The cognitive, affective and behavioural responses of self-oriented perfectionists following successive failure on a muscular endurance task

Abstract

Recent research suggests that self-oriented perfectionism may be a positive dimension of perfectionism. However, Flett and Hewitt (2005, 2006) have argued that while this dimension may appear to have some desirable consequences, it renders those high in the disposition vulnerable to psychological and motivational difficulties when personal standards are not met. The present investigation sought to examine this assertion by comparing the cognitive, affective and behavioural responses of those reporting higher and lower self-oriented perfectionism after experiencing two successive failures on a muscular endurance task. Sixty-eight student-athletes (M age = 19.75 years, SD = 1.25 years) performed a series of cycling trials in which they failed to meet personal performance targets. Providing some support for Flett and Hewitt's assertions, findings indicated that following failure in the first trial, those higher in self-oriented perfectionism experienced a more pronounced increase in threat and reported withdrawing effort from the subsequent trial.

Keywords:

• motivation
• achievement
• sport

Acknowledgements

The authors would like to acknowledge and thank Sally Fenton and Lily Badger for their contribution to the data collection involved in this investigation.

Notes

Preliminary analysis indicated that some of the variables were not normally distributed in each group. However, because ANOVA is generally considered robust when (i) groups are equal group size, (ii) there are 20 degrees of freedom for error and (iii) non-normality is not caused by the presence of outliers (Tabachnick & Fidell, 2007), we proceeded with the intended analyses. In some instances homogeneity of variances was also not supported. In all cases, however, variance ratios were below four. Because the group sizes were equal and variance ratios are below four, the degrees of heterogeneity of variances were considered unproblematic (Myers & Well, 2003).

Prior to conducting these analyses the additional assumptions of ANCOVA were examined (linear relationship between dependent variables and covariates, homogeneity of regression slopes, reliable measurement of covariate). Correlations between socially prescribed perfectionism (covariate) and the dependent variables were small for both threat (r = .27, p <.05, and r = .23, p = .065) and satisfaction (r = −.08, p = .527 and r = −.242, p = .051) and nominal for effort (r = .07, p = .561, and r = -.09, p = .480). The homogeneity of regression slopes (independent variable × Covariate interaction) assumption was checked for each ANCOVA. None of the interactions between self-oriented perfectionism group and socially prescribed perfectionism were statistically significant indicating that the regression slopes for each group are similar. In other words, the effect of self-oriented perfectionism on the dependent variables was not dependent on the level of socially prescribed perfectionism. Finally, the measurement of the covariate was considered sufficiently reliable (α = .74).