A group of high school self-reported dropouts and a group of simulated high school dropouts were compared with graduates in their participation rates, efforts required for data collection, and the overall quality of their responses in the National Longitudinal Study of the High School Class of 1972. The results indicate that dropouts are less likely to return mail questionnaires, have lower retention rates in a longitudinal study, need more effort to trace their residence and to complete editing checks, and generally provide data of lower quality. The results also indicate that no substantial differences in data quality exist between those dropouts and graduates who were interviewed. It is thus suggested that the interview approach is preferable as far as data quality is concerned. The combined use of mail survey, extensive telephoning, and interviews would also be advisable. In either case, questionnaires should be simple and straightforward. A complicated and lengthy questionnaire would probably discourage dropouts from completing the questionnaire. (Author/EVH)