A coaching strategy to decrease errors in swimming strokes with swimmers who had not improved under "standard" coaching procedures was investigated using a multiple baseline design across subjects and swimming strokes. The procedure resulted in a large decrease in errors on swimming strokes during sessions in a training pool. Stimulus generalization of improved performance to normal practice conditions in the regular pool was observed with all but one swimmer. This improvement was maintained during two maintenance phases lasting approximately 2 weeks, as well as under standard coaching conditions during at least a 2-week follow-up. For two swimmers, error rates on one of the strokes showed a gradual increase between the third and fifth week of follow-up, but brief remedial prompting sessions immediately corrected their performance. Some beneficial response generalization to other components of the stroke being trained was observed, but no improvements were found on untrained strokes. The error correction package did not disrupt practice, require excessive amounts of the coach's time, or necessitate the use of cumbersome apparatus. In addition, the coach and the swimmers considered the procedures to be effective, and expressed their willingness to participate in them again in the future.DESCRIPTORS: swimming, sports skills, behavioral coaching, stimulus generalization, response generalization During the last decade, behavioral techniques have begun to influence the area of sports and physical education. Books in this area have been We express our appreciation to Don Wells, Head Coach, and the coaches and swimmers of the Manitoba Marlin Swim Club for their excellent cooperation during this research. A special thanks to Roger LePage, the coach who conducted the standard coaching and the error correction procedures, for cheerfully supporting this research and making many excellent suggestions throughout the project. We are also grateful to Coach Diane Backman for administering the social validation interviews to the subjects.