A primer and empirical demonstration of network analysis for physical activity attributes

Purpose: A gold-standard measure for physical activity does not exist, partly because it is a latent system emerging from interactions between its attributes. This study aimed to validate a measurement model for physical activity using a systems approach, specifically a network model, assess its gender invariance, and examine its relationship with a psychological network. Methods: Forty healthy participants (18 males, 22 females; mean age = 27.9 years, 95% CI = 25.8–29.9) were recruited online. Physical activity was measured via self-report and smartphone accelerometry over one week. Three models were explored: (1) relationships between physical activity attributes in a holistic framework, (2) gender invariance, and (3) interactions with a psychological network (self-efficacy). Results: Centrality analyses identified key attributes, such as self-reported moderate activity, time spent sitting, and self-efficacy when access to facilities was limited. Conclusions: Findings support a systems-based measurement approach, highlighting the complex, dynamic nature of physical activity.