Three mediational theories of anxiety and performance, namely, cue utilization theory (Easterbrook, 1959) , attentional theory (Mandler & Sarason, 1952; Wine, 1971), and working memory capacity theory (M. W. Eysenck, 1979) , were compared for their efficacy in explaining anxiety-induced performance decrements on a task of analogical reasoning. One hundred two subjects who varied in their trait and state anxiety levels completed 100 geometric analogies under either relaxed (reassurance,

more »

... limited) or stressed (ego-threat, time-limited) conditions. Response time and error rate data for nine levels of task complexity (1-, 2-, and 3-element analogies with zero, one, or two transformations per element) were analyzed by means of multivariate analysis of variance. Results in the relaxed condition supported attentional theory in that the more anxious subjects were both slower and less accurate than were the less anxious subjects. In the stressed condition, none of the three anxiety-performance theories was supported. More anxious subjects were faster but made more errors than did less anxious subjects. Thus in the stressed condition, performance differences suggested differences in speed-accuracy trade-off strategies rather than differences in processing abilities. The limitations of attentional theory and the need to study the effects of anxiety and time stress on information processing are discussed.