GenAI‑Enhanced Software Engineering Conversion Programmes: A Longitudinal Quasi‑Experimental Study on Knowledge Architecture Construction for Non‑Computer Science Background Students

DOI

10.2991/978-94-6239-766-8_19How to use a DOI?

Keywords

Software engineering education; conversion programmes; generative AI; knowledge architecture; quasi‑experimental study; non‑CS background students; GenAI‑Scaffolded Conversion Framework

Abstract

The persistent shortage of qualified software engineers has led to a proliferation of conversion programmes that enable graduates from non‑computer science (non‑CS) disciplines to enter the field. While effective in building foundational knowledge, empirical evidence on how generative AI (GenAI) tools can accelerate the systematic construction of a coherent knowledge architecture remains limited. This longitudinal quasi‑experimental study examines a one‑year Master’s‑level software engineering conversion programme involving 78 mature students from diverse non‑CS backgrounds. Cluster randomisation (six tutorial groups as clusters) allocated participants to a control group (traditional curriculum, n = 39) or an experimental group (n = 39) using the GenAI‑Scaffolded Conversion Framework (GSCF). The intra‑cluster correlation coefficient (ICC = 0.07) confirmed low clustering effects, justifying the use of mixed‑effects models. Pre‑ and post‑intervention assessments, concept‑mapping exercises, and LLM interaction logs were analysed. The experimental group showed significantly higher gains in knowledge integration (Cohen’s d = 0.82, 95% CI [0.48, 1.16]), software engineering knowledge test scores (d = 0.74 [0.41, 1.07]), and capstone project performance (d = 0.69 [0.34, 1.04]). Analysis of 4,687 LLM interaction sessions revealed a three‑phase usage pattern via latent growth mixture modelling: exploratory (Weeks 1–6, revision rate 42%), structured (Weeks 7–20, revision rate 18%), and reflective (Weeks 21–36). Experience sampling method (ESM) data showed that frustration frequency declined from 2.4 to 0.7 episodes per module (p < 0.001), with 81% of participants converting initial frustration into productive struggle through structured reflection. The study contributes a practical framework that integrates GenAI as a scaffold for knowledge systematisation while addressing emotional and cognitive challenges. Implications for curriculum design, instructor facilitation, and responsible GenAI integration are discussed.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Guangxi Peng
AU  - Ningzhang
PY  - 2026
DA  - 2026/09/04
TI  - GenAI‑Enhanced Software Engineering Conversion Programmes: A Longitudinal Quasi‑Experimental Study on Knowledge Architecture Construction for Non‑Computer Science Background Students
BT  - Proceedings of the 2026 6th International Conference on Education, Information Management and Service Science  (EIMSS 2026)
PB  - Atlantis Press
SP  - 188
EP  - 196
SN  - 2589-4900
UR  - https://doi.org/10.2991/978-94-6239-766-8_19
DO  - 10.2991/978-94-6239-766-8_19
ID  - Peng2026
ER  -