Cross-level interaction mechanisms of exercise modality heterogeneity, self-efficacy, and social connectedness among older adults under a multi-level ecological systems framework: a psychological resource gain spiral perspective
Year: 2026
Type: article
Study design: Observational Study
Abstract: Introduction Sustained exercise participation in later life shapes psychological wellbeing in ways that single-level behavioral models capture only partially. This study takes Hobfoll's conservation-of-resources tradition as its principal frame, uses Bronfenbrenner's ecological systems scheme to organize the levels at which affordances are measured, and draws on Bandura's self-efficacy theory to specify the individual-level constructs. We ask how exercise modality heterogeneity relates to self-efficacy and social connectedness among older Chinese adults, and whether the two resources strengthen one another over time. Methods A four-stage stratified sample of 2,486 adults aged 60 and above, nested in 96 communities across 12 cities in eight provinces, was surveyed in four waves spanning 18 months; 1,927 respondents supplied complete records. Latent profile analysis, three-level hierarchical linear modeling, a random-intercept cross-lagged panel model and latent growth models were estimated, with pre-specified sensitivity analyses for the small city-level sample, for attrition, and for measurement non-invariance and non-linear time trends. Results Latent profile analysis identified four modality profiles, labeled community-square integrators, solitary walkers, home-based exercisers, and intensive group participants, which differed along sociality and venue openness rather than intensity alone. Three-level hierarchical modeling found cross-level moderation, with community peer-group density and institutional policy accessibility jointly conditioning the association between modality profile and psychological resources; because only 12 units are available at the city level, these estimates are accompanied by sensitivity analyses and should be read as provisional. The random-intercept cross-lagged panel model recovered reciprocal within-person effects whose magnitude increased across waves, and the increase survived corrections for attrition, measurement non-invariance, and non-linear time trends. Three of the four criteria fixed in advance for a gain spiral were met and the fourth only with qualification, so we describe the pattern as reciprocal strengthening with profile-specific acceleration rather than as a demonstrated spiral. Latent growth models located the steepest change between 6 and 12 months of sustained participation. Discussion The findings identify modality choice as a plausible leverage point for program design and argue for evaluation cadences longer than 6 months. Because the design is observational, none of the associations reported here should be read as a causal effect.
Source: Frontiers in Psychology
Institution: Central South University
Language: English
Topic: Physical Activity and Health
Field: Medicine
Domain: Health Sciences
Open Access status: gold
Estimated APC paid: $3,801
OpenAlex ID: W7225891694
