Project TALENT, the Equality of Educational Opportunity Survey, and the Base-Year Survey of the National Longitudinal Study tested national probability samples of American high school seniors in 1960, 1965, and 1972. This study is concerned with the feasibility of using these data for studying trends in educational attainment. Three separate aspects of the problem were considered. First, a detailed review of sample selection, weighting, and school participation showed that the data would provide a sound basis for comparing performance, and identified certain implications of the existing data for making and interpreting comparisons. Second, comparisons were made of students' high school experiences, their educational and occupational aspirations and plans, and their parents' education and occupation. The pattern of results for the three survey samples, although obscured by differences in questions and by population trends, indicated that comparisons of test performances of the groups would be warranted. Third, a detailed design was prepared for studying the equivalence of selected tests used in the three surveys and for equating test scores on pairs of equivalent tests. Application of these equating results to the test results of the three surveys would provide the desired trend data. (Author)