---

CDDRefactorER: Improving Code Comprehension through Cognitive-Load Aware Refactoring for Novice Programmers Welcome to the repository for the research paper, "Reducing Novice Programmer's Mental Burden: Improving Code Comprehension via Cognitive-Load Aware Automated Code Refactoring". This study explores how CodeRefactorER, a generative AI-powered tool built on Cognitive-Driven Development (CDD) principles, simplifies complex code to reduce cognitive load and enhance comprehension for novice programmers. A controlled study with 20 first-semester computer science students demonstrated significant improvements in code understanding, logic flow tracing, function identification, and reduced mental strain. Repository Overview This repository provides all scripts, datasets, and supplementary materials used in the study. Repository Structure CodeRefactorER-instructions.mdThe custom GPT instruction that is deployed in GPT platform. prompts/CDDRefactorER-prompt.mdThe custom prompt used to configure LLM/ChatGPT as a supportive mentor, applying CDD principles to refactor code by simplifying complex structures, enhancing readability, and minimizing cognitive load for novice programmers. prompts/Unguided-prompt-prompt.mdThe unguided-prompt used in this study. survey/pre-test_survey_questions.mdA set of survey questions designed to evaluate novice programmers' baseline understanding, perceived mental strain, and frustration when engaging with complex code prior to using CodeRefactorER. survey/post-test_survey_questions.mdA set of survey questions crafted to assess improvements in code comprehension, cognitive load, and user experience among novice programmers after utilizing CodeRefactorER. survey/task_list.csvA structured dataset listing 20 programming tasks from the Mostly Basic Python Problems (MBPP) dataset used in the study. survey/User_Study-Participant_Instructions.pdfDetailed, step-by-step instructions provided to first-semester computer science students for participating in the code comprehension study using CodeRefactorER. Usage and Reproducibility All materials in this repository are designed to support replication of the study or adaptation for related research. GPT Link Anonymized GPT Link: https://chatgpt.com/g/g-6801a23d42488191bd410191cd8512a1-beginner-code-refactor The GPT link as far as we know is anonymized. But please proceed with causion.