Objective We conducted a controlled randomized preliminary trial of a modified dissonance‐based eating disorder program (n = 24) compared to an assessment‐only control condition (n = 23) via a longitudinal design (baseline, postintervention, 2‐month follow‐up) in a community sample of women (N = 47) with clinical (n = 22) and subclinical (n = 25) eating disorder symptoms. Method The traditional content of the Body Project, a dissonance‐based eating disorder prevention program, was modified to include verbal, written, and behavioral exercises designed to dissuade self‐objectification and maladaptive social comparison. Women with clinical and subclinical symptoms were included in the target audience to investigate both the treatment and the indicated prevention utility of the modified dissonance program. Body dissatisfaction, self‐esteem, self‐objectification, thin‐ideal internalization, maladaptive social comparison, trait anxiety, and eating disorder symptoms were evaluated in the control and the modified dissonance condition at baseline, postintervention, and 2‐month follow‐up. Results We predicted a statistically significant 2 (condition: control, modified dissonance) x 3 (time: baseline, postintervention, 2‐month follow‐up) interaction in the mixed factorial multivariate analyses of variance results. Results confirmed this hypothesis. Eating disorder risk factors and symptoms decreased significantly among participants in the modified dissonance condition at postintervention and 2‐month follow‐up compared to baseline; symptom improvement was greater among participants in the modified compared to the control condition. A secondary analysis indicated symptom improvement did not vary as a function of symptom status (clinical, subclinical), suggesting the program is efficacious in both indicated prevention and treatment applications. Conclusion Results provide preliminary support for the modified dissonance program.