Availability: Plausible But Questionable.

Abstract

The present experiments focus on the role of the availability heuristic (Tversky and Kahneman, 1973, 1974) in judgments of frequency. We argue that the data that have been advanced in support of the availability model are (a) inconclusive, and (b) have probably been misinterpreted. We present four illustrative experiments (Experiments 1 through 4), which appear to support the availability hypothesis. However, causal modeling procedures reveal that three of the four experiments actually do not support the availability model. We then propose a hierarchical to explain when availability will and will not play a role in frequency judgment. According to this model, cognitive counters automatically record frequency information when there is actual variation in the frequency of individual stimulus items. When no such variation exists, secondary judgment strategies such as availability may come into play. Experiments 5 and 6 test this hierarchical model.