Date
Publisher
arXiv
The rapid development of generative artificial intelligence (GenAI) tools such as ChatGPT has intensified interest in their role in higher education, particularly in how students perceive and use them and how these perceptions may relate to educational outcomes. This study employs a hybrid methodological approach that combines a PRISMA-guided systematic literature review with simulation-based modeling to examine student perceptions of GenAI in higher education. Nineteen empirical articles published between 2023 and 2025 were identified through a Scopus-based review, and thematic synthesis was used to organize the emerging patterns in the literature. Of these, six studies reported item-level means and standard deviations suitable for probabilistic modeling. From this subset, one well-structured Likert-scale dataset was selected as a canonical example for inverse-variance-weighted Monte Carlo simulation. The simulation generated a composite perception-based Success Score, enabling estimation of both central tendency and uncertainty under different thematic configurations. The findings indicate that usability-related factors, particularly System Efficiency and Learning Burden, exert the greatest influence on the composite score under the specified weighting scheme, while other themes also contribute positively but more modestly. The study offers a transparent and privacy-preserving bridge between thematic synthesis and predictive probabilistic modeling, providing a reproducible framework for linking GenAI perceptions to educational outcomes in future research.
Study at a glance
- Question: How can published survey statistics be used in a simulation to quantify students' perceived success with generative AI?
- Setting and data: Undergraduate health sciences education survey data from published literature (Veras et al., 2024), yielding 10,000 synthetic observations.
- Method: Literature search and Monte Carlo simulation
Why it matters
Research on students' perceptions of generative AI remains largely descriptive. Developing quantitative models helps evaluate these perception data under uncertainty when individual-level student data cannot be shared.
Key findings
- The simulated perception-based student success score averaged 4.07 on a 5-point scale, indicating generally positive student evaluations of ChatGPT usability.
- System efficiency and learning burden had the largest influence on the composite score because it had the lowest variance among the evaluated themes.
- Ease of use and learnability and system efficiency both produced high weighted theme averages of approximately 4.12, whereas perceived complexity and integration averaged 3.71.
Summary drafted with AI from the full paper and may contain errors. See the full paper for details.
What is the application?
Who is the user?
Who age?
Why use AI?
Study design
