Background/Purpose: Despite well-documented health benefits of
regular physical activity (PA), many middle school students do not participate
in sufficient PA necessary to accrue the health benefits associated with PA
(USDHHS, 2008). Given the fact that a physically active lifestyle during
adolescence may track into adulthood, it is important to identify multilevel
theory-based correlates of PA (Sallis et al., 2000). Based on the social
ecological model (Sallis & Owen, 2002), multilevel correlates, such as
self-efficacy, school and home physical environments, social support from
parents, physical education (PE) teachers, and friends, have been consistently
cited to predict individuals' PA. However, researchers have not investigated
the direct and indirect relationships among these variables with middle school
students. This study attempted to fill this gap by testing a hypothesized
model.
Method: Participants were 539 students (189 boys, 350 girls; Mage
= 12.4 years) from two southeastern public schools. They completed previously
validated surveys assessing their perceptions of school and home physical
environments, social support from parents, PE teachers, and friends,
self-efficacy, and PA during their regular PE classes. The hypothesized model
was tested using path analysis (AMOS 16.0; Arbuckle, 2007).
Analysis/Results: The path analyses indicated that the hypothesized
model demonstrated a good fit to the data according to the various indices of
fit: ¦Ö² (3) = 4.0, p = .257; Comparative Fit Index (CFI) = .99;
Incremental Fit Index (IFI) = .99; Normed Fit Index (NFI) = .99; Root Mean
Square Errors of Approximation (RMSEA) = .023 (Hu & Bentler, 1999). The
model accounted for 23.6 % and 31.5 % of the variance in self-efficacy and PA,
respectively. Path coefficients suggested that school physical environments,
parents' support, and friends' support (b = .15, .25, .24; respectively) were
directly and positively associated with self-efficacy. Additionally, home
physical environments, PE teachers' support, and self-efficacy (b = .09, .13,
.27; respectively) were directly and positively related to PA. The effect of
the parents' support and friends' support on PA was mediated partially through
self-efficacy or otherwise was a direct effect (b = .17, .17; respectively). No
direct effect was found between school physical environments and students' PA.
Conclusions: The findings of this study provide empirical evidence
regarding the interactions among these social ecological variables. The results
highlight the importance of supportive physical and social environments on
students' self-efficacy and PA. These findings have significant implications
for PA interventions aimed at promoting middle school students' PA.