Eighty-seven students took part in a study (using three experimental conditions) of the effect of question generation, question answering, and rereading on an immediate posttest and on a delayed test of knowledge of a text passage. In condition G, subjects generated multiple-choice questions that would test knowledge of each of the six content areas of the passage; in condition A, subjects responded to multiple-choice questions in the six areas before checking and correcting their answers; and, in condition R, they reread the passage until they felt they understood the six main points. Results indicated that the question-generating group spent significantly more time on the task than did the other two treatment groups, but there were no significant differences between treatments on the posttest or the follow-up test, with or without covariance adjustments for differences in time spent. (Author/AA)