A multivariate predictive model of academic performance and licensure outcomes among technical-vocational education graduates using demographic and learning condition indicators

Despite the rapid shift to online and hybrid modalities during the COVID-19 pandemic, limited empirical evidence exists on how instructional modes, demographic factors, and competency-based performance collectively shape licensure outcomes among teacher education graduates. This study addressed this gap by examining 150 Technical-Vocational and Livelihood Education graduates from a state university in Northern Philippines, analyzing their demographic profiles, course-level grades, and board examination results to develop a multivariate associative and exploratory predictive framework. Using MANOVA, the study found significant multivariate differences across instructional modes for both programs (BTLED Pillai’s Trace = 0.7788, p<.001; BTVTED Pillai’s Trace = 0.4420, p<.001). Principal Component Analysis revealed that two components explained 59.74% of total variance in performance (PC1 = 33.77%, PC2 = 25.97%), while clustering analysis identified two natural learner groups (n = 73 and n = 77). Multiple regression showed that learning conditions were the most strongly associated with academic performance, particularly connectivity (β=–0.34), SHS strand (β=–0.28), and specialization (β=–0.31), all significant at p<.001. Logistic regression examining associations with board examination passing yielded acceptable fit indices (Accuracy = 0.827; AUC = 0.909; interpreted as exploratory given the absence of cross-validation), with Professional Education (OR = 0.197) and Mandated Courses (OR = 0.346) emerging as the strongest predictors of passing likelihood. Overall, the findings suggest that academic readiness and digital access are more impactful than fixed demographic traits, though causal inferences cannot be drawn from this retrospective design. The study points to opportunities for enhanced digital support, strengthened curricular alignment, and cautious exploratory use of learning analytics to better understand student performance patterns in teacher education.