Cited by 64EFL/ESL Teaching and LearningEnglish Language Learning and TeachingIntelligent Tutoring Systems and Adaptive Learning

Does Chatting with Chatbots Improve Language Learning Performance? A Meta-Analysis of Chatbot-Assisted Language Learning

Wang, Feifei, Cheung, Alan C. K., Neitzel, Amanda J., Chai, Ching Sing·2025·Review of Educational Research

Abstract

Given the importance of conversation practice in language learning, chatbots, especially ChatGPT, have attracted considerable attention for their ability to converse with learners using natural language. This review contributes to the literature by examining the currently unclear overall effect of using chatbots on language learning performance and comprehensively identifying important study characteristics that affect the overall effectiveness. We meta-analyzed 70 effect sizes from 28 studies, using robust variance estimation. The effects were assessed based on 18 study characteristics about learners, chatbots, learning objectives, context, communication/interaction, and methodological and pedagogical designs. Results indicated that using chatbots produced a positive overall effect on language learning performance ( g = 0.484), compared to nonchatbot conditions. Moreover, four characteristics (i.e., educational level, language level, interface design, and interaction capability) affected the overall effectiveness. In an in-depth discussion on how the 18 characteristics are related to the effectiveness, future implications for practice and research are presented.

Integrity signals

Registry

Crossref · 10.3102/00346543241255621

Cites

92 of 92 references verified0 citation errors · 0 not checkable

Statistics

Recomputing p-values needs the paper’s full text.

Claims the paper makes

This paper’s full text isn’t openly available, so its claims and self-check weren’t run. The registry and citation checks still apply.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • The prevalence of Conversational and NLP Agents reflects sustained interest in dialogue-based learning tools for language acquisition and tutoring.

    “The prevalence of Conversational and NLP Agents reflects sustained interest in dialogue-based learning tools for language acquisition and tutoring, consistent with prior findings on chatbot-mediated instruction [29].”

Cites

Flagged first, then by citations

1 of the 93 references couldn’t be matched to a known paper.

Cited by

Sorted by citations

Authors
Wang, F., Cheung, A. C. K., Neitzel, A. J., & Chai, C. S.
Affiliations
  • Chinese University of Hong Kong (HK)
  • Johns Hopkins University (US)
Corresponding author
Wang, Feifei
Venue
Review of Educational Research
Published
2024-06-14
Volume
95
Issue
4
Pages
623-660
ISSN
  • 0034-6543 (ISSN-L)
  • 1935-1046
OpenAlex
W4399681132
Cited by
64
References
93
Citations by year
2026: 31 · 2025: 30 · 2024: 3
FWCI
14.84
Type
journal article
Language
en
Open access
closed
Locations
  • Review of Educational Research (publishedVersion)
  • The HKU Scholars Hub (University of Hong Kong) (submittedVersion)
Topics
  • Social Sciences › Arts and Humanities › Language and Linguistics › EFL/ESL Teaching and Learning
  • Physical Sciences › Computer Science › Information Systems › English Language Learning and Teaching
  • Physical Sciences › Computer Science › Artificial Intelligence › Intelligent Tutoring Systems and Adaptive Learning
Keywords
chatbots, chatbot-assisted language learning, ChatGPT, language learning performance, conversation practice, interface design, learner characteristics, language proficiency levels, educational attainment

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.

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