A third followup survey within a National Longitudinal Study (NLS) of the high school class of 1972 finds that there is little difference in group variability in survey response when considering background characteristics, while significant differences in response rates are found when considering the group patterns of response to prior instruments. Prior response is given as the most important predictor of third followup response. Meant to measure postsecondary educational and vocational activities, the survey is intended to guide federal planning and implemention of educational and career programs. The scope and purpose and descriptive statistics are presented analyzing response rates, response modes and their quality. The prediction approaches and results of the variables; a discriminate function analysis of mail and personal interview modes; and the quality of data are discussed, referring to the importance of a response payment incentive procedure. The study is meant to facilitate future data collection efforts. Appendices and tables illustrate the analysis. (CM)