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

The DOI matches its Crossref record; 28 of the 28 references it cites were found in a registry or an index; 12 of the 40 sources its claims cite were checked for support.

Summarise, elaborate, try again: exploring the effect of feedback literacy on AI-enhanced essay writing

Open access
Hawkins, Brittany, Taylor-Griffiths, Daniel, Lodge, Jason M. · 2026 · Assessment & Evaluation in Higher Education

Abstract

Conservative estimates suggest that since ChatGPT was released in November 2022, more than 70% of university students are using AI in their studies. In particular, students have reported the benefits of using generative AI for real-time, personalised feedback. This study explores how students use generative AI for feedback using the novel methodology of a simulated assessment task. In individual, 1-hour sessions, 32 participants completed a screen-recorded, 25-minute essay writing task. Participants were asked to use generative AI to enhance their work. Following a brief questionnaire, participants completed a video-stimulated interview. While watching the essay screen recording, participants discussed their process, specifically focusing on how AI was used. A thematic analysis of transcribed interviews identified four distinct themes: feed-forward (initial content requests), feedback (requesting assessments of own work), feedback evaluation (making decisions based on AI output, including providing feedback to AI), and AI avoidance (deliberately not using AI). A multiple regression found feedback literacy to be a significant predictor of essay task performance. This exploratory study demonstrates the use of a novel methodology to help understand feedback literacy and evaluative judgement skills as students become equipped for interacting with AI.

Overview

CoChat Summary

TL;DR

In a simulated 25-minute essay task with 32 psychology students, feedback literacy was the only significant positive predictor of essay grade in a multiple regression, although the overall model was non-significant. Thematic analysis identified four themes in how students used and avoided generative AI for feedback: feedforward, feedback, feedback evaluation, and AI avoidance.

Key findings

  • The overall regression model accounted for a non-significant 35% of variance in essay grade, F(6, 25) = 2.26, p = .071. Feedback literacy was the only significant positive predictor of essay grade, β = 0.46, t(25) = 2.56, p = 0.017, 95% CI [0.09, 0.83], sr² = .17; feedback literacy and essay grade were also positively and significantly correlated (p = 0.019; Table 1 r = .41*).
  • All 32 participants engaged with AI at least once; all used ChatGPT, with one participant using their own premium account. Less than 20% of participants explicitly asked AI for overall essay feedback, and most feedback requests were for line-level language improvements such as paraphrasing, synonyms, and grammar.
  • All but one participant evaluated the accuracy or usefulness of AI content at least once. Half of the participants also expressed active attempts to avoid AI use during certain points of the task.
  • Thematic analysis produced four distinct themes: feedforward (initial content requests), feedback (requesting assessments of own work), feedback evaluation (making decisions based on AI output, including providing feedback to AI), and AI avoidance (deliberately not using AI).

Methods

  • Thirty-two psychology students (21 female, 11 male) at the University of Queensland participated; ages ranged from 17 to 46 (M = 21.5, SD = 6.1), and 68.75% were in their first year.
  • In individual 60-minute sessions, participants completed a screen-recorded, 25-minute essay task with unrestricted internet access and ChatGPT (free-tier) provided under researcher credentials. After the essay, they completed a questionnaire and a video-stimulated interview while watching back the screen recording.
  • Essays were graded blindly on criteria scored 0–10, summed, and converted to a percentage. A standard multiple regression tested essay grade against prior knowledge, squared task confidence, task effort, AI frequency, AI trust, and feedback literacy. Interview transcripts were analyzed using Braun and Clarke’s six-phase thematic analysis.

Limitations

  • The authors state three limitations: the statistical analysis was underpowered, so regression conclusions cannot be asserted with confidence; the sample came from the psychology program of a single university and does not represent university students broadly; and the simulated task had fidelity limitations because a real-world assignment would typically involve greater topic knowledge and more time, and students knew their performance would not affect actual grades.
  • The text also makes evident that the overall regression model was non-significant (p = .071) with N = 32, consistent with the authors’ stated concern about being underpowered.
  • The text notes that one researcher completed all transcriptions and data coding; the paper frames this as a strength of reflexive thematic analysis, but it also means coding was not independently checked by a second researcher.

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

Integrity signals

Registry

Crossref · 10.1080/02602938.2025.2492070

Cites

28 of 28 references verified0 citation errors · 0 not checkable

Statistics

Figures

Image 1
CC-BY-NC-ND · From the PDF
Image 2
CC-BY-NC-ND · From the PDF
Image 3
CC-BY-NC-ND · From the PDF
Image 4
CC-BY-NC-ND · From the PDF

Verification

Claims the paper makes

12 Valid · 2 Not valid · 23 Not checked · 88 Own results and interpretations

60 more claims

Create a free account to see all 125 claims and every source verdict.

See all claims — free account
  • Probably not a substantial claim. 43
  • It may state two separate claims as one. 30
  • Borderline: one of the tests for a substantial claim is unclear. 2
  • The quote only partly states this claim. 2

Create a free account to see these claims.

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.

  • ConsistentF(6, 25) = 2.26reported p = .071 · recomputed p = .070

    “F(6, 25) = 2.26, p = .071”

  • Consistentt(25) = 2.56reported p = .017 · recomputed p = .017

    “t(25) = 2.56, p = .017”

  • Consistentt(25) = 0.59reported p = .562 · recomputed p = .560

    “t(25) = 0.59, p = .562”

  • Consistentt(25) = 1.01reported p = .324 · recomputed p = .322

    “t(25) = 1.01, p = .324”

  • Consistentt(25) = -1.58reported p = .127 · recomputed p = .127

    “t(25) = -1.58,p = .127”

5 of 7 · the rest aren’t loaded

2 more recomputed tests

Create a free account to see every test the Stats check recomputed.

See all tests — free account

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

1 of 1

Citations

Cites

Show all 27 references

Cited by

See all 18

Metadata

Authors
Hawkins, B., Taylor-Griffiths, D., & Lodge, J. M.
Affiliations
The University of Queensland (AU)
Corresponding author
Lodge, Jason M.
Venue
Assessment & Evaluation in Higher Education
Published
2025-04-22
Volume
51
Issue
5
Pages
879-891
ISSN
  • 0260-2938 (ISSN-L)
  • 1469-297X
OpenAlex
W4409672219
Cited by
17
References
27
Citations by year
2026: 14 · 2025: 4
FWCI
11.62
Type
journal article
Language
en
Open access
Licence
Locations
Assessment & Evaluation in Higher Education (publishedVersion, cc-by-nc-nd)
Topics
  • Physical Sciences › Computer Science › Computer Science Applications › Online Learning and Analytics
  • Physical Sciences › Computer Science › Artificial Intelligence › Intelligent Tutoring Systems and Adaptive Learning
  • Social Sciences › Psychology › Developmental and Educational Psychology › Innovative Teaching and Learning Methods
Keywords
feedback literacy, generative AI, student use of AI, feedforward, feedback evaluation, university students

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 2026-10-05, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-05 (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: Summarise, elaborate, try again: exploring the effect of feedback literacy on AI-enhanced essay writing. https://app.cochat.ai/papers/10.1080/02602938.2025.2492070
The paper
Hawkins, B., Taylor-Griffiths, D., & Lodge, J. M. (2026). Summarise, elaborate, try again: exploring the effect of feedback literacy on AI-enhanced essay writing. Assessment & Evaluation in Higher Education. https://doi.org/10.1080/02602938.2025.2492070

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.