A Feature-Salience Analogue of the Inverse Base-rate Effect

A Feature-Salience Analogue of the Inverse Base-rate Effect

Classification learning requires integrating many properties of the items being learned including the base-rate probability that a category will occur as well as the salience of features. Previous research has demonstrated an inverse base-rate effect, in which people classify an item that has features predictive of both a high base-rate and low base-rate category into the rarer category. We suggest that this finding reflects that feature salience plays a greater role in classification than does base-rates. We tested this hypothesis by demonstrating that manipulations of feature salience determine the classification of ambiguous stimuli regardless of the underlying base-rates of the categories.