COMPARING HIGH SCHOOLS WITH RESPECT TO STUDENT PERFORMANCE IN UNIVERSITY

Donald B. Rubin,T. W. F. Stroud
DOI: https://doi.org/10.3102/10769986002002139
1977-01-01
Journal of Educational Statistics
Abstract:This report presents data analyses investigating the relationship of high school final average, X, and first year university average, Y, in a Canadian university. The statistical summaries are (1) a matrix of observed “relative grading standards” of each high school in each matriculation year, and (2) a quadratic regression relating Y to X in each year. These summaries suggest that (1) the grading standards for some of the schools are changing consistently in time, and (2) the general relationship between Y and X is becoming more quadratic and less linear in time. The implication is that the high school attended is a more important predictor of university performance for students with poorer high school records than for students with better high school records, especially in recent years.
What problem does this paper attempt to address?