Résumé
Regular physical activity (PA) offers substantial health benefits for older adults, enhancing longevity and quality of life. Social support may encourage PA engagement, while PA itself can create opportunities to foster social ties. However, longitudinal evidence on these reciprocal relationships is limited. This study used data from the Canadian Longitudinal Study on Aging (CLSA) to examine whether (1) nine social support factors predict PA over time, (2) PA predicts these social factors, and (3) these associations vary by demographic characteristics. The sample included 18,110 adults aged 65 and older (M = 73.02, SD = 5.67), assessed at two time points approximately three years apart. A cross-lagged path model tested prospective relationships, and invariance testing assessed moderation by sex, sexual orientation, living alone, country of birth, caregiving status, and income. Model fit was excellent (CFI = .97, SRMR = .18, RMSEA = .03). Social participation (β = .034), social networks (β = .061), affectionate support (β = .026), and social contact frequency (β = .017) significantly and positively predicted future PA. Conversely, PA significantly and positively predicted social participation (β = .042), social networks (β = .042), social contact frequency (β = .017), and domestic partnership status (β = .022), and significantly negatively predicted online communication (β = –.019). However, no moderation effects were found, suggesting these associations were consistent across demographic groups. These findings highlight the bidirectional relationship between PA and social connections in older adults and underscore the importance of fostering social networks to support healthy aging.
