Evaluating the Impact of Affirmative Action on Student Selection Outcomes: A Data Mining Approach Using SKD Test Performance at STMKG

Abstract

This study examines the impact of affirmative action on student selection outcomes at Sekolah Tinggi Meteorologi Klimatologi dan Geofisika (STMKG) in Indonesia, utilizing a data mining approach focused on SKD test performance. Affirmative action programs at STMKG aim to support applicants from marginalized regions, providing them with increased access to specialized academic opportunities. Using a dataset comprising TWK, TIU, TKP, and total SKD scores, this research compares the performance of affirmative action applicants against regular applicants, investigating differences in pass rates and score distributions. Exploratory data analysis reveals that regular applicants generally achieve higher mean scores across all test components. Statistical tests, including ANOVA, confirm significant differences in mean scores, particularly in TIU and total SKD scores. A Random Forest classification model was applied to predict selection outcomes, achieving an accuracy of 84.9%, although disparities in precision and recall indicate challenges in consistently classifying affirmative action applicants. Findings suggest that while affirmative action policies enhance access for underrepresented groups, additional support mechanisms may be required to bridge the performance gap. This research contributes valuable insights for policymakers and educational institutions seeking to refine affirmative action strategies, ensuring they promote both access and equity in competitive selection processes. Further studies could extend this analysis to include socio-economic factors and post-admission academic performance, offering a comprehensive evaluation of affirmative action’s long-term impact on student success.