Dual-submission homework, where students submit work, receive feedback and then revise has gained attention as a way to foster reflection and discourage reliance on online answer repositories. This study analyzes 13 years of exam data from a computer architecture course to compare student performance under single versus dual-submission homework conditions. Using pooled t-tests on matched exam questions, we found that dual-submission significantly improved outcomes in a majority of cases. The results suggest that reflective resubmission can meaningfully enhance learning and may serve as a useful strategy in today's AI-influenced academic environment. This full research paper also discusses pedagogical implications and study limitations.
@article{arxiv.2508.09314,
title = {Reflective Homework as a Learning Tool: Evidence from Comparing Thirteen Years of Dual vs. Single Submission},
author = {Madhur Dixit and Kavya Lalbahadur Joshi and Kaveri Bhalchandra Konde and Edward F. Gehringer},
journal= {arXiv preprint arXiv:2508.09314},
year = {2025}
}
Comments
Accepted for presentation at the Frontiers in Education Conference, Nashville, Tennessee, USA, 2-5 November 2025