In this article, an alternative for Tinto's integration theory of student persistence is proposed and tested. In the proposed theory, time available for individual study is considered a major determinant of both study duration and graduation rate of students in a particular curriculum. In this view, other activities in the curriculum, in particular lectures, constrain self-study time and therefore must have a negative impact on persistence. To test this theory, we collected study duration and graduation rate information of all--almost 14,000 students--enrolling in eight Dutch medical schools between 1989 and 1998. In addition, information was gathered regarding the timetables of each of these curricula in the particular period: lectures hours, hours spent in small-group tutorials, practicals, and time available for self-study. Structural equation modeling was used to study relations among these variables. In line with our predictions, time available for self-study was the only determinant of graduation rate and study duration. Lectures were negatively related to self-study time, negatively related to graduation rate, and positively related to study duration. The results suggest that extensive lecturing may be detrimental in higher education. However, in the curricula employing limited lecturing considerable energy was spent in supporting self-study activities of students and preventing postponement of learning. Given our findings, both activities will likely have large pay offs, in particular in curricula with low graduation rates. (Contains 4 tables and 1 footnote.)