What matters when children play: Influence of Social Cognitive Theory and perceived environment on levels of physical activity among elementary-aged youth

•Perceived environment variables contributed little to the models.•Sex, self-efficacy, and social support were most predictive of physical activity.•Sex was more strongly associated with physical activity among Hispanic students.•Self-efficacy was more strongly associated among non-Hispanic students.

ObjectivesSocial Cognitive Theory (SCT) has often been used as a guide to predict and modify physical activity (PA) behavior. We assessed the ability of commonly investigated SCT variables and perceived school environment variables to predict PA among elementary students. We also examined differences in influences between Hispanic and non-Hispanic students.DesignThis analysis used baseline data collected from eight schools who participated in a four-year study of a combined school-day curriculum and environmental intervention.MethodsData were collected from 393 students. A 3-step linear regression was used to measure associations between PA level, SCT variables (self-efficacy, social support, enjoyment), and perceived environment variables (schoolyard structures, condition, equipment/supervision). Logistic regression assessed associations between variables and whether students met PA recommendations.ResultsSchool and sex explained 6% of the moderate-to-vigorous PA models' variation. SCT variables explained an additional 15% of the models' variation, with much of the model's predictive ability coming from self-efficacy and social support. Sex was more strongly associated with PA level among Hispanic students, while self-efficacy was more strongly associated among non-Hispanic students. Perceived environment variables contributed little to the models.ConclusionsOur findings add to the literature on the influences of PA among elementary-aged students. The differences seen in the influence of sex and self-efficacy among non-Hispanic and Hispanic students suggests these are areas where PA interventions could be tailored to improve efficacy. Additional research is needed to understand if different measures of perceived environment or perceptions at different ages may better predict PA.