12,445 papers · continuously updated · last export: 10 Aug 2026livingmeta.ai
← Browse all papers
AI evidence extraction

The Impact of Chatgpt on Students’ Academic Writing Performance: A Meta-Analytic Study

Jinalyn B. Alvizo · Journal of Educational Analytics · 2025

AI-generated evidence extraction, verified across multiple analytical personas. Not a substitute for the peer-reviewed original.

8/10
Relevance
0/4
Quality (LMQS)
E
Evidence
0
Citations
0.00
FWCI

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.55927/jeda.v4i4.381

Methodology & findings

Study design

Systematic review with meta-analytic quantitative synthesis.

Primary method

Meta-analytic synthesis with pooled effect size calculation. Specific statistical software and heterogeneity testing methods are not reported in the abstract.

Main result

The study found that "the pooled effect size of 2.18 means it is a large positive effect of ChatGPT integration on students' writing performance, indicating improvements in grammar, coherence, and content elaboration."

Reports effect sizes.

Research paradigm

positivist

Author conclusions

The authors concluded that "ChatGPT is beneficial supplementary tool for improving students' writing skills when integrated into instructional design." They further recommended "structured implementation of ChatGPT in writing courses, professional development for instructors on AI integration, and emphasis on ethical use to uphold academic integrity."

Risk of bias

variations in study design across included studies; variations in assessment measures; potential publication bias not explicitly assessed; limited time window (2023-2025) may introduce recency bias; Variations in assessment measures across studies; Potential publication bias (not explicitly assessed); Limited timeframe of included studies (2023-2025); Publication bias not explicitly reported as assessed; Limited temporal range (2023-2025) may introduce recency bias; Variations in study design and assessment measures across included studies; Potential selection bias in inclusion criteria not fully detailed

Limitations

  • "Variations in study design, duration, and assessment measures suggest the need for further longitudinal research."

Open questions raised

  • The need for further longitudinal research examining variations in study design, duration, and assessment measures. Future research should evaluate long-term effects of ChatGPT integration and consistency across assessment methodologies.
  • The need for further longitudinal research to address variations in study design, duration, and assessment measures across studies examining ChatGPT's impact on writing performance.
  • Need for further longitudinal research
  • Need for standardized assessment measures across studies
  • Need for more consistent study design protocols
Data: not_statedCode: not_statedExtracted from: pdfAgreement 74%

Explore related topics

AI Manuscript WritingAcademic WritingMeta-Analysis

Related papers

← Browse all papers