Predicting Student Attrition in Competitive Programming: A Large-Scale Study Integrating Survey Insights and Global Behavioral Logs

Research combining large-scale behavioral data from Competitive Programming (CP) platforms with psychographic survey insights across multiple universities reveals patterns preceding student attrition. Key findings indicate a significant reduction in contest participation and an increase in 'struggle time' before disengagement. The study identifies a 'Skill-Application Paradox' as a contributing factor to attrition, highlighting the challenges students face in sustained engagement within competitive programming environments.

Why it matters

Understanding the dynamics of student attrition in specialized skill development environments like competitive programming is crucial for educational institutions and program organizers. Addressing these factors can improve retention rates, enhance skill development pipelines, and ensure a more robust talent pool in fields requiring advanced algorithmic reasoning.

Key insights

• Competitive programming (CP) helps computer science students develop algorithmic reasoning skills.
• Sustained student participation in CP is challenging, often due to skill plateaus or performance anxiety.
• Traditional educational data mining (EDM) often overlooks CP attrition, focusing more on MOOCs and academic courses.
• The study integrated large-scale Codeforces activity logs (n=1,816) with a multi-institutional psychographic survey from 10 Bangladeshi universities (n=64).
• True attrition is characterized by an 83.71% reduction in contest participation.
• A 15.6% increase in 'struggle time' also precedes attrition.
• The research identifies a 'Skill-Application Paradox' as a factor in student disengagement.

Source

arXiv — Computers and Society — https://arxiv.org/abs/2608.28618