Objective: We evaluated patterns and predictors of change from three efficacy trials of trauma-focused cognitive–behavioral treatments (TF-CBT) among service members (N = 702; mean age = 32.88; 89.4% male; 79.8% non-Hispanic/Latino). Rates of clinically significant change were also compared with other trials. Method: The trials were conducted in the same setting with identical measures. The primary outcome was symptom severity scores on the PTSD Symptom Scale—Interview Version (PSS-I; Foa, Riggs, Dancu, & Rothbaum, 1993). Results: Symptom change was best explained by baseline scores and individual slopes. TF-CBT was not associated with better slope change relative to Present-Centered Therapy, a comparison arm in 2 trials. Lower baseline scores (β = .33, p < .01) and higher ratings of treatment credibility (β = -.22, p < .01) and expectancy for change (β = -.16, p < .01) were associated with greater symptom change. Older service members also responded less well to treatment (β = .09, p < .05). Based on the Jacobson and Truax (1991) metric for clinically significant change, 31% of trial participants either recovered or improved. Conclusions: Clinicians should individually tailor treatment for service members with high baseline symptoms, older patients, and those with low levels of credibility and expectancy for change.