Overconfidence effects in category learning: A comparison of connectionistand exemplar memory models

Exemplar and connectionist models were compared on their ability to predict overconfidence effects in category learning data. In the standard task, pa rticipants learned to classify hypothetical patients with particular sympto m patterns into disease categories and reported confidence judgments in the form of probabilities. The connectionist model asserts that classification s and confidence are based on the strength of learned associations between symptoms and diseases. The exemplar retrieval model (ERM) proposes that peo ple learn by storing examples and that their judgments are often based on t he first example they happen to retrieve. Experiments 1 and 2 established t hat overconfidence increases when the classification step of the process is bypassed. Experiments 2 and 3 showed that a direct instruction to retrieve many exemplars reduces overconfidence. Only the ERM predicted the major qu alitative phenomena exhibited in these experiments.