A study investigated the effect of instruction about conventional expository text structure on fifth-graders' ability to learn from similarly structured social studies material. Subjects, 82 fifth graders from four heterogeneous classrooms in two schools in a small midwestern city, were assigned to either a structure training group which received direct instruction in recognizing and summarizing a conventional problem/solution text structure or to a traditional training group which read and discussed answers to questions about social studies passages. Subjects then completed three criterion tests of learning from independent reading, and one test of learning from structured discussion. Data analyses supported the major hypothesis that instruction in a problem/solution text structure, including summarization instruction, would facilitate the formation of macrostructures for text having that structure. Specifically, findings showed that, compared to the traditionally trained group, the structure trained group (1) recalled approximately 50% more of the macrostructure ideas on an essay (probed recall) test over the passage; (2) recalled approximately the same amount of information on a short answer test over specific information not necessarily included in the macrostructure; (3) wrote summaries that included more main ideas from the passages; and (4) wrote better organized summaries. However, findings did not support the hypothesis that using the problem/solution text structure as an organizational framework for classroom discussion would facilitate retention of the content discussed. (Possible reasons for this finding are examined. Four pages of references and five figures are included.) (JD)