Thirty-six reading specialists and classroom teachers enrolled in a graduate reading diagnosis course participated in a study to determine whether clinical performance could be improved by alterations in a clinician's memory and strategy. Overall procedures in this study included pretests, 30 hours of clinical training in diagnosis, and posttests. During the training segment, the subjects were divided into four different instructional groups: group one represented a traditional approach using real children with suspected reading problems; group two used simulated cases instead of live children; group three used simulated cases as well as decision aids; group four, originally scheduled to use computer based simulations, instead used procedures similar to those of group three. Results of the study suggest that clinical memory and diagnostic performance are related, and that training can improve both. (Tabular material is included.) (AEA)