A theory-driven scale for assessing text-based generative AI literacy from a self-efficacy perspective (T-GASE)
Authors:
- Gürhan DurakGDGürhan Durak
- Serkan ÇankayaSÇSerkan Çankaya
- Semiral ÖncüSÖSemiral Öncü
Top Cited Papers
Open Access
- 4 May 2026
- journal article
- Published by Springer Nature in Education and Information Technologies
- Vol. 31 (14) , 5671-5694
Abstract
This study aimed to develop and validate the Text-Based Generative AI Literacy Scale (T-GASE)—a multidimensional instrument designed to assess individuals’ competencies in interacting with, evaluating, and ethically engaging with text-based Generative AI (GenAI) tools. The scale development followed Hinkin’s five-step framework, including item generation, expert validation, pilot testing, exploratory factor analysis (EFA), and confirmatory factor analysis (CFA). A total of 645 undergraduate students participated in the study, with 332 students in the EFA phase and 313 in the CFA phase, all from diverse academic disciplines such as engineering, education, architecture, economics, and social sciences. Grounded in four well-established theoretical frameworks—the Theory of Planned Behavior (TPB), Technology Acceptance Model (TAM), Expectancy-Value Theory (EVT), and Diffusion of Innovations Theory (DIT)—the final 23-item scale revealed a robust four-factor structure: Application, Expectations, Ethics, and Evaluation. The scale demonstrated strong psychometric properties, with high internal consistency and excellent model fit indices. Results showed that students had high expectations for the future of GenAI and were actively using it but exhibited lower confidence in critically evaluating AI-generated content. Moreover, students from STEM-oriented disciplines outperformed those from non-STEM fields, suggesting disciplinary differences in GenAI literacy levels. Unlike existing AI literacy scales that broadly assess technical knowledge or general attitudes, the T-GASE uniquely addresses the practical, ethical, and evaluative skills specific to text-based GenAI tools, including prompt crafting, misinformation detection, and adaptive usage. These findings highlight the T-GASE as a pioneering and theoretically grounded instrument for assessing GenAI literacy, with significant implications for curriculum design and digital competence development across higher education contexts.Keywords
Funding Information
- Balikesir University
This publication has 45 references indexed in Scilit:
- The critical role of structured training in GenAI adoption: a longitudinal case study on student perceptions and skills developmentInternational Journal of Technology and Design Education, 2025
- Developing and Validating the Artificial Intelligence Literacy Concept Inventory: an Instrument to Assess Artificial Intelligence Literacy among Middle School StudentsInternational Journal of Artificial Intelligence in Education, 2024
- Development and validation of ChatGPT literacy scaleCurrent Psychology, 2024
- What Students Can Learn About Artificial Intelligence – Recommendations for K-12 Computing Education2023
- Artificial intelligence self-efficacy: Scale development and validationEducation and Information Technologies, 2023
- ChatGPT: evolution or revolution?Medicine, Health Care and Philosophy, 2023
- GPT-3: Its Nature, Scope, Limits, and ConsequencesMinds and Machines, 2020
- Grundkurs Künstliche Intelligenz2016
- Empirical research methods in operations managementJournal of Operations Management, 1990
- From Intentions to Actions: A Theory of Planned Behavior1985