The concept of "practice makes perfect" was examined in this work in the context of effective learning. Specifically, we wanted to know how much practice was needed for students to demonstrate mastery of learning outcomes. Student learning patterns in two different university courses that use a similar education approach involving problem based learning, outcomes based assessment, and problem sequencing based on elaboration theory were examined. Learning outcomes for each course were explicitly defined and students were repeatedly assessed through sequential assignments. The cumulative proportion of criteria successfully demonstrated for each problem-solving attempt was determined using data retrospectively obtained from instructor grading records. Learning followed a typical growth pattern--it increased rapidly at first and more slowly with succeeding attempts. The precise shape of the learning curve differed between the two courses and is thought to be the result of problem difficulty and problem sequencing. Depending on these two factors, at least one more attempt than the number of times criteria need to be demonstrated is required and often more are needed to demonstrate mastery. This paper presents class-level data and future work should investigate individual performance and particularly why some students learn more quickly than others. Two additional issues for future consideration are the effect of the number of attempts on long-term retention and on the transferability of the learning to other problems. (Contains 4 tables and 2 figures.)