Background/Purpose It is widely acknowledged that students' physical activity (PA) levels declines in middle school and PA is consistently lower among adolescent girls than boys (Cavill et al., 2001). To increase students' PA levels and achieve public health goals, there is a trend calling for social ecological approaches that incorporate multilevel factors to complement conventional research paradigms centered on individual influences on PA (Sallis & Owen, 2002). Guided by the youth physical activity promotion (YPAP) model, the primary purpose of this study was to investigate the predictive strength of individual, social, and physical environmental factors toward students' PA behaviors within and beyond physical education (PE) classes. A secondary purpose was to examine gender and grade differences in the variables.
Method Participants were 290 students (101 6th graders; 92 7th graders; 97 8th graders; M age =12.45 years, SD = .97; 71 % White) from two suburban public schools. They responded to previously validated surveys assessing their perceptions of self-efficacy, enjoyment, school social and physical environmental factors, and self-reported PA. Students' percentage of time engaged in moderate to vigorous physical activity (MVPA) in three PE classes was measured by Actical accelerometers (Mini-Mitter Co., Inc., Bend, OR).
Analysis/Results Correlation analyses revealed that, except for school physical environment, variables were significantly related to one another (r's ranging from .17 to .48, p <.01). Hierarchical multiple regression analyses revealed that self-efficacy, enjoyment, and school social environment were positive predictors of self-reported physical activity, explaining 18.6 % of the variance. Self-efficacy was the only predictor of students' percentage of time in MVPA in PE classes, accounting for only 3% of the variance. The 2 x 3 (gender x grade) MANOVA yielded a significant interaction effect (Wilks' Lambda = .88, F (12, 558) = 3.12, p < .01), the main effects for gender (Wilks' Lambda = .64, F (6, 279) = 26.55, p < .01) and grade (Wilks' Lambda = .85, F (12, 558) = 4.06, p < .01). Boys demonstrated a higher percentage of time in MVPA than girls (p < .01). Sixth graders had less time in MVPA than 7th and 8th graders (p < .01).
Conclusions These findings provided empirical evidence on relationships among multilevel correlates and students' PA behaviors, as well as grade and gender differences in MVPA. The results shed light on the importance of designing supportive school environments in order to enhance students' PA in addition to individual factors.