This study examined the effects of self-questioning (question-generation) on reading comprehension and on self-evaluation of comprehension. Seventh-grade students were assigned to one of 4 treatment groups : a question-generation group (Gr. G), an answering questions generated by an experimenter group (Gr. A1)(Exp. 1), an answering questions generated by Gr. G group (Gr. A2) (Exp. 2), or a read-reread control group (Gr. C). Verbal ability, as measured by the Siba Vocabulary Test, was used to group Ss into 3 levels. The quality of questions generated by Gr. G and task performances were analyzed in terms of comprehension of macrostructure. The major results were as follows. a) Question-generation facilitated the comprehension of main ideas. In particular, this effect was larger for lower than for higher verbal ability students. Such effect was caused not by the contents of questions generated, but from the process of generating those questions. b) Gr. G seemed to evaluate more adequately on their comprehension though without any significant results. To examine such result, more valid measures should be planned in the future.