Studies of distributed practice. XI. An attempt to resolve conflicting facts on retention of serial nonsense lists.

CO In 1940 Hcvlsnd (3) published a it in which retention of serial nonsense lists was measured after 2 min., 10 min., and 24 hr. In one case lists were learned by biassed practice (6-sec. intertrial interval) and in another by distributed practice (2min. intertrial interval). The results clearly showed that retention as measured by both recall and releaming was better following distributed practice than following massed practice. We have reported a similar experiment (5) in which retention after 24 hr. was much better following massed practice than following distributed practice. Since the two studies are in direct conflict, it is important that attempts be made to determine the source of this conflict. The most apparent difference between Hovland's experiment and ours was the difference in length of intertrial rests defining distributed practice. The longest intertrial rest we used was 1 min., as compared with the 2-min. interval used by Hovfecd. Consequently, in one of the experiments to be reported we have used 60, 90, and 120 sec. as intertrial rents. If length of rest interval is the critical difference between the two studies, we would expect sharp increments in retention between 60 and 120 sec. The second experiment to be re-