A THEORY OF COMPARATIVE RESPONSE-TIMES AND DIFFERENCE JUDGMENTS

In two experiments, subjects performed three tasks: First, they learned associations between names of hypothetical persons and adjectives that described them. Second, they judged "differences" in likeableness between pairs of names. Third, they pressed one of two keys to indicate which person of each pair was more (or less) likeable. Response times showed the traditional distance effect, end effect, and semantic congruity effect. A simple model was developed to describe response times and "difference" ratings, using a single scale of subjective magnitude. The second experiment applied this model to measure the likeableness of persons described by two or four adjectives. This experiment indicated that adjectives do not combine by a parallel-averaging model. Instead, "difference" ratings and comparative response times were consistent with Birnbaum's configural-weight theory, in which the weight that a stimulus receives depends on its relation to the other stimuli to be combined. In the case of impression formation, the less favorable information receives greater weight. Results are consistent with the hypothesis that ratings of "differences" and choice response times are mediated by the same scale of subjective value. © 1990.