Abstract
The characteristics of dropouts in a four-year longitudinal study of Mexican Americans and Anglos aged 60 years and over are investigated. Dropouts are found to be significantly older, to be less healthy, to view themselves as older, and to be less active than persons remaining in the study. Subgroups of dropouts differ from each other: deceased persons and those who could not be located were found to be generally less advantaged than restudied persons, while refusers were found to be advantaged. Implications for longitudinal studies on aging are discussed.