Structural constraints and age differentiation: an empirical study on urban and rural residents’ physical exercise participation

Using self-reported physical exercise frequency data from the 2023 Chinese General Social Survey (CGSS 2023), this study examines differences in exercise participation across age groups and between urban and rural residents. Ordinal regression, Karlson–Holm–Breen (KHB) associational decomposition, propensity score matching, and Oaxaca–Blinder decomposition are applied to examine the statistical associations between socioeconomic resources and time budgets. The results indicate a 12.8-percentage-point urban–rural gap in high-frequency exercise participation among younger adults. In the middle-aged group, the marginal associations between years of education and log personal annual income are weaker than those in the younger group by 0.6 and 0.9 percentage points, respectively. The joint decomposition shares associated with working and housework hours are 29.4% for the overall association across age groups, 29.8% for the interaction between education and age, and 40.5% for the interaction between income and age. These shares pertain to distinct decomposition targets and should be interpreted separately. The interaction between health knowledge and age is statistically significant but has a small effect on the probability scale, ranging from approximately 0.1 to 0.2 percentage points, suggesting limited substantive public-health importance. Since the data are derived from a single cross-sectional survey, the findings highlight conditional associations and differences between groups but do not identify causal mediation pathways or net policy effects.