Distinct learner profiles of attitudes toward biostatistics in medical education: a latent profile analysis

Abstract

Biostatistics is essential for interpreting evidence and clinical decision-making, yet medical learners may value it while experiencing anxiety, low enjoyment, and perceived difficulty. Average scores may fail to capture distinct attitude patterns. This study aimed to identify latent profiles of attitudes toward biostatistics among first-year medical students and medical residents and compare their sociodemographic and academic characteristics. This cross-sectional study included 193 participants: 132 first-year medical students and 61 medical residents. Attitudes were assessed using the 33-item Statistics Attitude Scale, with responses framed in relation to biostatistics in medical education. Item-level responses were treated as continuous indicators in latent profile analysis. Models with one to seven profiles were estimated using an equal-variance, zero-covariance parameterisation. Model selection considered fit indices, profile size, and interpretability. Characteristics were compared across profiles. A five-profile solution was retained because it had the lowest Bayesian information criterion, high classification quality (entropy = 0.970), adequate profile sizes, and clear interpretation. The six- and seven-profile solutions included profiles with only one participant. The retained profiles were Low-Engagement and Anxious (24.4%), Value-Aware but Anxious (28.0%), Moderately Receptive (18.7%), Highly Disengaged and Anxious (10.9%), and Highly Receptive and Confident (18.1%). All five attitude dimensions differed across profiles (all p < 0.001), with the largest between-profile difference observed for Relation of Statistics to Professional Life (ε² = 0.821). The Value-Aware but Anxious profile combined high professional relevance and importance with greater anxiety, lower enjoyment, and greater perceived difficulty. Educational level also differed across profiles ( p < 0.001; Cramér’s V = 0.449): first-year students were more common in three profiles, whereas residents were more common in the Value-Aware but Anxious and Highly Receptive and Confident profiles. Attitudes toward biostatistics among medical learners varied and could not be represented on a single scale from negative to positive. Distinguishing learners who valued biostatistics but remained anxious from those showing low engagement may help inform clinically relevant and tailored educational support. The profiles should be regarded as cross-sectional and preliminary rather than as fixed learner types.