The DOI matches its Crossref record; 71 of the 74 references it cites were found in a registry or an index.
Development and validation of generative artificial intelligence attitude scale for students
Open accessAbstract
Introduction Generative artificial intelligence (AI) tools, such as ChatGPT, have gained significant traction in educational settings, offering novel opportunities for enhanced learning experiences. However, limited research has investigated how students perceive and accept these emerging technologies. This study addresses this gap by developing a scale to assess university students’ attitudes toward generative AI tools in education. Methods A three-stage process was employed to develop and validate the Generative AI Attitude Scale. Data were collected from 664 students from various faculties during the 2022–2023 academic year. Expert evaluations were conducted to establish face and content validity. An exploratory factor analysis (EFA) was performed on a subset of 400 participants, revealing a two-factor, 14-item structure that explained 78.440% of the variance. A subsequent confirmatory factor analysis (CFA) was conducted on a separate sample of 264 students to validate this structure, resulting in the removal of one item and a final 13-item scale. Results The 13-item scale demonstrated strong reliability, evidenced by a Cronbach’s alpha of 0.84 and a test–retest reliability of 0.90. Discriminative power was confirmed through corrected item-total correlations between lower and upper percentile groups. These findings indicate that the scale effectively differentiates student attitudes toward generative AI tools in educational contexts. Discussion The newly developed Generative AI Attitude Scale offers a valid and reliable instrument for measuring university students’ perspectives on integrating generative AI tools, such as ChatGPT, into educational environments. These results highlight the potential for more targeted research and informed implementation strategies to enhance learning outcomes through generative AI.
Overview
CoChat Summary
TL;DR
This study developed and validated a 13-item Generative AI Attitude Scale for university students, yielding a two-factor structure of positive and negative attitudes toward generative AI in education. The scale showed strong reliability and construct validity in a Turkish university sample.
Key findings
- EFA on 400 students produced a two-factor, 14-item structure explaining 78.440% of total variance; CFA on 264 students led to removal of one item, resulting in a final 13-item scale.
- Reliability was strong: overall Cronbach’s alpha = 0.84 and test–retest reliability = 0.90; positive attitude subscale α = 0.97, test–retest = 0.98; negative attitude subscale α = 0.94, test–retest = 0.96.
- CFA fit indices for the two-factor model were χ²/df = 2.126, CFI = 0.98, GFI = 0.93, IFI = 0.98, TLI = 0.98, RMSEA = 0.065, SRMR = 0.0367.
- Corrected item–total correlations after item removal ranged from 0.576 to 0.684 for positive attitude and 0.344 to 0.440 for negative attitude; all items showed significant discriminative power between upper and lower 27% groups.
Methods
- Sequential exploratory mixed-method scale development: literature-based item generation (21 items), review by 15 students with generative AI experience, expert review by eight experts, pilot testing with 14 students, and final 14-item 5-point Likert scale.
- Participants were 664 students from nine faculties, three vocational schools, and a graduate school at one Turkish state university during 2022–2023; EFA used n = 400 (217 female, 183 male), CFA used n = 264 (137 female, 127 male), and test–retest used n = 25 (13 female, 12 male) after excluding two of 27 initial participants.
- Analyses included EFA with principal component analysis and varimax rotation, CFA in AMOS, Cronbach’s alpha, test–retest reliability, corrected item–total correlations, and comparisons between lower and upper 27% groups.
Limitations
- Stated: the sample came from a single state university across various faculties and grade levels, limiting generalizability; many participants reported using ChatGPT specifically; the study focused on university students with generative AI experience and on short-term attitudes.
- Stated: future research should include diverse universities, different generative AI applications, broader student populations, and longitudinal designs to examine long-term effects on learning outcomes, critical thinking, and motivation.
- Evident from the text: the test–retest sample was small (n = 25), and the scale relies on self-reported Likert responses, which may not capture behavioral or contextual factors.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
0 Valid · 1 Not valid · 2 Not checked · 80 Own results and interpretations
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“It can be observed from Table 8 that the correlation between the sub-dimensions of the scale is −0.172.”
§ 1. One-Factor Solution Analysis > 4.3 Correlations between the sub-dimensions
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“As the scale explains, the total variance is 78.440%, representing that the scale effectively accounts for the measured feature.”
§ 4 Results > 4.1 Construct validity
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Internal consistency was assessed through Cronbach’s alpha reliability, resulting in a value of 0.84, indicating high reliability.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Removal of Item 9 improved the overall Cronbach’s alpha from.82 to.84”
§ 1. One-Factor Solution Analysis > 4.2 Item analysis
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“A subsequent confirmatory factor analysis (CFA) was conducted on a separate sample of 264 students to validate this structure, resulting in the removal of one item and a final 13-item scale.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
27 more claims
Create a free account to see all 83 claims and every source verdict.
- It may state two separate claims as one. 16
- Not yet classified. 21
- The quote only partly states this claim. 3
- Probably not a substantial claim. 13
Statistics recomputed
Every reported test result, recomputed from its test statistic and degrees of freedom. A result shows as not adding up only once an independent check has confirmed it.
- Consistentχ²(91) = 5788.732reported p = 0 · recomputed p < .001
“χ2 (91) = 5788.732, p = .000”
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ 4 Results > 4.1 Construct validity
“The first dimension, labelled positive attitude, accounts for 6.489% of the variance and comprises eight items with factor loads ranging from 0.874 to 0.925. On the other hand, the second dimension, termed negative attitude, explains 4.493% of the variance”
§ 4 Results > 4.1 Construct validity
“% eigenvalue (total = 10.982) 6,489 4.493 % variance explained (total = 78.440) 46.351 32.089”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ 1. One-Factor Solution Analysis
“2. Negative attitude factor analysis: the negative attitude dimension (Items 10–13) revealed three distinct concern categories:”
§ 4 Results > 4.1 Construct validity
“Negative attitude Item 12 0.901 0.82 Item 14 0.900 0.81 Item 10 0.871 0.76 Item 11 0.869 0.76 Item 13 0.859 0.74 Item 9 0.768 0.6”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ 1. One-Factor Solution Analysis > 4.2 Item analysis
“The upper 27% and lower 27% of participants were compared”
§ 5 Discussions and conclusion
“Comparisons were made between the top and bottom quartiles of the participants”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ 1. One-Factor Solution Analysis
“a) Future impact concerns (Items 10, 13) Apprehension about long-term implications Societal impact considerations Factor loadings of.900 and.859, respectively”
§ 4 Results > 4.1 Construct validity
“Item 14 0.900 0.81 Item 10 0.871 0.76 Item 11 0.869 0.76 Item 13 0.859 0.74”
Citations
Cites
Hu, L., & Bentler, P. M. (1999) · Structural Equation Modeling A Multidisciplinary Journal · doi:10.1080/10705519909540118 · Cited by 108,788 · cited 2×
“To evaluate the adequacy of the proposed measurement model in relation to the observed data, several fit indices were examined, as suggested by researchers (Byrne, 2010; Hu and Bentler, 1999; Kline, 2005).”
Tabachnick, B. G., & Fidell, L. S. (1983) · Cited by 80,032
Perceived usefulness, perceived ease of use, and user acceptance of information technology
Davis, F. D. (1989) · MIS Quarterly · doi:10.2307/249008 · Cited by 68,483 · cited 4×
“…framework integrates three key theoretical perspectives: 1 Tripartite model of attitudes (Eagly and Chaiken, 1993 2 Technology acceptance variables (Davis, 1989; Venkatesh et al., 2023): - Perceived usefulness in educational contexts - Ease of use in learning environments - Social influence…”
InfluentialHow to Design and Evaluate Research in Education
Fraenkel, J. R., & Wallen, N. E. (1990) · Cited by 14,998
Structural Equation Modelling: Guidelines for Determining Model Fit
Hooper, D., Coughlan, J., & Mullen, M. R. (2008) · Dublin Institute of Technology · doi:10.21427/d7cf7r · Cited by 8,746
Zawacki-Richter, O., Marín, V. I., Bond, M., & Gouverneur, F. (2019) · International Journal of Educational Technology in Higher Education · doi:10.1186/s41239-019-0171-0 · Cited by 7,242
Applied multivariate statistics for the social sciences
Stevens, J. P., & Stevens, J. P. (2001) · doi:10.4324/9781410604491 · Cited by 4,993
Structural Equation Modeling With AMOS: Basic Concepts, Applications, and Programming, Third Edition
Byrne, B. M. (2016) · doi:10.4324/9781315757421 · Cited by 4,653
Dwivedi, Y. K., Kshetri, N., Hughes, L., Slade, E. L., Jeyaraj, A., Kar, A. K., Baabdullah, A. M., Koohang, A., Raghavan, V., Ahuja, M., et al. (2023) · International Journal of Information Management · doi:10.1016/j.ijinfomgt.2023.102642 · Cited by 4,584
Artificial Intelligence in Education: A Review
Chen, L., Chen, P., & Lin, Z. (2020) · IEEE Access · doi:10.1109/access.2020.2988510 · Cited by 4,133 · cited 1×
“The study acknowledges the importance of understanding not merely the technological facet of AI tools but also how they fit into the educational ecosystem, influencing the cognitive and creative dynamics of students (Chen et al., 2020).”
Research works printed but not linked: 8
- Akbulut, 2010, Sosyal bilimlerde SPSS uygulamaları [SPSS applications in social sciences]
- Buyukozturk, 2010, Sosyal bilimler için veri analizi el kitabı [the data analysis handbook for social sciences]
- Eagly, 1993, The psychology of attitudes
- Hooper, 2008, Structural equation modelling: guidelines for determining modelfit, Electron. J. Bus. Res. Methods, 6, 53
- Kline, 2005, Principle and practice of structural equation modelling
- Rogers, 2003, Diffusion of Innovations
- Tabachnick, 2007, Using multivariate statistics
- Yavuz, 2006, Developing a technology attitude scale for pre-service, Chemistry Teachers, 4
Cited by
Goh, A. Y. H., Hartanto, A., & Majeed, N. M. (2025) · Computers in Human Behavior Reports · doi:10.1016/j.chbr.2025.100845 · Cited by 88
Yılmaz, F. G. K., & Yılmaz, R. (2025) · Innovations in Education and Teaching International · doi:10.1080/14703297.2025.2484613 · Cited by 63
“The GenAI attitude scale was developed by Marengo et al. (2025) to assess university students’ attitudes towards GenAI tools, such as ChatGPT, in educational settings.”
Self-citationChen, S., & Cheung, A. C. K. (2025) · Educational Research Review · doi:10.1016/j.edurev.2025.100737 · Cited by 78
Hao, Z., Cao, J., Li, R., Yu, J., Liu, Z., & Zhang, Y. (2025) · Computers & Education · doi:10.1016/j.compedu.2025.105472 · Cited by 73
Lee, H., & Wu, T. (2025) · Computers & Education · doi:10.1016/j.compedu.2025.105375 · Cited by 38
Metadata
- Authors
- Marengo, A., Karaoglan-Yilmaz, F. G., Yılmaz, R., & Ceylan, M.
- Affiliations
- University of Foggia (IT)
- Bartin University (TR)
- Corresponding author
- Marengo, Agostino
- Venue
- Frontiers in Computer Science
- Publisher
- Frontiers Media SA
- Published
- 2025-02-25
- Volume
- 7
- ISSN
- 2624-9898 (ISSN-L)
- OpenAlex
- W4407929143
- Cited by
- 54
- References
- 73
- Citations by year
- 2026: 33 · 2025: 21
- FWCI
- 55.66
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Frontiers in Computer Science (publishedVersion, cc-by)
- SSRN Electronic Journal (acceptedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Topics
- Social Sciences › Education › Artificial Intelligence in Education
- Social Sciences › Gender Studies › Gender and Technology in Education
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Keywords
- generative AI, student attitudes, scale development, ChatGPT, university students, artificial intelligence in education, technology acceptance, psychometric validation
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-09-28, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-09-28 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Development and validation of generative artificial intelligence attitude scale for students. https://app.cochat.ai/papers/10.3389/fcomp.2025.1528455
- The paper
- Marengo, A., Karaoglan-Yilmaz, F. G., Yılmaz, R., & Ceylan, M. (2025). Development and validation of generative artificial intelligence attitude scale for students. Frontiers in Computer Science. https://doi.org/10.3389/fcomp.2025.1528455
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.