Studies that treat creativity as operant behavior were critically reviewed. Of the twenty studies, most met minimal requirements for methodological adequacy; all provided at least some evidence for increased creative responding. Major difficulties involved potential confounds between instructions and contingencies, lack of an independent record of the training interaction, lack of social validation data, and very limited evidence for generalization. Several issues were discussed: problems in the behavioral definition of creativity, objections to the use of contingent reinforcement, and the need for empirical analysis of the creative process.As noted by Bijou (1979) and Sloane, Endo, and Della-Piana (1980), the experimental analysis of behavior is often criticized for failure to deal with complex human phenomena such as creativity. Such a criticism is no longer tenable: Studies conducted over the past 13 years clearly address the experimental analysis ofcreativity. What remains unclear is the methodological and conceptual adequacy of these studies and whether the research has dealt with the special problems of creativity. Recent reviews by Goetz (1982) and Sloane et al. (1980) have summarized some ofthe behavioral work but have not provided a comprehensive and critical review.In this paper, we review published studies on the training of creativity as operant behavior. The behavior analytic conception of creativity will first be described, followed by a review of the methodological characteristics and findings of the empirical studies. Finally, we will discuss some ofthe difficulties raised by the analysis of such a complex, subjective, and highly valued activity as creativity. Specifically, the discussion will focus on problems in (a) the definition of a creative response, (b) the use ofcontingent reinforcement, and (c) the analysis of the creative process.The studies included in this review were