Influences of anxiety on golf performance: A field test of catastrophe theory

The purpose of this study was to test the basic tenets of Fazey and Hardy's (1988) catastrophe model. Specifically, the purpose was to examine the interaction effects of cognitive anxiety (worry) and physiological arousal (activation) on golf performance. Previous research has had difficulty in assessing the basic constructs of the catastrophe because of the lack of independence between arousal and cognitive anxiety. Four amateur golfers were tested using the Sport Grid-Revised (Ward & Cox, 2001) in competitive play. The Sport Grid-Revised is the most reliable measurement inventory available for catastrophe theory assessment. Physiological arousal and cognitive anxiety ratings as well as performance measures were collected prior to each golf shot taken in four competitive rounds for each of the participants. Multiple regression procedures were utilized to analyze the data.;In the linear analyses of all participants, it can be observed that Wor has a significant negative effect on golf performance. This would lead to support of Martens' et al. (1990) multidimensional theory. Also, in the linear analyses, it is noticed that activation is a positive or neutral effect for the higher skilled golfers. Participant 1's (male, high skill) activation score yielded a positive beta and Participant 3's (female, high skill) activation score yielded a very slight negative beta. A better skilled golfer is able to cope with increased activation (Spence, 1956).;No evidence for catastrophe theory was observed in the analyses of the full model of all participants. In order for catastrophe theory to be supported, a significant quadratic relationship between activation (Act) and performance must be observed and a significant interaction between the variables, Act and Wor (Act/Wor).;In all of the participants, increased worry (Wor) is associated with a decrement in performance and activation is associated with a curvilinear relationship to performance in all participants, except participant 3 (female, high skill). In only one case (female, low skill) was the interaction variable (Act/Wor) significant. In order for the catastrophe theory to be verified the interaction variable must be significant and the quadratic of activation (Act2) must be significant. In this case, an interaction is observed in Figure 10. But, in Figure 10 you would expect a drop in performance on the high activation line as worry increases. Instead, we see an increase in performance. There is a drop in performance on the low activation line as worry increases, but that is predicted in the multidimensional theory of anxiety (Martens et al., 1990).