Attrition may pose a threat to the validity of a study, specifically in lengthy longitudinal studies in which retaining participants may be more difficult. The purpose of the present study was to determine, using retrospective data (n = 415), which variables (i.e., gender, ethnicity, level of education, primary language, alcohol dependence, and employment) may increase the probability of attrition in a study on depression among depressed ethnic minority patients seen in a public safety net primary care setting. Of the 6 variables explored, there were no specific hypotheses for gender, ethnicity, primary language, or employment. These variables were explorative. The following were specific hypotheses for the remaining 2 variables: (a) participants with more education would be less likely to drop out of the study, and (b) participants at high risk of alcohol abuse would be more likely to drop out of the study. Pearson chi-square analyses showed significant results only for ethnicity: participants who self-identified as Black/African American were the most likely to drop out at 33.7% (n = 29), whereas White participants were the least likely to drop out at 17.2% (n = 5). The 2 specific hypotheses were therefore not supported. Limitations and implications of the present study are discussed.