Human subjects exposed to a particular schedule of reinforcement will frequently show response stereotypy, even at an early stage of a session (Schwartz, 1982). The stereotypy, however, will not be revealed if we focus only on the response rate of the subject, but can be revealed by topographic analysis such as response sequences, or by descriptions of the rules stated by the subjects (Heltzer & Vyse, 1994). Very little has been studied on response stereotypy itself in the analysis of human behavior.The present study examines what kind of stereotyped behavior could develop under response-dependent schedules of reinforcement, and we discuss whether the stereotyped behavior can be regarded as superstitious behavior through the investigation of its cause, that is, why and how the stereotyped behavior comes to appear.Because Skinner (1948) defined superstitious behavior in terms of behavior generated by response-independent schedules of reinforcement, subsequent studies on superstitious behavior have mainly used those contingencies (Davis & Hubbard, 1972;Eldridge, Pear, Torgrud, & Evers, 1988;Fenner, 1980;Justice & Looney, 1990;Ono, 1987;Staddon & Simmelhag;1971;Timberlake & Lucas, 1985;Wagner & Morris, 1987). Some researchers have however demonstrated that superstitious behavior can develop under other contingencies (Blue, Sherman, & Pierrel, 1971;Bruner & Revusky, 1961;Catania & Cutts, 1963;Herrnstein, 1966;Kello, Innis, & Staddon, 1975;Kieffer, 1965; Morse & Skinner, Japanese Psychological Research 1997, Volume 39, No. 4, 277-290 Response stereotypy in humans maintained by response-contingent events KOICHI ONO 1 Department of Psychology, Komazawa University, Komazawa, Setagaya-ku, Tokyo 154, JapanAbstract: The present study examined stereotyped behaviors developed during human performances that were generated by response-dependent intermittent schedules of reinforcement. Thirty university students were assigned to either fixed-interval 30-s or fixed-ratio 30-s schedules in which points or monetary reinforcers were produced only by presses on the number keys of a 41-key computer keyboard. Behavior patterns developed by all subjects were classified into four categories: optimal, random, unique, and general stereotypes. The general stereotypes category was further subdivided into five idiosyncratic types: connection, order, shift, repeat, and restriction. Analysis of the data demonstrated the role of contiguity: Whatever behavior happened to precede reinforcers was repeated even though reinforcers did not depend on that behavior. These findings support the argument that much of idiosyncratic and stereotyped human behavior is produced and maintained by contingencies of reinforcement, rather than schedule-induced or adjunctive behavior.Key words: response stereotypy, superstitious behavior, response-dependent contingency, key press, university students. 1957; Ono, 1994;Starr & Staddon, 1982;Zeiler, 1970).Concurrent superstition, in which responses in one class are maintained by reinforcers produced by concu...