A randomized trial of 265 consenting students was conducted within an introductory biostatistics course: 69 received eight small group cooperative learning sessions; 97 accessed internet learning sessions; 96 received no intervention. Effect on examination score (95% CI) was assessed by intent-to-treat analysis and by incorporating reported participation. No difference was found by intent-to-treat analysis. After incorporating reported participation, adjusted average improvement was 1.7 points (-1.8, 5.2) per cooperative session and 2.1 points (-1.4, 5.5) per internet session after one examination. After four examinations, adjusted average improvement for four study sessions was 5.3 points (0.4, 10.3) per examination for cooperative learning and 8.1 points (3.0, 13.2) for internet learning. Consistent participation in active learning may improve understanding beyond the traditional classroom. (Contains 3 tables and 2 figures.)