing to you about our study of social work students experience with client violence. Three weeks ago the Client Violence Quest ionnaire was mailed to you. This is the first nationwide study to ask social work stude nts about their experience with client violence and its possible impact on their career de cisions. It is hoped that by increasing our understanding about social work students expos ure to client violence, we can help to reduce incidents of client violence and improve tra ining that is offered on this subject. If you have already completed and mailed the survey please accept my sincere thanks. If not, please do so today. Because the survey was onl y sent to a small random sample of social work students it is very important that your opinion be included in the study if the results are to be representative of social wo rk student in the United States. Participation in the research is strictly voluntary You may withdraw at any time without penalty. You may skip any questions that you do not wish to answer. All responses are anonymous. Potential risks of participation include possible negative feelings related to the recall of incidents of client violence. All ret urned questionnaires are secured in a locked cabinet, accessible only to the primary rese archer. Upon completion of the study, the questionnaires will be shredded and disposed of In the event that your questionnaire has been mispl aced, a replacement is being enclosed. I would be happy to answer any questions or concern s you have about the study. Please contact me by email at firstname.lastname@example.org or by phone at (863) 667-5153. If you are interested in the results of the study I will b e happy to forward the results as soon as they are available.
346 Appendix F (Continued) Your contribution to the success of this study is g reatly appreciated. Sincerely, Pam Criss, MSW, LCSW Doctoral Candidate University of South Florida School of Social Work Tampa Florida
347 Appendix G: Pre-Mailing Post Card March 11, 2008 Dear Social Wor k Student: You have been randomly selected to participate in a n important national study of client violence against social work students. Be watching in the coming week for the su rvey instrument and more details. Thank you in advance for your assistance with this study. Pam Criss, MSW, LCSW Doctoral candidate University of S outh Florida Tampa, Florida
348 Appendix H: Emailed Permissions to Use Scales/Quest ions Email from Dr. Christina Newhill Dear Ms. Criss, It is a great idea to look at the issue of client v iolence with social work students. Not much has been done in that realm just a couple of studies that I am aware of. You are welcome to use my questionniare in my book and addi ng a few questions to speak specifically to the issues relevant to students is good. I would suggest questions addressing the following: (1) have the students rec eived any education in the classroom about working with violent and aggressive clients particularly those who are involuntary and how to do a violence risk assessment?; (2) Do es the agency where they do their field placement have a safety policy in place and w as the student appraised of such a policy? (3) Is the issue of safety part of the stud ent's field learning plan? Those are just a few things I can think of immediately. Questions re lated to whether they have experienced incidents of violence are included in t he questionnaire and the wording can be modified to be relevant to the student experienc e. Let me know if I can be of further assistance. Best of luck with your dissertation! Christina Newhill Christina E. Newhill, Ph.D.,LCSW Associate Professor School of Social Work 2217F Cathedral of Learning University of Pittsburgh Pittsburgh, PA 15260 Telephone: (412) 624-6330 Fax: (412) 624-1159
349 Appendix H (Continued) Email concerning Fear of Future Violence Scale from Dr. Kevin Kelloway Pam Sorry about this, I was using my memory and got th e wrong email for aaron. I am now back in the office. You certainly have per mission to use the fear and/or support scales in your own w ork. I am attaching a file containing the support measures we used in the study Best of luck with your research kevin -----Original Message----From: Pam Criss [ mailto:email@example.com ] Sent: Wednesday, June 13, 2007 12:54 PM To: kevin.kelloway@SMU.CA Subject: RE: Fear of future violence scale Thanks for your response. I have not heard from Dr. Schat. I am not in a great hurry to get the scales, so I c an wait until you return. I actually have seen one of the s cales, Fear of Future Violence Scale, in a dissertation by Ki-bum Song, Columbia University, 2005. However, I wanted to get permission from you to consider using the scale in my research. I would still like to see the scales on Instrumental and Informational support, when you ha ve a chance to send them. Thanks for your help with this Pam Criss, MSW, LCSW Social Work Program Field Coordinator Southeastern University 1000 Longfellow Blvd. Lakeland, FL 33801 (863) 667-5153 Fax (863) 667-5200
About the Author Pamela Myatt Criss is Associate Professor and Field Coordinator for the Social Work Program at Southeastern University in Lakeland Florida. She obtained her B.A. from Florida Southern College and her M.S.W. from U niversity of South Florida. Ms. Criss has over 30 years experience as a social work er, predominantly in the field of child welfare and she is a licensed clinical social worke r. She co-wrote accreditation documents that culminated in Council on Social Work Education accreditation of the Social Work Program at Southeastern University in F ebruary, 2006. Ms. Criss was nominated for Whos Who among Americas Teachers in 2005 and 2007. Additionally she was named Social Worker of the Year for her loc al unit of National Association of Social Workers in 2003.
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 2200397Ka 4500
controlfield tag 001 002029019
007 cr mnu|||uuuuu
008 090915s2009 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002843
Criss, Pamela Myatt.
Prevalence of client violence against social work students and its effects on fear of future violence, occupational commitment, and career withdrawal intentions
h [electronic resource] /
by Pamela Myatt Criss.
[Tampa, Fla] :
b University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains 349 pages.
Dissertation (Ph.D.)--University of South Florida, 2009.
Includes bibliographical references.
Text (Electronic dissertation) in PDF format.
ABSTRACT: Social work literature has documented that social workers may be the victims of client violence. However, to date, no studies have documented the nationwide prevalence of client violence towards social work students. This study examined direct and indirect incidents of physical assault, threats of physical harm, verbal abuse, threats of lawsuit, and property damage. The randomly selected national sample of social work students were selected from the National Association of Social Workers (N = 595). Findings revealed that 41.7% of social work students directly experienced client violence during their practicum. The highest rate of the violence reported by students was verbal abuse (37.5%) while the lowest rate of reported violence was physical assault (3.5%). Being male was the most significant predictor of social work students' exposure to client violence. Other factors related to increased violence were found, such as ethnicity and degree program. This study also examined whether students received safety training in 17 content areas and where they received the training. Fewer than 50% of students received training in most training content areas, regardless of where training was received. Furthermore, increased safety training in the field agency was significantly related to increased threats of physical harm and overall client violence. When training from all venues was totaled, increased training was significantly related to increased verbal abuse, property damage, and overall direct client violence. This study found that when students experience client violence directly or indirectly, they have increased fear of future violence in social work practice. Implications for social work programs, field agencies and educators and social work students are discussed. Training content and strategies are suggested.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
Advisor: Lisa Rapp-Paglicci, Ph.D.
Career turnover intentions
x Social Work
t USF Electronic Theses and Dissertations.