Effect of information feedback upon intertrial consistency of time judgment

Male college students made verbal estimations of time intervals ranging from five to 80 sec on six blocks of five trials. Some subjects were given information feedback (IF) about judgment errors after each trial and the remaining subjects were given no IF. Judgments of the no-IF group changed from initial overestimations to final underestimations of the intervals across blocks of trials. Judgments of the IF group did not significantly change across blocks. In the no-IF group, each block of trials correlated strongest with immediately adjacent blocks; i.e., the correlation matrix had a superdiagonal form. Interblock correlations were weakened by IF, and the superdiagonal form was not obtained in the matrix of the IF group. Previous theories of the superdiagonal form are considered inadequate to explain these results. An alternative hypothesis, based on changing decision criteria with blocks, which IF attenuates, is proposed to explain these data.