Research in reading comprehension was reviewed in order to characterize, summarize, and evaluate its contribution to principles of instructional practice. Studies were divided into four major research traditions: existential descriptions, existential proofs, pedagogical experiments, and program evaluations. An examination of these four frameworks revealed several generalizations. For example, existential proofs comparing good and poor readers or older and younger readers established that several behaviors related to strategy use and monitoring discriminate between mature and novice or good and poor readers. On the other hand, existential descriptions of classroom practices and manual suggestions revealed little explicit teaching of either comprehension or comprehension monitoring strategies. The studies also indicated one model of instruction that emphasized guided practice, independent practice, and feedback. Using this model, teachers can train students to summarize an expository passage, ask questions about it, detect difficult portions, and make predictions about following passages, eventually assuming responsibility for monitoring these tasks themselves. (HOD)