Task and Environmental Change SelfEfficacy for Physical Activity Scale: Analysis of Factorial Invariance Across Gender

Karteroliotis, Konstantinos1; Dzewaltowski, David A.2; Gyurcsik, Nancy C.2; Estabrooks, Paul A.3; Johnson, Judy A.4

Medicine & Science in Sports & Exercise: May 2004 - Volume 36 - Issue 5 - p S62
Annual Meeting Abstracts: B-26 - Free Communication/Poster: Determinants of Physical Activity

1University of Athens, Department of Physical Education & Sport Science, Athens, Greece.

2Kansas State University, Community Health Institute, Department of Kinesiology, Manhattan, KS.

3Kaiser Permanente, Clinical Research Unit, Denver, CO.

4University of Kansas, School of Medicine, Wichita, KS.

Email: ckarter@cc.uoa.gr

(Sponsor: Thomas Barstow, FACSM)


Despite growing research into the role of self-efficacy in the prediction of physical activity in youth, minimal research attention has been given to illuminating components of this global construct. PURPOSE: The study examined the factorial validity of task and environmental change of self-efficacy related to physical activity in a sample of sixth grade youth. The factor structure of the measure was examined by testing its factorial invariance and latent mean structure across gender. METHODS: Participant were 2039 sixth graders (976 boys and 1063 girls), aged 11–14 years, enrolled in 16 middle schools in rural and urban areas of a Midwestern state. Informed consent from students and their legal guardians were obtained prior to data collection. Conflrmatory factor analyses were performed using full-information maximal likelihood (FILM) estimation in AMOS 5.0. RESULTS: Single group analyses indicated an acceptable four-factor structure model for boys and girls. The four-factor structure composed of self-efficacy for moderate to vigorous physical activity (MVPA), vigorous physical activity (VPA), home environmental change (HEC), and school environmental change (SEC). Multigroup analysis demonstrated that factor structure, factor loadings, and unique variances for the scale was invariant across gender. The analysis of latent mean structure indicated that girls exhibited significantly different latent mean scores on VPA (ê = −0.31, SE = 0.06, p < 0.01) and HEC (ê = 0.21, SE = 0.05, p < 0.01) measures. CONCLUSION: Generally, the study provided evidence for the appropriateness of a four-factor structure of task and environmental change selfefficacy for physical activity. Further work is needed to validate its structure across ethnicity, an area in which empirical research is lacking.

©2004The American College of Sports Medicine