A study evaluated the effectiveness of a direct instructional paradigm for teaching children to comprehend main ideas of prose passages. Subjects, 69 children in three fourth grade classes, were randomly assigned to an experimental "strategy" group receiving direct instruction in main ideas, an experimental "basal" group receiving massed basal reader instruction in main ideas, and a control group receiving vocabulary development exercises. All groups received eight 54-minute lessons over a period of three weeks; posttests were administered in the fourth week. The strategy group lessons consisted of a sequence of main skills progressing from main ideas in groups of words to main ideas in passages. Lessons in the basal group were taken directly from a current basal reader series. Results of the posttests revealed no significant differences between the three groups' skills at comprehending main ideas. Given the proven power of direct instruction, differences in the intelligence of the groups and discipline problems with the strategy group may explain these unexpected results. (JL)