This study examined pretreatment factors associated with attrition from a clinical trial for panic disorder. The study group consisted of 162 patients who began 11-visit treatments. Six domains (demography, panic disorder severity, psychiatric comorbidity, illness/treatment attributions, coping styles, and personality styles) with 52 variables were used to predict attrition. One hundred twenty-two patients completed and 40 dropped out from treatment. Final multivariate regression analyses showed that the following two variables were independently associated with attrition: lower household income and negative treatment attitudes; attributing the panic disorder to life stressors and greater age were independently associated with attrition at the trend level. Preliminary analyses suggested, in addition, associations between attrition and lower education, shorter length of prior treatment, higher anxiety sensitivity, lower agoraphobic avoidance, and a coping style of seeking social support that were not confirmed by best predictor analysis. Psychiatric comorbidity and personality styles were unrelated to attrition. The implications of these findings for future research and clinical practice are discussed.