A Bayesian multilevel path model of college students' pickleball participation motivation: Theory-informed priors and hierarchical random intercepts.

Despite explosive global growth of pickleball, motivational processes underlying college student participation remain poorly understood. We present a Bayesian multilevel path model integrating theory-informed priors with hierarchical random intercepts for grade and major. A total of 1992 college students from 12 higher-education institutions across six provinces of China completed validated measures. Using brms (v2.21.0) in R (v4.3.1) with Stan (v2.32), results showed psychological needs as the dominant predictor (β = 0.613, 95% CrI [0.532, 0.692]), followed by environmental factors (β = 0.118, 95% CrI [0.052, 0.184]) and social media use (β = 0.180, 95% CrI [0.140, 0.223]). Personal and social factors showed marginal or non-significant associations after controlling for psychological needs and environment. Hierarchical random intercepts revealed exploratory between-grade variation, with freshman students reporting the highest motivation. Bayesian mediation confirmed social media use showed a small indirect association (8.5% of total effect for psychological needs). The model achieved elpd_loo = -1425, decisively outperforming seven alternatives (model weight = 1.000). Prior sensitivity analysis demonstrated robustness across prior specifications. These findings suggest psychological need satisfaction and environmental support as primary levers for promoting pickleball on university campuses, while grade-related patterns identify grade-adapted programming as a possibility warranting future longitudinal investigation.