.1. Differences in associations between objectively and subjectively measured MVPA
As previous research on psychosocial correlates of PA has relied heavily on self-report measures, this study highlights the importance of providing comparative objective measures. While accelerometry data may offer a more accurate indication of actual MVPA involvement, the objectively measured MVPA in this study showed no significant, independent correlations with any of the psychosocial variables were examined in the univariate analyses in boys. Yet, when PA was measured subjectively, each perceptual variable was significantly and positively correlated with MVPA in girls, while PA enjoyment and PA self-efficacy was significantly correlated with MVPA in boys. Since psychosocial measures are subjective in nature, those who have more positive perceptions may also perceive themselves as being more physically active than they actually are. Research has already indicated that children tend to overestimate their PA levels,25 which may account for this difference.
When looking at regression models, the models that included the four psychosocial variables as predictors of objective MVPA/total PA indicated that none of them was a significant predictor of either. These results were surprising as PA enjoyment is often considered intrinsic motivation for PA and is often reported as a correlate of PA in youth.3 The models that used the psychosocial variables to predict subjective MVPA did show that PA enjoyment and PA self-efficacy were significant predictors. While not consistent with the objectively measured PA models, the findings hold promise for PA promotion activities since enjoyment is potentially modifiable through intervention. For example, Dishman et al.9 found that a school-based intervention increased PA in high school youth and that this increase was mediated by changes in enjoyment of PA. While MVPA was measured using a well-validated self-report measure results should be interpreted with caution in light of the present findings. As suggested by both Dishman et al.9 and the results of the current study, future research should employ objective measures of PA, especially when seeking to identify mediators of intervention effects in youth.