Models for Speed and Time-Limit Tests

Abstract

There is a long tradition in the ability testing literature, going back at least to Spearman (1927) and Thorndike et al. (1927), that response speed is as much an indicator of ability as is the correctness of responses to items of increasing difficulty (Berger, 1982; Eysenck, 1982; Brand and Dreary, 1982). Some tests are pure speed tests (consisting of items that are virtually always correctly solved, and where completion time is the recorded variable), others are pure power tests (consisting of items of increasing difficulty, administered without time limit, where the number of correct responses is the recorded variable), but most tests are a mixture: They consist of items of varying difficulty and are administered with a time limit.