A study tracked 3,192 University of Iowa freshmen through their first year and into their second year on campus. Logistic regression analyses using multiple data sources (admissions and registrar files, a standardized entrance test (the American College Testing Program Assessment) student profile section, an entering freshman survey) were conducted to determine models of student persistence at two points: freshman year spring re-enrollment and sophomore year fall re-enrollment. Two relatively successful models for predicting sophomore persistence were derived, although both models predicted correctly rather small percentages of non-persisters. A followup study on the reasons given for withdrawal by false-positives in the model is recommended. For this sample, it was found that modeling persistence was related to college-level academic indicators; the only non-academic factor to enter the prediction equation was students' perceived need for financial aid. (Contains 17 references.) (MSE)