Models of the effects of prior knowledge on category learning.

When people learn about a new category, they are influenced by prior knowledge of other categories. In 5 experiments, subjects made categorization judgments after observing descriptions of persons from a location referred to as City W. In these experiments, prior knowledge as well as observations within City W were manipulated. The integration, weighting, and distortion models of categorization explain prior knowledge effects in different ways. The integration model, which assumes that categorization is influenced by prior examples from other categories, predicted the results of the experiments. It was found that the effect of prior knowledge was independent of the observed proportion of category membership in City W, that the prior knowledge effect was diminished with more observations, and that learning about City W affected subjects' judgments about the general population. The weighting and distortion models could not account for all of the results.