Just below the line: appeal behaviour and outcomes at a national exam cutoffs

This paper investigates how marginal failure at critical grading thresholds influences contestation behaviour and academic outcomes using 2022–2025 Romanian Baccalaureate data, thereby excluding the COVID-19 disruption period. We estimate reduced-form effects at the 5.00 (subject-level) and 6.00 (overall-average) pass marks using robust local linear regression discontinuity designs (RDD) with data-driven bandwidth selection. Robustness checks including donut-hole RDD and placebo cut-offs ensure result credibility. By focusing on students near the passing margin, the study provides evidence on the behavioural salience of high-stakes thresholds. These findings emphasize the necessity for targeted counselling and refined contestation protocols in national examination systems to support students at the margin.