Abstract
Despite changes made in domestic violence (DV) programs, attrition continues to be a major problem. For this study on DV rehabilitation attrition, 62 male batterers and 31 female victims were recruited during a six month time frame from an existing batterers' program. Of the 62 batterers, one man was removed from the study, 38 dropped out of the program, and 23 made the transition from rehabilitation to the maintenance phase of the program. A logistical regression to predict completion status resulted in a Model Chi-square statistic of 31.08 (p =. 000). Completers were more likely young, court-monitored, had lower levels of stress (SOS Inventory) and posttraumatic stress (PCL), and had higher levels of mutuality (MPDQ) in their relationships than noncompleters. The model predicted 88.89% of the noncompleters, 78.26% of the completers, and had an overall predictive ability of 84.75% for the study sample.