SCIENTIFIC Representative longitudinal studies of older adults are important for identifying risk factors for adverse mental and physical health outcomes. The Chicago Health, Aging, and Social Relations Study (CHASRS) is a 10-wave longitudinal, population-based study of 229 Caucasians, African American, and Hispanic men and women who ranged from 50 to 68 years of age at baseline. We describe the sample, design, methods, and measures used in CHASRS to promote dissemination and secondary analysis of de-identified data, and we identify characteristics of respondents that predicted attrition to permit evaluation of potential selection biases. Participation required respondents to spend one day per wave to complete interviews, surveys, and physiological testing in a university laboratory. Analyses of baseline measures of demographic factors, health, cognitive function, loneliness, and social contact indicated that attrition was higher for older respondents and lower for respondents with better cognitive function and more social contacts. These results, which were observed in an ethnically diverse sample, replicate and extend prior studies of the predictors of attrition. (PsycINFO Database Record (c) 2018 APA, all rights reserved)