Recent studies have indicated variable reductions in criminal thinking for justice-involved persons with mental illness exposed to cognitive-behavioral treatments. To date, however, no studies have identified risk factors for limited response or modeled observed disparities in responsivity to interventions aimed at reducing criminal thinking. Using an archival data set of 162 probationers with a dual diagnosis who were exposed to changing lives and changing outcomes, a latent profile analysis modeled unobserved heterogeneity in treatment response per observed changes in criminal thinking. Most participants endorsed significant changes in self-reported reactive criminal thinking with minimal changes in proactive criminal thinking. Neither self-reported pretreatment severity of psychopathology nor self-reported compliance with psychotropic medication predicted response to treatment. Although diagnosis also did not predict responsiveness, more favorable views of treatment predicted greater decreases in criminal thinking. Moreover, those expressing fewer levels of criminal thinking after treatment were also found to express a decrease in attitudes supportive of violence. Limitations and treatment recommendations are discussed, including the need for correctional treatments to improve responsiveness to individual treatment needs. (PsycInfo Database Record (c) 2022 APA, all rights reserved).