<![CDATA[Low achievement in statistics courses is a persistent concern in higher education, yet the contributions of study time and attendance remain underexplored, particularly in Indonesian undergraduate contexts and without correcting for predictor multicollinearity. This study addresses this gap by applying Ridge regularization to estimate the independent effects of daily study hours and attendance rate on statistics course scores among 20 Indonesian undergraduates. Classical OLS assumptions were tested; Ridge regression with Leave-One-Out Cross-Validation was applied when violations occurred. Study hours and attendance jointly explained 97.54% of score variance (R² = 0.975; F = 337.70; p < 0.001); however, severe multicollinearity (VIF = 11.77) destabilized OLS estimates. Ridge regression (λ = 1.0) produced stable coefficients (b₁ = 2.87; b₂ = 0.80) with negligible accuracy loss (R² = 0.974). The near-perfect collinearity of the predictors indicates that study hours and attendance are part of the same underlying construct academic engagement rather than acting through independent pathways. These findings challenge the conventional additive model of learning time and imply that interventions targeting self-regulated learning may be more effective than policies addressing attendance or study time in isolation.]]>
Copyrights © 2026
