The effect of anxiety upon word association performance under nonstress and stress conditions

The purpose of this experiment was to investigate the relationship between anxiety and word association performance under nonstress and stress conditions with regard to three variables: commonality scores, reaction times, and faults. The procedure involved presenting two equated word lists comprised of low, moderate, and high commonality word stimuli to three groups designated low, medium, and high anxious on the basis of Neuroticism scale scores of the Eysenck Personality Inventory. The first list was presented under time pressure with a reproduction phase. Then an electrical shock apparatus was demonstrated, and the subjects were told that they would receive a shock sometime during the second word list. It was hypothesized that the moderately anxious subjects would have higher commonality scores, shorter reaction times, and fewer faults than the other two groups and that these differences would be most predominant on the medium commonality words. It was further hypothesized that under stress the low anxious group would improve their performance in terms of the three variables, the medium anxious group would do more poorly than previously, with no differences between conditions in the high anxious group. The predictions were based upon the reaction potential ceiling hypothesis, a modification of Hull-Spence drive theory. Although statistical analysis failed to confirm any of the hypotheses, several findings seemed important. Most of the significant differences were found on high commonality words. With regard to the commonality variable, there was a significant difference between groups, between conditions, and an interaction between these main effects on high commonality words. Reaction times did not differentiate groups, but significantly longer reaction times occurred on low commonality words and shorter latencies on high commonality words. In addition there was a trend toward significance between conditions at the high commonality level where reaction times decreased in all three groups under stress. Significantly more faults occurred on low commonality words than either of the other two levels, but no differences were found between groups. The major finding in this experiment was a consistent, systematic relationship between the three variables under consideration (l.e., commonality, latencies, and faults) in the anxiety groups on high commonality words. Under the stress condition, the reaction times decreased for all three groups. Low anxious subjects produced lower commonality scores and more faults. The moderate anxious group showed the same tendency to a lesser degree. The high anxious subjects improved their performance with higher commonality scores and fewer faults. An attempt was made to explicate these results from a phenomenological point of view in which different modes of perceiving and responding to stimulus cues in the environment were conceptualized as a function of anxiety level.