Background Anxiety is common among medical undergraduates, yet modifiable behavioral pathways underlying the association between physical activity motivation and anxiety remain unclear. Guided by self-determination theory, the Interaction of Person–Affect–Cognition–Execution (I-PACE) model, and the two-process/hyperarousal framework of sleep, we tested smartphone addiction symptoms and sleep quality as parallel and serial statistical mediators of this association. Methods In a cross-sectional survey of 1,276 Chinese medical undergraduates, we assessed physical activity motivation (Motives for Physical Activity Measure–Revised, 15 items), smartphone addiction symptoms (Smartphone Addiction Scale–Short Version), sleep quality (Sleep Quality Questionnaire−9; higher scores indicate poorer sleep), and anxiety (Generalized Anxiety Disorder−7). All instruments were administered using a harmonized 5-point Likert response format. Indirect effects were estimated using Hayes' PROCESS macro (Model 6) with 5,000 bootstrap resamples, controlling for sex and grade (academic year). Results In bivariate analyses, physical activity motivation was inversely correlated with anxiety ( r = −0.261, p 0.001). In the covariate-adjusted serial mediation models, the total effect on anxiety was c = −0.311 (95% CI −0.385, −0.237) and remained statistically significant after accounting for smartphone addiction symptoms and sleep quality (direct effect: c ′ = −0.216, 95% CI −0.293, −0.139). The total indirect effect was B = −0.095 (95% CI −0.126, −0.066). Specific indirect effects were significant via smartphone addiction symptoms ( B = −0.062, 95% CI −0.086, −0.040) and via sleep quality ( B = −0.028, 95% CI −0.046, −0.012), as well as via the serial pathway (physical activity motivation → smartphone addiction symptoms → sleep quality → anxiety; B = −0.006, 95% CI −0.010, −0.002; 1.80% of the total effect). Conclusion Higher physical activity motivation was associated with lower anxiety, with this association partially statistically accounted for by lower smartphone addiction symptoms and better sleep quality (lower sleep-quality scores), including a modest serial indirect pathway. Given the cross-sectional design, these indirect effects should be interpreted as statistical associations rather than causal mechanisms; longitudinal and experimental studies using objective smartphone-use and sleep metrics are needed to test temporal ordering.
Ma et al. (2026) studied this question.