Predictive Effects of Core Subject Grades on Senior High School Strand Selection

Abstract

This study examined whether Grade 10 core-subject grades (English, Mathematics, and Science) can be used to provide a preliminary, grades-only screening of Senior High School (SHS) strand options (ABM, HUMSS, STEM, and TVL). Using multinomial logistic regression on 400 student records, English emerged as the strongest predictor, with Mathematics and Science contributing more modestly in selected comparisons. In classification, the grades-only model correctly classified 215 of 400 cases (overall accuracy = 53.75%). Preliminary grades-only screening achieved an overall accuracy of 53.75%, constrained by grade profile overlap across strands. Strand-level performance varied substantially (ABM = 17.4%, HUMSS = 49.2%, STEM = 75.0%, TVL = 74.6%), indicating limited separability for strands with similar grade patterns. These results support the use of grades as an initial filter to narrow counseling conversations, but not as a stand-alone placement or decision-making tool without complementary measures such as interest inventories, aptitude assessments, and structured guidance inputs.