Health Disparities
Quantile Regression Analysis of Socioeconomic and Demographic Determinants of Neighborhood Physical Activity in U.S. 500 Cities Jong Cheol Shin* Jong Cheol Shin
Objective: This study examines how socioeconomic and demographic factors influence the percentage of individuals reporting no leisure physical activity, offering insights across varying levels of inactivity using quantile regression.
Methods: Quantile regression was applied at the 10th, 25th, 50th, 75th, and 90th percentiles of no leisure physical activity. Independent variables included poverty rate, median age, and racial/ethnic majority of census tracts (Asian, Black/African American, Hispanic, and White as reference). The analysis utilized data from CDC PLACES (2014–2021), covering 20,202 U.S. census tracts after eliminating missing values. Model fit was assessed using Pseudo-R² and mean absolute error (MAE).
Results: Pseudo-R² values increased from 0.274 (10th percentile) to 0.391 (90th percentile), indicating better model fit at higher inactivity levels. Poverty was positively associated with inactivity across quantiles (β=0.384 to 0.540), as was median age (β=0.120 to 0.162). Census tracts with a majority Black/African American or Hispanic population were linked to higher inactivity across all quantiles, with the strongest effects at the 10th percentile (β=6.909 and β=5.667, respectively). Majority Asian tracts showed mixed effects, with significant associations at the 90th percentile (β=1.396).
Conclusion: Disparities in physical inactivity across U.S. census tracts are influenced by poverty and racial/ethnic composition, varying by level of inactivity. Structural and community-level interventions are essential to promote active lifestyles in high-inactivity areas.