LAUSR

Introduction: Physical inactivity has become a public health issue as it can trigger many chronic diseases. Studies have found that an individual’s social networks (SNs) influence their engagement in physical activity (PA). However, it remains unclear how the influence varies between different sociodemographic groups. This study examined the associations between the SN structures and the PA of young adults across sociodemographic groups. Method: Data on 14,595 young adults aged 24–32 were obtained from Wave IV of the National Longitudinal Study of Adolescent to Adult Health. Latent class analysis was conducted to identify heterogeneous subgroups of respondents with respect to their engagement in PA and SN structures were measured in terms of the relationship quality, contact frequency, network size, and spatial accessibility. Logistic regression and chi-square tests were used to further test the relationships between PA and SNs as well as the sociodemographic factors. Results: SNs were found to have a significant influence on PA. Relationship quality was identified as the most important feature of an individual’s SN, followed by network size, contact frequency, and accessibility. The effects of SNs on PA varied with the individuals’ gender, ethnicity, income, and educational attainment. For example, male participants tended to be physically active if they had frequent connections with their neighbors, while the engagement in PA of the female participants was strongly related to the frequency of their contact with their children. Conclusions: This research has important implications for practitioners as it suggests that PA intervention programs should be designed to target specific population groups.