The effect of post-passage question generation on comprehension question performance was examined. Students who were taught to generate "think"-type questions for reading comprehension performed better than students who were told to read and reread a story for comprehension on higher order inferential comprehension test items, but not on literal items. (Author/CB)