This study investigated whether the one-year retention rate for the cohort of full-time, baccalaureate-degree-seeking, first-time freshmen could be predicted from institutional and aggregate cohort characteristics, including institutional type and control, institutional size, cohort size, average entrance exam score, percentage of part-time undergraduates, percentage of cohort residing on campus, percentage of nontraditional students, and percentage of minority students. Eight different models, based on institution type, size, and geographic location, were applied to 1995 data from 230 public and private institutions. While a completely satisfactory general model did not emerge, four of the models, all based on public four-year institutions, were particularly strong predictors of student retention. Average entrance exam score, percentage of cohort residing on campus, and percentage of nontraditional students were the most consistently significant predictor variables across the eight models. Private institutions and Research I institutions produced higher retention rates than did public institutions and other Carnegie-rated institutions. Results suggest that it is not possible to evaluate institutional effectiveness in student retention without factoring in these variables. By using these models, institutions can calculate their expected retention rate and measure actual over expected retention. (Contains 55 references.) (DB)