This dissertation analyzes the relationships between criminal thinking, alliance, and psychological functioning. Secondary data for 1,589 participants was analyzed using correlations, regressions, and path analysis. It was hypothesized that variables of criminal thinking would be associated with poorer alliance and psychological functioning, and that alliance would mediate the relationship between criminal thinking and psychological functioning. Criminal thinking was moderately correlated with alliance and outcome variables. In linear regression models, alliance and criminal thinking variables were significantly predictive of most outcome variables, controlling for sociodemographic variables. Criminal thinking showed moderate negative pathways to alliance and several of the outcome variables, alliance partially mediated only one outcome variable model. Findings indicate that offenders entering treatment with higher levels of criminal thinking are less able to develop alliance with their counselor and treatment, and will have poorer behavioral functioning and higher likelihood of relapse and recidivism post-treatment. Research and practice implications are discussed.