The purpose of this dissertation research is to analyze criminal justice offender outcomes from child sexual abuse (GSA) cases interviewed by the Children's Advocacy Services of Greater St. Louis (CASGSL). The study includes 203 child sexual abuse cases referred to the CASGSL for forensic interviews from the County of St. Louis, MO, from 2003-2003. The overarching research purpose is to determine what effects, if any, a variety of child/victim, family, incident, offender, and evidentiary characteristics have on the likelihood of conviction. The research collected a wide variety of victim and incident data from information directly from child/victims and their caregivers, prior to referral to prosecution, tracked these victim's cases from interview to prosecution and conviction, and developed statistical models that control for a variety of variables. The study uses a theoretical framework of uncertainty avoidance and victim credibility and blameworthiness. Findings are compared with other studies to cross-validate the results.

In the final logistic regression models, the presence of a verbal or mental disability in the child was associated with a decreased likelihood of conviction. The family characteristics of intimate partner violence (IPV) reported in the household and caregiver support were related with conviction; specifically, IPV decreased the likelihood of conviction, and caregiver support increased the likelihood of conviction. Marital status of caregivers, employed caregiver, and legal action pending (divorce, custody) were not associated with conviction. None of the offender characteristics for which data was available were related to conviction; however, offender data availability was limited. Several incident characteristics were assessed including the victim/offender relationship, nature of the abuse, use of coercive tactics by the perpetrator including force, and duration of abuse. The only incident characteristic that was significantly associated with conviction was child being forced to masturbate the abuser. Children's Division as the referral source was associated with a decreased likelihood of conviction, while conversely, law enforcement increased the likelihood. The final model explained 77% of the variance; thus, predicting which cases are most likely to result in a conviction is not clear cut as much of the variation in this outcome appears to be randomly distributed.

This research further proposes a decision-making framework of focal concerns related to the prosecutorial uncertainty avoidance in child sexual abuse cases. In this framework, prosecutors are concerned with a variety of considerations associated with: (1) the victim, (2) the victim's family, (3) the victim/offender relationship, (4) incident/evidentiary characteristics, and (5) protection of the community. Future research should examine victim blameworthiness, particularly related to teen victims. Second, the influence of intimate partner violence and other family dysfunctions requires further research. Third, qualitative research is needed on the organizational structures and processes within which decisions related to CSA cases are made. Fourth, research is needed into the role of law enforcement in these cases. Research has consistently found a positive association between law enforcement involvement and prosecution outcomes, but there is little or no empirical discussion as to why this is the case. Finally, there is a need for research on child and family outcomes.

Keywords. Children's Advocacy Center, child sexual abuse, prosecutorial decision making, victimization, specialized prosecution, uncertainty avoidance, focal concerns theory, intimate partner violence.