Risk Markers for Sexual Predation and Victimization in Prison

Abstract

Our research was exploratory and sought for the first time to apply empirically validated static and dynamic risk makers for violence in the community to sexual predation and victimization in prisons. We conducted this exploration using five outcome measures including experiences of coerced sexuality while imprisoned as a perpetrator or victim, involvement in sexual exchanges that were based on bartering of goods or protection as either a perpetrator or victim, and consensual acts that occurred with other inmates or members of the correctional staff. We sought to identify the most relevant risk markers through a review of clinical violence risk research, criminological research concerning sexual violence in the community and in prison, and risk classification schemes for incarcerated men and women. This review directed us to the potential relevance of an array of risk factors that included early adverse life experiences; prior adolescent and domestic violence; criminality and prison violence; hyper-sexuality and impersonal sexuality; affective and cognitive states including anger, impulsivity and thoughts of harm; personality traits and disorders including psychopathy; extant risk instruments for violence including the COVR, HCR:20 and VRAG; and the social environment as reflected in the attitudes and behaviors of the correctional staff in each institution. Each of these domains of risk markers was examined in terms of its bivariate association with the various outcome measures and later entered into a multivariate classification analysis using Chi Square Automated Interaction Detector (CHAID). This analytic process resulted in the construction of nine valid conceptual models for predatory, victimized, bartered, and consensual sex in prison, each of which was different for the male and female inmates.The models were developed using a sample of 471 inmates (288 males/ 183 females) from two states. The models were characterized by good to excellent (.70 to.99) accuracy with high levels of sensitivity and relatively low levels of specificity. In developing our models, we found that male inmates engaged in more predatory, bartered, and consensual sex with visitors or staff while the female inmates engaged in more bartered and consensual sex with other inmates. However, these different types of sex were highly related suggesting that inmates who engage in one type of sex are far more likely to be involved in the other types of sex and that these inmates are different in many ways from inmates who adopt a largely asexual stance while incarcerated. Similarly, many of the risk makers for sex in prison have been previously identified as predicting violent behavior in the community. Inmates who experienced conduct problems as juveniles, had problems establishing relationship with others during adolescence, had a past history of early violence, and suffered from one or more of the Cluster B personality diagnoses tended to be more sexually active in prison. The HCR: 20 used to predict violence risk in the community ...