A study of misbehavior: token reinforcement in the rat

The purpose of this research was to investigate the phenomenon of misbehavior described by Breland and Breland (1961). Rats were trained to obtain ball-bearings and drop them in a hole for food or water reinforcers. In confirmation of the Brelands' observation, many subjects were slow to deliver the balls, and frequently attempted to chew them before they were dropped. A series of four experiments, in which the same rats were used throughout, showed that delivery times tended to be longer with food than with water, and that these times increased when nylon balls were substituted. The effect of motivational level was investigated by varying both deprivation and amount of prefeeding; no effect on delivery time was detected, although other measures of performance were affected by motivational factors. Similar results were obtained in a final experiment that employed a new set of naive subjects. The studies demonstrated that misbehavior can be studied in an experimental situation, and the results supported an analysis in terms of competition between stimulus-reinforcer and response-reinforcer contingencies. The question of why such effects have not been reported in previous token reinforcer studies was unanswered.