3 experiments are reported in which Ss were asked to judge the degree of contingency between responses and outcomes. They were exposed to 60 trials on which a choice between 2 responses was followed by 1 of 2 possible outcomes. Each S judged both contingent and noncontingent problems. Some Ss actually made response choices while others simply viewed the events. Judgments were made by Ss who attempted to produce a single favorable outcome or, on the other hand, to control the occurrence of two neutral outcomes. In all conditions the amount of contingency judged was correlated with the number of successful trials, but was entirely unrelated to the actual degree of contingency. Accuracy of judgment was not improved by pretraining Ss on selected examples, even though it was possible to remove the correlation between judgment and successes by means of an appropriate selection of pretraining problems. The relation between everyday judgments of causal relations and the present experiment is considered.
Jenkins, H. M., & Ward, W. C. (1965). Judgment of contingency between responses and outcomes. Psychological Monographs: General and Applied, 79(1), 1-17.