RESUMEN
This two-wave prospective study applied the Social Influence in Sport Model to investigate whether the social influences of parents, physical education (PE) teachers, and peers were predictive of students' intention to engage in leisure-time physical activity (PA). Participants were 2,484 secondary school students (11-18 years old) who completed a questionnaire assessing positive influence, punishment, and dysfunction from the three social agents (parents, PE teachers, and peers) at baseline, and PA intention at a 1-month follow-up. Structural equation modelling (SEM) yielded excellent goodness-of-fit and consistent pathways between the three social agents. Students' leisure-time PA intention (R2 = .103 to 0.112) was positively associated with positive influence (ß = .223 to 0.236, p < .001) and punishment (ß = .214 to 0.256, p < .01), and negatively associated with dysfunction (ß = - 0.281 to -.335, p < .001). Multi-group SEM showed that the predictions were invariant between parents, PE teachers, and peers. Furthermore, no significant differences in students' gender were found between perceived social influence and PA intention. The findings supported the application of the Social Influence in Sport Model in explaining the role of significant others on students' intention to take part in leisure-time PA.
RESUMEN
Parents are often regarded as one of the significant social agents who are important to the participation of physical activity (PA) among children and adolescents. However, within the literature, the relationships between parental influences and child and adolescent PA have been inconclusive and discordant. The purpose of this meta-analysis was to quantify and synthesize the associations between parental social influences (positive parental influence, punishment, and discouragement) and the PA level of children and adolescents. Through a systematic literature search using PsycINFO, Web of Science, PubMed, ProQuest, and SPORTDiscus databases, we identified 112 eligible studies and subsequently extracted 741 effect sizes for our analysis. Multilevel meta-analysis showed that the corrected zero-order correlation of positive parental influence was positive and statistically significant, r = 0.202, SE = 0.014, t = 14.975, p < 0.001, 95% confidence interval (CI) = [0.176, 0.228]. Further moderation analysis also found that this was significantly moderated by parental gender (maternal vs. paternal), respondent of influence measure (parent-reported vs. child-reported), and type of PA measure (subjective vs. objective). The corrected zero-order correlations of negative parental influences (i.e., punishment and discouragement) were not statistically significant, and no significant moderation effects were observed. The findings of our meta-analysis showed that children and adolescents had higher PA levels when their parents supported PA participation by exerting positive social influence. Punishment and discouragement against PA by parents did not appear to be significantly associated with the PA level of children and adolescents. The findings of negative parental social influence were mixed and required further investigations.