This study compares different versions of mirror exposure (ME), a body image intervention with research support. ME protocols were adapted to maximize control and comparability, and scripted for delivery by research assistants. Female undergraduates (N = 168) were randomly assigned to receive mindfulness-based (MB; n = 58), nonjudgmental (NJ; n = 55), or cognitive dissonance-based (CD, n = 55) ME. Participants completed the Body Image Avoidance Questionnaire (BIAQ), Body Checking Questionnaire (BCQ), Satisfaction with Body Parts Scale (SBPS), Beck Depression Inventory-II (BDI-II), and Eating Disorders Examination Questionnaire (EDE-Q) at pre-treatment, post-treatment, and 1-month follow-up. Mixed models ANOVAs revealed a significant main effect of time on all measures, and no significant time by condition interaction for any measures except the SBPS. Post-hoc analysis revealed that only CD ME significantly improved SBPS outcome. Results suggest that all versions of ME reduce eating disorder risk factors, but only CD ME improves body satisfaction.