Publication Type

Conference Proceeding Article

Version

acceptedVersion

Publication Date

5-2026

Abstract

Programming errors and misconceptions are pervasive in novice programmers which causes difficulty in the learning of computer programming. Large Language Models (LLMs), with their ability to comprehend and generate programming codes have shown promising results in the automatic identification of errors. This can potentially benefit student programmers by providing them with timely formative feedback at efficiencies and scale that were not attainable previously. In this study, we leveraged an LLM - OpenAI o4-mini for the generation of elaborated, targeted feedback for novice programmers across PHP and JavaScript exercises. We contend that the feedback needs to be effective and targeted other than being correct to address novice programmers’ misconceptions. We designed a two-level prompting system and incorporated a custom tool to overcome the issues of verbose generated feedback and incorrect line numbers extracted for erroneous code blocks respectively. Our evaluation showed that the generated feedback were targeted and specific with no cases of false positives where non-erroneous codes were wrongly picked up. Notably, some of the generated feedback were well explained and uncommon. We thus conclude that elaborated, targeted and thus effective feedback for novice programmers across different programming languages can be achieved by leveraging on LLM.

Keywords

programming, Large Language Model, personalized feedback, error, misconception

Discipline

Instructional Media Design | Programming Languages and Compilers

Research Areas

Data Science and Engineering

Publication

Proceedings of the 18th International Conference on Computer Supported Education (CSEDU 2026), Benidorm, Spain, May 18-20

First Page

1

Last Page

8

Identifier

10.5220/0015019100004021

Publisher

Scitepress

City or Country

Spain

Additional URL

https://doi.org/10.5220/0015019100004021

Download

Share

COinS