Objective: Most previous research based on the Theory of Planned Behaviour (TPB) to explain children's physical activity (PA) remains cross-sectional and the moderating role of demographic variables on the link between TPB variables is rarely explored. This study aims to (1) test a longitudinal cross-lagged panel model based on the TPB to explain PA in children, and (2) examine in what extent this model operates equivalently across age and gender.
Design: Children (N = 363, M = 8.6 years) completed measures of TPB constructs and PA at two measurement times (T1 and T2) within three months.
Main outcome measures: Self-reported PA.
Results: Path analyses revealed that the model tested accounted for 25% of the variance in PA at T2. Subgroups analyses revealed that age significantly moderated the links between subjective norms at T1 and T2, between PA at T1 and T2, and between PA at T1 and subjective norms at T2. Gender significantly moderated the links between intentions at T1 and T2, and between PA at T1 and perceived behavioural control at T2.
Conclusion: The temporal stability of the variables of the TPB and the link between initial PA and subsequent TPB variables may vary by children's gender and age.
Keywords: Youth; exercise; longitudinal study; moderation analyses; psychosocial theories.