Multiple-choice questions as a means of assessment in the face of generative AI: a Python analysis of master’s students’ perceptions

Abstract

The justifiable multiple-choice question, which requires students to select an answer and then justify it, is often presented as a defence against the unauthorised use of generative AI. This case study focuses on 41 of the 50 Master’s students at the Higher School of Education and Training (Ibnou Zohr University) who sat an exam in this format (response rate: 82 per cent). Data from a mixed-methods questionnaire were processed using a reproducible analysis pipeline in Python: descriptive statistics, reliability, non-parametric tests, k-means clustering and automated analysis of open-ended responses. The students recognise the format’s high educational value (71 per cent believe it forces them to understand better; α = 0.85), but only 25 per cent consider it more difficult to circumvent using AI (Wilcoxon, p < 0.001, r = 0.62). This scepticism is shared by both identified profiles and does not depend on the frequency of AI use. Students attribute the resistance to the contextual nature of the questions (7 per cent consider a question relating to classroom experience to be vulnerable) and the majority suggest redesigning the tasks rather than increasing supervision (71 per cent versus 22 per cent).

Keywords: Justification-based multiple-choice questions; generative AI; academic integrity; authentic assessment; educational data analysis; Python; case study.

Received Date: August 17, 2026

Accepted Date: September 09, 2026

Published Date: October 01, 2026

Available Online at: https://www.ijsrisjournal.com/index.php/ojsfiles/article/view/986