This study assessed the relative importance of problem gambling, cocaine use and selected psychosocial characteristics in predicting criminal convictions. A self-administered questionnaire was completed by clients when admitted to a treatment program for a primary problem with cocaine (n = 300), gambling (n = 199), or tobacco (n = 249). The questionnaire included questions on various psychosocial characteristics and criminal conviction(s) in the past 3 years. Those in treatment for cocaine use or gambling had higher rates of criminal conviction (21.8 % and 11.6 %, respectively) than those in treatment for tobacco (1.6 %). In a multivariate analysis, being in treatment for cocaine or gambling, compared to those in treatment for tobacco, was significantly associated with having had a criminal conviction; while having high levels of social support and being employed full-time were protective factors against criminal conviction. Due to the cross-sectional nature of the study causation cannot be inferred.