The purpose of this paper is twofold. First, the development of psychometrically sound process indicators of student performance is described. Second, the results are reported of a test of the utility of these indicators with six samples of men and women students (n = 911 in each group) from 19 baccalaureate, 27 master's, and 29 doctoral-granting institutions. Process indicators represent behaviors associated with desired outcomes of college and estimate the extent to which students are engaged in these activities. Selected items from the College Student Experiences Questionnaire were used to create measures of three good educational practices: faculty-student contact, cooperation among students, and active learning. Active learning and cooperation among students in that order were the best predictors of gains for both women and men at all three types of institutions. Student background characteristics had only trivial influences on educational gains. Implications are discussed for institutional policy and further research.