Clinical and subclinical eating pathology are common, especially among female undergraduates. Such problems are often chronic and associated with a range of negative medical and psychological outcomes. Thus, it is important to develop effective prevention programs to reduce eating disorder risk. Numerous studies suggest that dissonance-based prevention programs are the most successful in reducing eating disorder risk factors, however, such programs might not be convenient for students limited by scheduling restraints or geographic proximity. Further, some students may be reluctant to attend such groups due to lack of anonymity. One way to address these potential barriers is to adapt dissonance-based programs for online use. However, no extant studies have examined the feasibility of this mode of delivery for dissonance-based programs. The current study examined the effectiveness of an online dissonance-based program, and compared it with traditional face-to-face delivery and assessment-only control conditions. It was hypothesized that: 1) online and face-to-face dissonance programs would produce comparable results; and 2) both of these active treatments would yield improvements in eating disorder outcomes (e.g. reduced thin ideal internalization, body dissatisfaction, dieting, negative affect, and eating disorder symptoms) compared with an assessment-only control condition. Results partially supported the original hypotheses. Modified intent-to-treat analyses (MITT) indicated that participants in both the face-to-face and online intervention groups showed less body dissatisfaction at post-intervention assessment compared to assessment only participants. Further, when analyses were conducted using a non-intent-to-treat (non-ITT) approach (examining only the outcomes of participants who completed the intervention), significant post-intervention differences were observed for all outcome variables. Specifically, individuals in both intervention groups showed lower thin-ideal internalization, body dissatisfaction, restraint, negative affect, and fewer eating disorder symptoms compared to assessment only participants. This study indicates that there may be some promise in adapting dissonance-based eating disorder prevention programs for online use. Future studies should continue to refine online adaptations of such programs and examine the effects of such programs with different populations.