A Latent Class Analysis of Childhood Victimization and Intimate Partner Violence Among Women on Probation and Parole: Examining Mental Health and Health Risk Behaviors

Schedule:
Saturday, January 17, 2015: 8:00 AM
La Galeries 4, Second Floor (New Orleans Marriott)
* noted as presenting author
Katherine M. Winham, MS, MSSW, Doctoral Candidate, University of Louisville, Jacksonville, FL
Tanya Renn, MSSW, MPH, Doctoral Candidate, University of Louisville, Louisville, KY
Martin T. Hall, PhD, MSSW, Assistant Professor, University of Louisville, Louisville, KY
Seana Golder, PhD, Associate Professor, University of Louisville, Louisville, KY
George E. Higgins, PhD, Professor, University of Louisville, Louisville, KY
TK Logan, PhD, Professor, University of Kentucky, Lexington, KY
Purpose: Women represent the fastest growing segment of the criminal justice population in the United States. The vast majority (85%) of criminal justice involved women are sanctioned in the community (i.e., probation or parole). Victimization is known to be a common experience among women in the criminal justice system; 70% report a history of childhood victimization and 75% report experiencing intimate partner violence (IPV). These rates are higher than women in the general population where 52% report childhood victimization and 22% report IPV. While the prevalence of victimization among this population is well documented, little is known regarding heterogeneity among women’s victimization experiences and the impact of these different profiles on mental health (anxiety, depression, somatization) and health risk behaviors (substance use, HIV sexual risk behaviors). To this end, this study used latent class analysis to: 1) identify meaningful subgroups of women based on indicators of childhood victimization and IPV; 2) examine differences among identified groups based on indicators of socio-demographics, mental health, and health risk behaviors.

 

Method: Women (N=406), on probation and parole, with a history of victimization (physical or sexual as a child or an adult) were recruited from an urban Midwest county.  Women were recruited through probation/parole offices or flyers throughout the community, then participated in face-to-face, audio computer assisted interviews. Childhood victimization (physical, sexual, and psychological abuse; neglect) and IPV (physical, sexual, and psychological) were measured through items from the Revised Conflict Tactics Scale, Tolman’s Psychological Maltreatment of Women, and Mullings childhood neglect measure. Latent class analysis was used to identify subgroups of women based on patterns of childhood victimization and IPV. Between group differences were then examined across indicators of mental health (Brief Symptom Inventory subscales: anxiety, depression, somatization) and health risk behaviors (substance use, HIV sexual risk behaviors).

 

Results: LCA yielded a three-class solution (AIC =8979.97, BIC=9136.22, ABIC =9012.47, LMR=62.03, p=.79). Class 1 (9%; n =39) was comprised of women with moderate levels of childhood victimization and high levels of IPV. Class 2 women (17%; n =68) reported high levels of childhood victimization and moderate levels of IPV. Class 3 (73%; n =299) women evidenced low childhood victimization and low IPV (n =299). Chi-squared test results indicated that women in the first two classes reported significantly greater depression, anxiety, and somatization when compared to women in the third class (low childhood victimization, low IPV). Also, women in the second class were significantly more likely to report at least one year of crack/cocaine use when compared to women in the third class. No other socio-demographic, substance use, or HIV sexual risk behavior differences were noted.

Conclusions: The findings highlight the heterogeneity of victimization experiences among women on probation and parole and underscore the clinical utility of adapting treatment to meet the mental health needs of women with specific victimization profiles. Implications for policy and practice will be discussed.