Date of Award

2014

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Graduate Group

Criminology

First Advisor

John M. MacDonald

Abstract

This dissertation is a collection of three distinct chapters. The first chapter explores the associations between various assimilation measures and outcomes of delinquent risk factor development for immigrant youth from Los Angeles County, CA, using both logit regression and doubly robust estimation methods. Youth with a low level of assimilation were more likely to have a high level of attachment to family and school than those with a high level of assimilation. However, the associations were less evident for the other delinquent risk factors. Among the assimilation measures used, the generation status measure was not predictive of the outcomes. The two language-related assimilation measures were less strongly associated with the outcomes than the immigrant status measure.

The second chapter investigates crime impacts of the increased alcohol availability made through repealing a Sunday off-premise liquor sales ban. Since 2003, Pennsylvania permitted a part of its state-run liquor stores to open on Sunday by repealing the ban, which enables a quasi-experimental triple difference design. Some evidence of local crime pattern changes after the repeal was found. The repeal was associated with an increase in total crime incidents occurring in the immediate vicinity of the Sunday-open liquor stores in Philadelphia. At the same time, total crime incidents occurring relatively farther away from the stores decreased. These pattern changes were present in low socioeconomic status (SES) neighborhoods, but not evident in high SES ones.

The third chapter evaluates whether the effect of a cognitive behavioral therapy (CBT) program on recidivism differ depending on probationer characteristics from a Bayesian perspective. Using Philadelphia CBT randomized controlled experiment data with a Bayesian hierarchical Gamma-Poisson model, the study compares average recidivism rates between the CBT and non-CBT groups, conditional on probationer characteristics. The Bayesian analysis showed that the effects of the CBT program were statistically meaningfully more evident for the high-risk probationers who were between 10-19 and 30-39 years old, who had more extensive prior experience on probation, and who had a higher ratio of high risk predictions. These results contradict the frequentist evaluation results that the CBT program had no statistically significant effect on recidivism.

Share

COinS