Some methodological issues in the study of conditioned reinforcement

Behavior-analytic research has developed two different procedures to condition stimuli as reinforcers. One procedure uses operant discrimination. A second procedure pairs the stimulus with a reinforcer (classical conditioning). Clinical observations suggest that the “pairing procedure” may not always be effective. In this experiment we used 4 rats to compare the “pairing procedure” with an explicit operant discrimination procedure. A previously neutral stimulus (left light) was established as a discriminative stimulus for a response that produced a reinforcer. Another previously neutral stimulus (right light) was repeatedly paired with the same reinforcer. The left and the right stimuli were alternatingly presented with the same duration throughout each session. Second, to test preferences, the rats could operate left and right levers to produce the left and right light respectively. Third, the contingencies were reversed so that presses on left lever produced right light and vice versa. The results showed a preference for responding that produced the paired light (S-paired) over responding that produced the SD. This was also evident during the reversed contingency, where the response preference also reversed. Finally, a test was conducted with the same conditions except that the S-paired was not presented. The rats showed no changes in preference during the test. In order to test if the SD and the S-paired could function as conditioned reinforcer for the establishment of novel behavior, three novel response types produced the left light, the right light, or a novel center light, respectively. The results showed no evidence that either light had become conditioned reinforcers.