In this paper we analyze longitudinal data from a nationally representative panel of college entrants to test and compare several theoretical explanations of college degree attainment and noncompletion. So far, relatively little emphasis has been placed on determining the relative and combined predictive power of competing explanations or mechanisms of college noncompletion. We utilize a methodological tool--the sheaf coefficient--to combine multiple variables into single conceptual predictors and estimate their relative effect sizes across college types. We then draw implications of these findings both for theory and for policy. (Contains 2 notes, 1 figure and 1 table.)