Fast categorization of stimuli with multivalued dimensions

Abstract

The extended generalized context model's (EGCM's) ability to account for the time course of categorization of stimuli with multivalued dimensions was tested in two experiments. In each experiment, the participants were first trained to classify stimuli (semicircles of variable size with a radial line of variable orientation) into two categories. In the subsequent transfer stage, they categorized a set of transfer stimuli. The time available on each transfer trial was manipulated. Responses had to be given within 400 msec, within 700 msec, or without time pressure. Different category structures were used in Experiments 1 and 2. The results of both experiments showed reliable effects of response deadline. The EGCM accounted for the data and performed consistently better than an alternative model.