The efficacy of individual components of an online course on positive course outcome was examined via stepwise multiple regression analysis. Outcome was measured as the student's total score on all exams given during the course. The predictors were page hits, discussion posts, and discussion reads. The vast majority of the variance of outcome was accounted for by total page hits. Participation in discussion had little to no effect on performance as measured by outcome on exams. The results were double cross-validated with a sample chosen from another class. There was no shrinkage, indicating that the equations derived from the two samples were very reliable.