Sixty-six grade six students were subjects in a study to evaluate the effectiveness of a direct instruction paradigm for teaching children the reading comprehension skill of main idea identification. Both experimental groups (strategy and basal) and the control group received eight lessons of 30 minutes each. All instruction occurred over a two-week period, with posttesting immediately following instruction. The strategy group received intensive, direct instruction in main ideas that consisted of five steps: introduction, example, direct instruction, teacher-directed application, and independent practice. To compare the performance of students, basal group members were administered all the lessons on main ideas and supporting details from a current basal reader series. Students in the control group received an equivalent amount of instructional time on unrelated vocabulary development exercises. Results indicated that (1) the strategy group outperformed both the basal and the control groups in ability to recognize explicit and implicit paragraph main ideas, recognize details that support main ideas, recognize explicit and implicit passage main ideas, and compose paragraph and passage main ideas for a main idea outline; (2) the basal group outperformed the control group in ability to recognize explicit paragraph main ideas; (3) treatment groups did not differ in ability to free recall main ideas or in total free recall; (4) more capable readers consistently outperformed less capable readers on all dependent variables; and (5) there were no interactions between treatments and achievement levels. (HOD)