Paper Number
ECIS2026-1540
Paper Type
SP
Abstract
Code review is a critical software engineering practice for improving software quality, enforcing coding standards and supporting knowledge sharing. Incorrect code reviews that deviate from established standards can provoke negative affect and perceptions of unfairness among code contributors, thereby impacting their code revision performance. Advancement in large language models have enabled automated code review tools, offering opportunities to support review processes and address interpersonal challenges. This study examines the mechanisms linking review accuracy to contributors’ code revision performance, with a focus on differences between incorrect reviews produced by human reviewers and automated code review tools. We propose a research model and a 2 (review accuracy) ×2 (review source) experimental design to test these mechanisms. The study has potential contributions to affective events theory and organizational justice theory, and practical insights for integrating automated code reviews for engineers’ wellbeing and performance.
Recommended Citation
Li, Qianyi; Fernandoe, Niroshinie; Turel, Ofir; Alam, Sultana Lubna; and Loke, Seng W., "Harnessing Automated Code Review For Better Code Revision: The Lens Of Affect And Fairness" (2026). ECIS 2026 Proceedings. 3.
https://aisel.aisnet.org/ecis2026/genai/genai/3
Harnessing Automated Code Review For Better Code Revision: The Lens Of Affect And Fairness
Code review is a critical software engineering practice for improving software quality, enforcing coding standards and supporting knowledge sharing. Incorrect code reviews that deviate from established standards can provoke negative affect and perceptions of unfairness among code contributors, thereby impacting their code revision performance. Advancement in large language models have enabled automated code review tools, offering opportunities to support review processes and address interpersonal challenges. This study examines the mechanisms linking review accuracy to contributors’ code revision performance, with a focus on differences between incorrect reviews produced by human reviewers and automated code review tools. We propose a research model and a 2 (review accuracy) ×2 (review source) experimental design to test these mechanisms. The study has potential contributions to affective events theory and organizational justice theory, and practical insights for integrating automated code reviews for engineers’ wellbeing and performance.