Physical function factors associated with the modified star excursion balance test in healthy university students.

Cross-sectional design: correlations cannot be read as cause-and-effect; a low balance score may reflect deconditioning rather than a specific deficit. Small sample (35 participants) makes the regression models vulnerable to overfitting and unstable coefficient estimates, as the authors themselves note. Variables were selected for regression based on which ones showed significant bivariate correlations, introducing selection bias and inflating the chance of false-positive findings. No adjustment for multiple comparisons across the many correlation tests performed. Adjusted R² values around 0.2 mean roughly 80% of the variation in balance performance is unexplained by the variables measured. No electromyography, motion analysis, or perturbation testing, so the neuromuscular mechanisms behind the correlations are speculative. Body composition was measured by bioelectrical impedance, which is sensitive to hydration, meal timing, and time of day. Left and right limb scores were averaged, potentially masking meaningful side-to-side asymmetry. Sex, height, and sport participation were not controlled for, and the sample is limited to healthy young adults, so findings may not generalise to clinical or athletic populations.

Funded by JSPS KAKENHI Grant Number JP24K20382 (a Japanese government research council). The authors declare no conflicts of interest. No commercial sponsor designed, conducted, or wrote the study.

Do not read the significant bivariate correlations (e.g., ankle plantarflexion strength r=0.49 in the posterolateral direction) as evidence that strengthening that one muscle group will fix a patient's balance deficit. The multiple regression, which is the more rigorous test, found no independent predictor in any direction, and the authors themselves state that dynamic balance reflects the combined influence of multiple interrelated factors rather than a single dominant one.