Which has an impact on which? Academic performance or mental health. A nationwide empirical study of 50 universities in China

Background While numerous studies have identified a strong association between academic performance and mental health, the direction of causality remains unclear. This study aims to examine whether academic performance impacts mental health or vice versa. Methods A multistage sampling survey was conducted among 11,659 medical students from 50 universities across the country. Two mediation models were constructed for analysis: (1) Academic stress → Academic performance → Mental health (AAM), and (2) Academic stress → Mental health → Academic performance (AMA). Data were analyzed using descriptive statistics and mediation analysis. Results The Bootstrap test showed that the 95% confidence interval (CI) ranged from 0.000 to 0.002 in AAM model and 0.003 to 0.008 in AMA model. Academic performance does not have a mediating effect in AAM, whereas in the AMA, mental health serves as a mediator. Although both the AAM and AMA models showed mediating effects by using the three-parameter comparison method, the ratio of the mediating effect to the total effect was 1.6% in the AAM model and 17.7% in the AMA model. Model fit indices in SEM revealed that, aside from the goodness-of-fit index (GFI), the AAM model did not meet acceptable fit criteria. In contrast, the AMA model showed good fit across most indices. Conclusion The findings support the AMA model, suggesting a more plausible causal pathway in which mental health significantly affects academic performance. This insight underscores the importance of addressing mental health concerns to enhance students’ academic outcomes, offering valuable guidance for educational policy and student support interventions.