Re-Examination of the Factor Structure and Composition of the Self-Presentation in Exercise Questionnaire (SPEQ)

Abstract

The purpose of the present study was to re-examine the factorial integrity of the 11-item two factor Self-Presentation in Exercise Questionnaire (SPEQ; Conroy, Motl, & Hall, 2000). Initial attempts to confirm the factor structure and composition in a Canadian university sample were unsuccessful. Subsequent scale reconstitution yielded an 8-item model, which retained the two original factors labeled impression motivation and impression construction. Cross-validation of the reconstituted model in both a Canadian and a New Zealand sample produced an adequate approximation of the data, but conceptual problems with both the impression construction and impression motivation subscales were identified. Specifically, the SPEQ items tended to focus too much on physical appearance as both a self-presentational motive and behavior. Several items also confounded these two constructs. Thus, although the SPEQ represents an important first step in investigating the role of self-presentation in exercise behavior, further work must continue in order to develop a psychometrically sound instrument.

Notes

1 The 11 items were collected as part of the original 20-item version of the SPEQ. Hence, it is possible that completion of the other 9 items may have influenced the answering of the 11 items. We doubt, however, that our methodology can account for the extremely poor fit of the data.

2 High multivariate kurtosis was indicated for the 11-item model (Mardia's normalization estimate = 16.05). Observations farthest from the centroid (Mahalanobis distance) were therefore eliminated and the ML estimator was once again applied to the data. Fit indices paralleled those reported in text.

3 Correlation and variance-covariance matrices can be obtained from the lead author for all data sets.

4 Factor loadings, standard errors, and squared multiple correlations data for the New Zealand sample can be obtained from the lead author.