On the ability to predict one's own responses while learning

In a multiple-trial, paired associate procedure, subjects predicted “Yes” or “No” on each trial. Significant prediction accuracy was attributed (a) to knowledge of items correctly recalled on Trial N-1, leading to “Yes” predictions and correct recall on Trial N, and (b) to estimates of item difficulty. In Experiment 2, subjects served in triads. The Learner received 48 sentences in a four-trial, study-test paradigm, predicting his own recall on each trial for each sentence. The Listener predicted and heard the Learner's recall. The Observer predicted but did not hear the Learner's recall. The Learner had virtually no advantage in predicting his own recall compared to the Listener, with both performing better than the Observer. The results were related to “feeling-of-knowing” research, to self-responding as the basis of self-knowledge, and to the concept of privileged access.