Concurrent versus retrospective assessment in test anxiety research.

Abstract

This study investigated whether retrospective and concurrent assessments in test anxiety research yield comparable data. A 3 X 2 X 2 (Test Anxiety Level X Type of Assessment X Final Exam Point) design was used. Results indicated that concurrent and retrospective assessment groups did not differ on number of positive thoughts, negative thoughts, bodily sensations, or anxiety levels. Concurrent assessment did not affect test performance. Additional analyses revealed a few significant differences on rank-ordered positive and negative thoughts for the concurrent and retrospective groups assessed at the beginning but not the middle of the test. Other results generally replicated those of a previous study. Progress in cognitive behavior therapy is linked to progress in assessment. However, developments in assessment have lagged behind therapeutic developments (Kendall & Korgeski, 1979), and the need for sound assessment procedures is clear. This study investigated a fundamental assessment problem. Using an actual test, the objective was to determine if retrospective assessment of cognitions yields data that arc comparable to those obtained by interrupting the subject (concurrent assessment) during the test. Cacioppo and Petty (1981) have called attention to the limited evidence about whether cognitive responses differ as a function of when they are obtained. Previous research (Galassi, Frierson, & Sharer, 1981) revealed that negative cognitions and bodily sensations indicative of arousal accumulate toward the end of a test. Because related studies (Deffenbachcr, 1978; Holroyd, Westbrook, Wolf, & Badhorn, 1978) investigated cognitions retrospectively, it seemed possible that their frequency had been overemphasized. In general, retrospective assessment offers the advantage of not interrupting the task (Cacioppo & Petty, 1981;