Short-term forecasts of the number and scholastic ability of enrolling freshmen by academic divisions

Abstract

The difference between accepted and enrolling students was modeled over a 30-week period using total number of students accepted, mean composite SAT scores, and mean high school quarter rank. Time until the application deadline was regressed onto the difference function to produce week-by-week forecasts. The enrollment yield and academic ability difference functions were collectively modeled for the university and separately modeled for each academic college. The 33 regressions were applied to cumulative values for the subsequent year, and the forecasts proved to be reasonably accurate. These short-term forecasts can be used to assess the likelihood of reaching both freshman enrollment count objectives and objectives concerning the academic ability of entering students. Using this method, undesirable trends can be identified well before application deadlines which allows recruitment efforts to be directly aimed at appropriate populations.