The effect of distributed questioning with varied examples on exam performance on inference questions

The purpose of this study was to investigate whether the distributed presentation of different versions of a question would produce better performance on a new version of the question than distributed presentation of the same version of the question. A total of 48 four question sets of five alternative multiple-choice questions were presented during a college lecture course. The answers to all the four questions in each set required an inference from the same fact statement. One question in each set appeared on both a block exam and the final. Each of the other three questions was presented either on an online quiz before class, during class or on an online quiz a week after class. Answering different versions of a question before class, during class and after class resulted in better performance on a fourth version of the question on the exam and final than answering the same question before class, during class and after class. On the final exam, 44 questions were answered twice, first under a time constraint of 90 seconds and then when response time was unlimited. The two responses were virtually identical.