Exploring the Performance Drivers in Competitive Exams using Statistical Inference

Abstract Large-scale educational assessments generate extensive datasets that require advanced statistical approaches to extract meaningful insights into learner performance, assessment quality, and demographic disparities. The Maharashtra Talent Search Examination (MTSE) is a competitive assessment designed to evaluate the academic potential of students across multiple subjects. This study presents a comprehensive statistical analysis of MTSE datasets from recent years, focusing on subject-wise performance patterns, regional variations, and year-on-year trends in student outcomes. Reliability analysis was conducted to assess the internal consistency of subject components. Exploratory data analysis was employed to examine score distributions, central tendency, and variability across academic subjects and demographic groups. Inferential statistical techniques were applied to identify significant differences in performance across gender and rural–urban categories. The results reveal notable associations between demographic factors and academic performance, as well as strong interrelationships among subjects. These findings provide data-driven insights into student strengths, weaknesses, and regional disparities, supporting informed educational planning, targeted academic interventions, and evidence-based policy formulation in large-scale assessments such as MTSE.