Investigating the Role of a Conversational Artificial Intelligence Model in Promoting Self-Regulated Learning in Project-Based Learning for Software Engineering Students for Vocational Education

Kwok, Yu Ho and Passey, Don (2025) Investigating the Role of a Conversational Artificial Intelligence Model in Promoting Self-Regulated Learning in Project-Based Learning for Software Engineering Students for Vocational Education. PhD thesis, Lancaster University.

Text (PhD_Thesis_Kwok_2025)
PhD_Thesis_Kwok_Final.pdf - Published Version
Available under License Creative Commons Attribution-NonCommercial.
Download (3MB)

Abstract

This PhD thesis investigates the integration of ChatGPT, a Conversational Artificial Intelligence (AI) model, into Project-Based Learning (PBL) environments to enhance Self-Regulated Learning (SRL) among software engineering students at V-Institute, a leading Vocational and Professional Education and Training (VPET) institution in Hong Kong. Employing a mixed-methods comparative design, the study combines quantitative data from the Motivated Strategies for Learning Questionnaire (MSLQ) and academic assessments with qualitative insights from interviews and ChatGPT logs. Findings reveal ChatGPT’s comparable, slightly enhanced support for SRL compared to traditional PBL, with significant gains in effort regulation and help-seeking (p<0.05) and modest academic improvements (p=0.041). It fosters SRL through personalised guidance, debugging, and reflective reviews, supporting what are regarded as sophisticated projects aligned with industry standards. However, challenges such as dependency risks, technical limitations, and mixed teamwork impacts necessitate strategic curriculum design. The study proposes embedding AI literacy and prompt engineering to balance autonomy and collaboration, addressing VPET’s resource constraints. Theoretical contributions enrich SRL and PBL frameworks, while practical implications advocate curriculum redesign for Hong Kong’s higher and vocational institutes. Despite limitations such as the reliance on a single AI tool (ChatGPT), which may limit the generalisability to other AI models, the research highlights ChatGPT’s transformative potential, offering a scalable, innovative approach to prepare students for AI-augmented workplaces while fostering critical thinking and industry-ready skills.