On the Testability of the Availability Heuristic

TVERSKY and KAHNEMAN's availability heuristic, although originally intended to account for only frequency and probability judgments, has been used to explain almost all kinds of social judgments. Accordingly, the process of judgment formation is mediated by the availability of memorized information, that is, by the ease with which relevant material can be recalled at the time when the judgment is made. Recall operations - either pure retrieval or reconstructive recall - are regarded as the determining subprocess within the process of judgment formation. In this article, empirical evidence is presented which is hardly compatible with such an account. A reaction time experiment, on the egocentric attribution phenomenon is described suggesting that Ss' claim to contribute more to various social activities than their partner is not caused by Ss' tendency to predominantly recall examples of their own activities. Within-judge correlations of recall and judgment latencies are rather low., and an analysis of the facilitating effect of prior judgments on subsequent recall latencies (for the same issues) does not reveal the kind of priming effect that would be expected if recall operations were already involved in the preceding judgments. Negative evidence from some other experiments on illusory correlations is also mentioned. These results are discussed in the context of methodological problems inherent in testing so-called judgmental heuristics.