Reinforcement schedules in habit reversal—a confirmation.

Abstract

To investigate partial versus continuous reinforcement in a habit-reversal situation, two groups of rats, one partially reinforced and the other continuously reinforced, were trained to make a simple white-black discrimination. In this phase, the white choice was positive and the black negative, with a correction procedure. After 14 days of six trials per day, reversal training was given for ten days with six trials per day, with the black positive. The mean number of correct reversal responses per day for the ten-day period for the continuously reinforced group was 3.47; in contrast, the mean number of reversal responses for the partially reinforced group was 2.32. The difference is significant at the .01 level of confidence. Results confirm recent research which indicates that the effects of partial reinforcement are apparent in reversal situations as well as in the traditional extinction procedure; they also suggest that the retraining or habit-reversal situation may be a more sensitive and appropriate procedure for comparing response strengths acquired under partial or continuous reinforcement than conventional extinction procedures.