Social support and exercise persistence among chinese university students: integrating social ecological and self-determination theories

Abstract

Under the macro-guidance of the "Healthy China 2030" strategy, promoting students’ physical health and cultivating lifelong sports habits has become a long-term core issue and important strategic goal of China’s educational policies. This study examined whether social support was associated with exercise persistence among university students and whether exercise motivation and exercise investment statistically accounted for this association in sequence. Social ecological theory specifies the environmental level of the framework, whereas self-determination theory explains how a supportive context may correspond to more self-endorsed motivation and sustained behavioral investment. A cross-sectional survey design was used. A convenience sampling and snowball sampling method recruited 430 college students from a university in China. Data were collected using the "College Student Social Support Scale," "Exercise Motivation Scale," "Exercise Investment Self-Assessment Scale," and "Exercise Persistence Scale." The PROCESS macro was used to conduct a chain mediation model test. At the focal-construct level, social support, exercise motivation, exercise investment, and exercise persistence were positively associated. The four-factor measurement model fit substantially better than a single-factor model; HTMT values ranged from 0.298 to 0.416, and the Fornell–Larcker criterion was met. All three bootstrap indirect associations excluded zero. In the unadjusted model, the direct association remained small but significant (B = 0.102, 95% CI [0.003, 0.200], p = .043), indicating partial statistical mediation. After adjustment for gender, year of study, residence, and field of study, all indirect associations remained significant, while the direct association was attenuated (B = 0.096, p = .060). In this single-university cross-sectional sample, social support was associated with exercise persistence through statistically significant indirect pathways involving exercise motivation and exercise investment. The focal constructs showed acceptable discriminant validity, although their original subdimensions did not and were therefore not interpreted separately. These statistical associations do not establish a temporal or causal mechanism.