Conditioning of Self-Reinforcing Responses: An Analogue to Self-Confidence Training

Abstract

Sixty Ss guessed words flashed on a screen. No stimuli were actually presented. When Ss thought their response was correct, they asked for a token (reward) from E. SR-positive Ss were encouraged and SR-negative Ss were discouraged for judging their responses as accurate. After training, rate of “I am correct” responses was 58% in the former and 0.2% in the latter group. On a subsequent learning task, generalization of these responses was tested. Ss learned to a low criterion; half were asked to take tokens when thinking they were correct, half were on extinction. SR-positive Ss reinforced themselves more often than SR-negative Ss. Accuracy of performance did not significantly differ among the groups.