Comparisons between Recidivists
This chapter presents the statistical analyses comparing recidivists of domestic violence with non-recidivists using bivariate cross tabulations. To examine differences in the means for interval/ratio or continuous variables, this study employs t-tests. To measure effect sizes, this study uses Hedge’s g. The chi-square tests of significance determine the probability of associations of nominal-level and dichotomous variables with recidivism of domestic violence. The phi coefficient reflects the strength of the associations. In addition to the descriptive statistics and analysis of mean differences, this study uses Cox proportional hazards regression analysis to predict the risk of recidivism of domestic violence over time, comparing the number of community legal sanctions (or dose-response) imposed for the index case.
During the follow-up period, 69.8% of the 607 men in the sample were not involved with either the civil or the criminal courts for domestic violence related offenses. The remaining 184 men in the sample (30.2%) were involved with the courts and met the study’s criteria for recidivism of domestic violence: any cases filed in the county for domestic violence related criminal offenses, or any protective orders processed by the county District Attorney’s office after the index case. The number of days from the index case to recidivism ranges from 1 day to 2805 days. The mean number of days