Response Time in Computer-Aided Testing: A "Verbal Memory" Test for Routes and Maps

Abstract

Computerized testing, be it assisted or adaptive, provides not only the test scores but also the item response times. It is therefore particularly interesting to investigate what additional meaning can be gained from the item response times. This paper draws partly on the testand item-based results for 1622 young men who took part in a computerized verbal memory test. Since the early 1930’s response times have been considered as indicators of personality traits which should be differentiated from ability scores. Recent psychometric models for response times have also adopted this approach. However, the question as to whether any significant diagnostic information can be gained from CAT with detailed and programmed test-taking protocols, remains unresolved. Earlier studies show that the response times for wrong solutions is noticeably longer than for correct solutions. This was confirmed in the present study. Average item response times for wrong and right responses do not show a marked correlation with scores (.387 and .290 respectively). However, times for wrong answers correlate higher with scores than times for correct answers.