Student or educator? Sign up with your school email and get 6 months of Pro free.

The DOI matches its Crossref record, with a small difference in title or year; 72 of the 72 references it cites were found in a registry or an index; 11 of the 60 sources its claims cite were checked for support.

College students’ literacy, ChatGPT activities, educational outcomes, and trust from a digital divide perspective

Open access
Zhang, Ceciley (Xinyi), Rice, Ronald E, Wang, Laurent H·2026·New Media & Society

Abstract

This study investigates the association of socioeconomic status (SES) and digital and AI literacy with types of Chat GPT use by college students, with subsequent implications for academic self-efficacy and creativity, conditioned by trust. Analyses of a survey of U.S. college students (N = 947) show that SES has a greater association with AI literacy than with general digital literacy. Two dimensions of Chat GPT activities emerge: academic support and displacement. Structural equation modeling reveals that AI literacy is positively associated with both activity dimensions, while digital literacy is unexpectedly a negative contributor. Further, academic support is strongly linked to positive outcomes whereas academic displacement is negatively associated. Attitudinal trust in Chat GPT moderates the overall relationships. Our findings suggest that conventional digital inequality persists and evolves with generative AI, traditional digital literacy becomes insufficient in the age of AI, and trust in this new and opaque digital technology influences these relationships.

Overview

CoChat Summary

TL;DR

Applying a digital divide framework to 947 U.S. college ChatGPT users, the study identifies two activity dimensions—academic support and academic displacement—and finds family SES and AI literacy predict greater use while student SES and digital literacy predict less. Support activities are positively associated, and displacement negatively associated, with academic self-efficacy and creativity, with trust moderating several paths.

Key findings

  • CFA confirmed two ChatGPT academic activity dimensions: academic support (8 items; Omega = .94) and academic displacement (4 items; Omega = .91).
  • Family SES was positively associated with AI literacy (β = .08, p = .05), academic support (β = .21, p < .001), and academic displacement (β = .26, p < .001); student SES was negatively associated with digital literacy (β = -.08, p = .01), AI literacy (β = -.12, p = .01), support (β = -.24, p < .001), and displacement (β = -.26, p < .001). H1 was contradicted.
  • Digital literacy was negatively related to support (β = -.33, p < .001) and displacement (β = -.35, p < .001); AI literacy was positively related to support (β = .53, p < .001) and displacement (β = .33, p < .001).
  • Support was positively associated with ASE and PAC (β = .93 for both, p < .001), while displacement was negatively associated with ASE (β = -.46, p < .001) and PAC (β = -.38, p < .001). Under low trust, support-outcome links were stronger (ASE β = .60 vs .36; PAC β = .65 vs .46), and displacement was negatively associated with outcomes only under low trust (ASE β = -.24, p < .001; PAC β = -.17, p = .007).

Methods

  • Pilot survey N = 360 college students at one university (Oct 9–Nov 14) assessed measurement validity/reliability; full study recruited a purposive, non-representative national sample of 947 U.S. ChatGPT users via Qualtrics in Nov 2023, aged 18–35, from 4-year colleges/universities (excluding city/community colleges, online programs, and international exchange students).
  • Measures included SES (7 questions; EFA yielded family SES, 5 items, Omega = .72, and student SES, first-generation status and % loans, Spearman’s rho = .16, p < .001), digital literacy (8 seven-point items, Omega = .90), AI literacy (12 items from Wang et al., second-order Omega = .89 after removing awareness dimension), 12 ChatGPT activity frequencies (5-point scale), ASE (4 items, Omega = .89), PAC (4 items, Omega = .91), and trust (4 items, Omega = .88).
  • Analysis used CFA and SEM in Mplus 8.10 with MLR due to non-normality; model fit RMSEA = .04, CFI = .95, SRMR = .07. Moderation was examined via split-sample path analysis (low vs. high trust).

Limitations

  • Stated: the full sample was national but not representative of U.S. college students; the study is cross-sectional; future research should use longitudinal data, objective measures of ChatGPT use and outcomes, and qualitative/mixed-methods designs; gender and race/ethnicity divides were not hypothesized and should be explored.
  • Stated/evident: split-sample path analysis cannot test the significance of moderation/interaction terms, only provide an overview of path differences.
  • Evident: the sample is limited to ChatGPT users aged 18–35 at 4-year colleges/universities, excluding community colleges, online programs, and international exchange students, so findings do not generalize to non-users or those student populations; activity and outcome measures are self-reported (the paper calls for objective measures).

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1177/14614448241301741

Cites

72 of 72 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1 The conceptual model depicting potential inequalities in the context of ChatGPT
Figure 1 The conceptual model depicting potential inequalities in the context of ChatGPTFrom the PDF
Results Figure 2. Significant SEM paths in the tested model
Results Figure 2. Significant SEM paths in the tested modelFrom the PDF
Figure 3 Significant paths under low vs. high trust
Figure 3 Significant paths under low vs. high trustFrom the PDF

Verification

Claims the paper makes

11 Valid · 0 Not valid · 34 Not checked · 64 Own results and interpretations

44 more claims

Create a free account to see all 109 claims and to verify the 1 reference still open.

See all claims — free account
  • Probably not a substantial claim. 41
  • It may state two separate claims as one. 30
  • The quote only partly states this claim. 1

Create a free account to see these claims.

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.

3 of 3

Citations

Cites

Show all 61 references

Research works printed but not linked: 2

  1. Atlas S (2023) ChatGPT for Higher Education and Professional Development: A Guide to Conversational AI. Kingston, RI: University of Rhode Island.
  2. Van Dijk J (2020) The Digital Divide. Hoboken, NJ: John Wiley.

Cited by

See all 47

Metadata

Authors
Zhang, C. (., Rice, R. E., & Wang, L. H.
Affiliations
  • University of California, Santa Barbara, USA
  • University of California, Santa Barbara (US)
Venue
New Media & Society
Published
2024-12-05
Volume
28
Issue
2
Pages
673-695
ISSN
  • 1461-4448 (ISSN-L)
  • 1461-7315
OpenAlex
W4405083449
Cited by
47
References
61
Citations by year
2026: 36 · 2025: 11
FWCI
17.97
Type
journal article
Language
en
Open access
Licence
Locations
  • New Media & Society (publishedVersion, cc-by-nc)
  • eScholarship (California Digital Library) (submittedVersion, cc-by-nc-nd)
Funders
The Arthur N. Rupe Endowed Professorship in the Social Effects of Mass Communication
Topics
  • Social Sciences › Education › Artificial Intelligence in Education
  • Social Sciences › Education › Digital literacy in education
  • Social Sciences › Gender Studies › Gender and Technology in Education
Keywords
ChatGPT, generative AI, AI literacy, digital literacy, socioeconomic status, digital divide, trust in AI, university students, academic self-efficacy, creativity, academic support, educational outcomes

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.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-02, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-02 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

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: College students’ literacy, ChatGPT activities, educational outcomes, and trust from a digital divide perspective. https://app.cochat.ai/papers/10.1177/14614448241301741
The paper
Zhang, C. (., Rice, R. E., & Wang, L. H. (2026). College students’ literacy, ChatGPT activities, educational outcomes, and trust from a digital divide perspective. New Media & Society. https://doi.org/10.1177/14614448241301741

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.