MOTL, R. W., and D. E. CONROY. Validity and factorial invariance of the Social Physique Anxiety Scale. Med. Sci. Sports Exerc., Vol. 32, No. 5, pp. 1007–1017, 2000.
The present study 1) tested whether the two-factor model to the 12-item Social Physique Anxiety Scale (SPAS) was substantively meaningful or a methodological artifact representing positively and negatively worded items, 2) assessed the factorial validity of the nine-item unidimensional model to the SPAS, 3) examined whether modifying the number of SPAS items would improve the factorial validity, 4) evaluated the factorial invariance of the SPAS across gender, and 5) explored the construct validity of SPAS scores.
Female (N = 146) and male (N = 166) college students (22.2 ± 4.0 yr) in lecture (N = 103) and physical activity (N = 209) courses completed the SPAS, Physical Self-Efficacy Scale (PSES), Surveillance subscale of the Objectified Body Consciousness Scale (S-OBCS), and short form of the Marlowe-Crowne Social Desirability Scale (SDS-C).
Confirmatory factor analyses (CFA) revealed that the two-factor model to the 12-item SPAS was a methodological artifact representing positively and negatively worded items. CFA indicated that the nine-item unidimensional model represented an acceptable fit to the SPAS, but it also could be improved. Modifications based on standardized residuals and item content led to the removal of two items and a seven-item unidimensional solution to the SPAS. The nine- and seven-item models demonstrated factorial invariance across gender. Correlation analyses between nine- and seven-item SPAS scores to PSES, S-OBCS, and SDS-C provided support for the construct validity.
The nine- and seven-item unidimensional models to the SPAS demonstrated evidence of factorial validity, factorial invariance, and construct validity; the two-factor model to the SPAS represented a methodological artifact.
Department of Exercise Science, The University of Georgia, Athens, GA; and Department of Kinesiology, The Pennsylvania State University, University Park, PA
Submitted for publication April 1999.
Accepted for publication September 1999.
Address for correspondence: Robert W. Motl, Department of Exercise Science, The University of Georgia, 300 River Road, Athens, GA 30602; E-mail: firstname.lastname@example.org.
1 We would like to thank an anonymous reviewer for pointing out that the magnitude of Mardia’s coefficient is positively related to sample size, suggesting that Mardia’s coefficient is not very informative with large samples.