OBJECTIVE: To identify predictors of outcome of a multicomponent behavioral training program for urge and stress incontinence in women. METHODS: This report is a secondary analysis of data from three prospective, randomized, clinical trials testing behavioral interventions for urinary incontinence. Participants were a volunteer sample of 258 ambulatory, nondemented, community-dwelling women, aged 40-92 years, with stress, urge, or mixed urinary incontinence. Participants received 8 weeks (four visits) of multicomponent behavioral training in each study. The relationship between a number of variables and treatment success were explored by univariate and multivariable logistic regression analyses. RESULTS: Successful treatment of predominantly urge incontinence (75% reduction of incontinent episodes as recorded on bladder diary) was associated with not wearing any form of protection for incontinence (P = .045; 95% confidence interval [CI] .282, .987). Achieving total continence (100% reduction) was associated with fewer incontinent episodes at baseline (P < .001; 95% CI .138, .557), previous surgery for incontinence (P = .021; 95% CI 1.169, 6.543), and lower education level (P = .022; 95% CI .175, .871). Successful treatment of predominantly stress incontinence (75% reduction) was related to not having previous evaluation or treatment for incontinence (P = .001; 95% CI .026, .415), and fewer incontinent episodes on baseline bladder diary (P = .026; 95% CI .210, .907). Outcomes were not associated with age, race, type of incontinence, or a number of other variables reflecting medical history, obstetric history, medications, pelvic examination, body mass index, urodynamic parameters, or psychological distress. CONCLUSION: Aside from indicators of severity and previous treatment, there were few associations between baseline clinical variables and outcome of behavioral treatment.