The purpose of this study is to describe and compare the performances of students in a beginning course in educational statistics. Modified mastery learning strategies such as the use of a pretest, diagnostic tests, parallel midterms, and an extended time allowance to master the content were employed during the course. Data was available on 359 graduate students over a 2-year period. The results of this nonexperimental study support the usefulness of the pretest; show that grades improve when more diagnostic tests are taken; scores almost always increase with the second parallel midterm; increased time is accompanied by decreased performance; masters students always do better than doctoral students; and the classes in which the mastery strategies were used were about the same overall as previous classes. (Author)