The Influence of Attitude towards Statistics on Statistical Proficiency among Medical Students: A Chain Mediating Effect of Statistical Anxiety and Statistics Self-efficacy

Background Statistics education is crucial for medical students as it lays the foundation for evidence-based practice and clinical research. Prior studies have identified various affective and cognitive factors that may influence statistical proficiency, yet little is known about the chain-mediated mechanism by which attitudes toward statistics sequentially affect statistics anxiety, self-efficacy, and ultimately proficiency. Aiming to fill the gap, the present study investigates the complex mediating pathways among medical students. Methods A total of 576 medical students from Shandong University participated in this cross-sectional study. Participants completed validated Chinese-adapted versions of the Statistics Anxiety Rating Scale (STARS) and the Current Statistics Self-Efficacy (CSSE) scale. Statistical proficiency was measured using final examination scores. Hypothesized mediation pathways were tested via structural equation modeling (SEM), with bootstrap methods applied to verify indirect effects. Results The mean total STARS score was 2.126 ± 0.597, with computation self-concept identified as the weakest dimension. A good-fitted SEM showed that negative attitudes directly impaired proficiency (standardized efficient = -0.120, 95% CI: [-0.224,-0.027]). Meanwhile, its effect was also sequentially mediated by statistical anxiety and statistics self-efficacy (indirect effect = -0.049, 95% CI [–0.096, − 0.001]). Statistics self-efficacy also independently mediated the effect of attitude on proficiency (indirect effect = − 0.035, 95% CI [–0.069, − 0.008]). However, the direct effect of anxiety on proficiency was not significant (standardized efficient = -0.018, 95% CI [–0.049, 0.033]). Conclusions These findings reveal a distinctive “high-competence but low-confidence” phenomenon among Chinese medical students and highlight the complex interplay between cognitive, emotional factors, and statistical proficiency in statistics education. The complete mediation of self-efficacy on anxiety to proficiency indicates that educational interventions should prioritize enhancing statistics self-efficacy and fostering positive attitudes toward statistics rather than focusing solely on reducing anxiety, to improve learning outcomes in medical statistics education.