This study compared the achievement of high-and low-reading-ability students in two classes of a one-semester introductory college biology course after they either received and used teacher-provided questions posed at various taxonomic levels on textual reading assigned over a five-week period (Group 1) or learned via training to generate and use their own questions at the identical taxonomic levels based on the same text over the same period of time (Group 2). Independent variables included (a) form of questioning (teacher provided and self-generated), (b) reading ability (low and high), and (c) question type (referent, literal, interpretive, inferential, and self-critical). Dependent variables included scores from weekly quizzes and from a summative examination. Results indicated that (a) training students to generate and answer their own questions based on text reading had a favorable effect on their midrange (weekly quiz) performance; (b) relative to long-range (summative exam) performance, training students to generate and answer their own questions based on study reading was no more efficacious than providing students with questions based on the same text; and (c) teacher-provided questions at the literal level facilitated the acquisition of intended and incidental discrimination material better than teacher-provided questions at any other taxonomic level.From the time of Socrates, questioning has been a major (if not the major) strategy for teaching and learning. The dominant form of this strategy in classrooms has been the use of teacher questions as an integral part of instruction. "Recognizing this phenomenon, researchers for the past fifty years have quantified the number of teacher questions asked in classrooms, classified the types of questions asked by teachers, and have examined the effectiveness of different types of questions on student behavior" (Evenson, 1981, p.