Essays on the political economy of the centralized provision of local public goods

Background and objectives. This dissertation's analyses examined the association of caregiver, child, abuse and investigation characteristics with maternal response and emotional harm among families for whom child sexual abuse (CSA) was suspected or substantiated.;Results. According to social worker assessment, the majority of female non-offending caregivers (87.1%) of children investigated for suspected or substantiated child sexual abuse responded with belief of the abuse disclosure, emotional support and protection of the child victim. The overall maternal response model was significant and accounted for 40.8% of the variance (Nagelkerke R2). Factors significantly associated with maternal response in the multivariate model included maternal mental health, age of the child, child's manifestation of sexualized behavior, child's relationship to the perpetrator, duration of abuse and co-occurring maltreatment. The overall emotional harm model was significant and accounted for 18.3% of the variance (Nagelkerke R2). Age of the child at the time of investigation, inappropriate sexualized behavior and substantiation level were significant predictors in the final block of the emotional harm regression equation. Maternal response was no longer significantly associated with emotional harm when the analysis adjusted for child characteristics.;Conclusion. Consistent with previous research, the majority of non-offending mothers investigated as part of the CIS-98 responded to CSA disclosure with belief, emotional support and protection as determined by the social worker's assessment. The current study supports the cumulative evidence that caregiver mental health, age of the child and the child's relationship with the offender are significant predictors of negative maternal response and emotional harm. Although negative maternal response failed to predict emotional harm among children investigated for CSA, continued examination of the risk factors associated with maternal response is warranted to ensure the safety of a small, yet vulnerable segment of children.;Method. This study was based on secondary analysis of data collected in the Canadian Incidence Study of Report Child Abuse and Neglect 1998. The current analysis was limited to 373 CSA investigations for which there was a female non-offending caregiver and complete data on maternal response. Bivariate and hierarchical logistic regression analysis was conducted for two outcomes; maternal response and emotional harm.