CoChat

The DOI matches its Crossref record; 84 of the 88 references it cites were found in a registry or an index, and 4 are not checked yet; none of the 51 sources its claims cite is checked for support yet.

A Meta‐Analysis and Meta‐Regression of Incidental Second Language Word Learning from Spoken Input

Open access

Abstract

We meta‐analyzed the effectiveness of incidental second language word learning from spoken input. Our sample contained 105 effect sizes from 32 primary studies employing meaning‐focused word‐learning activities with 1,964 participants with typical cognitive functioning. The random‐effects meta‐analysis yielded a mean effect size of g = 1.05, reflecting generally large vocabulary gains from spoken input in meaning‐focused activities. A meta‐regression with three substantive and two methodological predictors also revealed that adult participants outperformed children in terms of word learning and that interactive learning tasks were more effective than noninteractive ones. Furthermore, learning scores were higher when measured with recognition than with recall tests. Methodologically, the use of a no‐input control group seemed to protect against an overestimation of learning effects, evidenced by smaller effect sizes. Finally, whether a pretest–posttest design was used did not influence effect sizes. All data and the analysis script are publicly available. Open Practices: This article has been awarded an Open Data badge. All data and the analysis script are publicly accessible via the Open Science Framework at https://osf.io/92vfw . Learn more about the Open Practices badges from the Center for Open Science: https://osf.io/tvyxz/wiki .

Checked 2026-11-19 · 4 references not checked yet

Overview

CoChat Summary

TL;DR

This meta-analysis finds that meaning-focused exposure to spoken second language input yields large vocabulary gains, with a mean effect size of Hedges’ g = 1.05. Meta-regression shows advantages for university over younger learners, interactive over noninteractive tasks, recognition over recall tests, and smaller effects when a true no-input control group is used.

Key findings

  • Across 105 effect sizes from 32 primary studies and 1,964 participants, the random-effects meta-analysis yielded a mean effect size of g = 1.05, 95% CI [0.81, 1.28], SE = 0.12, z = 8.77, p < .0001, described as a large effect.
  • University participants outperformed kindergarten/elementary school participants by 0.92 standard deviation units; the high school versus kindergarten/elementary difference was nonsignificant, β = 0.18, 95% CI [−0.78, 1.13], p = .72.
  • Task/+interaction treatments were more effective than task/−interaction treatments, β = 0.63, 95% CI [0.26, 1.00], SE = 0.19, z = 3.37, p < .001; all other treatment contrasts were nonsignificant (all ps > .11).
  • Recognition tests produced effect sizes 0.42 standard deviation units higher than recall tests, and studies with a true control group produced effect sizes 0.47 standard deviation units lower than studies without one; gain scores showed no significant effect.

Methods

  • Meta-analysis and meta-regression of 32 primary studies (from 30 sources; 24 peer reviewed) with 44 independent treatment groups, eight true control groups, 105 effect sizes, and 1,964 participants; studies were located through four electronic databases searched through August 2017 plus manual reference and journal searches.
  • Included studies had to involve spoken-only target word input embedded in a meaning-focused activity, typical cognitive functioning, controlled preexisting knowledge, and sufficient data to calculate standardized effect sizes.
  • Hedges’ g was used as the effect-size measure; random-effects models included random intercepts at study and sample levels, and a mixed-effects meta-regression tested five predictors: age group, treatment type, mode of testing, use of gain scores, and use of a true control group. Three influential cases with Cook’s distance > 0.85 were excluded from the final meta-regression.

Limitations

  • Stated: the authors depended on primary-study reporting; some relevant studies could not be included because means or standard deviations were missing, and correlations between repeated tests were often not reported. They also could not include exposure frequency, retention interval, or proficiency as predictors, and the age effect is confounded with education type, L2 experience/proficiency, motivation, and possible IQ differences.
  • Stated: the study originally aimed to examine incidental learning but often could not determine whether participants expected a posttest, so it analyzed meaning-focused learning instead; only about a quarter of studies included a true control group.
  • Evident from the text: meta-regression cannot establish causal effects of the predictors, and several treatment and age contrasts had small samples or were not directly manipulated within studies, limiting precision and generalizability.

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

Integrity signals

Registry

Crossref · 10.1111/lang.12296

Cites

84 of 88 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

0 Valid · 1 Not valid · 35 Not checked · 94 Own results and interpretations

65 more claims

Create a free account to see all 130 claims and to verify the 32 references still open.

See all claims — free account
  • It may state two separate claims as one. 31
  • Probably not a substantial claim. 41

Create a free account to see these claims.

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?

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.

  • Consistentz = 0.36reported p = .72 · recomputed p = .719

    “z = 0.36, p = .72”

  • Consistentz = 3.37reported p < .001 · recomputed p < .001

    “z = 3.37, p < .001”

  • Consistentχ²(1) = 461.74reported p < .0001 · recomputed p < .001

    “X 2 (1) = 461.74,p < .0001”

  • Consistentχ²(1) = 536.03reported p < .0001 · recomputed p < .001

    “X 2 (1) = 536.03,p < .0001”

  • Consistentz = 8.77reported p < .0001 · recomputed p < .001

    “z = 8.77, p < .0001”

5 of 5

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.

1 of 1

Citations

Cites

Show all 85 references

Research works printed but not linked: 2

  1. *Donkaewbua S., 2009, Vocabulary learning through listening in another language
  2. *Toya M., 1993, Occasional Paper #23

Cited by

See all 110

Metadata

Authors
de Vos, J. F., Schriefers, H., Nivard, M. G., & Lemhöfer, K.
Affiliations
  • Radboud University Nijmegen (NL)
  • Vrije Universiteit Amsterdam (NL)
Corresponding author
de Vos, Johanna F.
Venue
Language Learning
Publisher
Wiley
Published
2018-06-11
Volume
68
Issue
4
Pages
906-941
ISSN
  • 0023-8333 (ISSN-L)
  • 1467-9922
OpenAlex
W2807788276
Cited by
110
References
85
Citations by year
2026: 7 · 2025: 5 · 2024: 12 · 2023: 11 · 2022: 49 · 2021: 11 · 2020: 8 · 2019: 7
FWCI
8.68
Type
journal article
Language
en
Open access
Locations
  • Language Learning (publishedVersion, cc-by)
  • VU Research Portal (publishedVersion, other-oa)
  • MPG.PuRe (Max Planck Society) (submittedVersion, cc-by)
  • Radboud Repository (Radboud University) (submittedVersion)
  • OSF Preprints (OSF Preprints) (submittedVersion, cc-by)
  • OSF Preprints (OSF Preprints) (submittedVersion)
  • Data Archiving and Networked Services (DANS) (submittedVersion, other-oa)
Funders
Nederlandse Organisatie voor Wetenschappelijk Onderzoek (276-89-004)
Topics
  • Social Sciences › Psychology › Developmental and Educational Psychology › Second Language Acquisition and Learning
  • Social Sciences › Arts and Humanities › Language and Linguistics › EFL/ESL Teaching and Learning
  • Social Sciences › Arts and Humanities › Language and Linguistics › Subtitles and Audiovisual Media
Keywords
incidental vocabulary learning, second language acquisition, vocabulary acquisition, interactive learning activities, recall versus recognition, age differences, pretest-posttest design, effect size overestimation

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 2026-11-19, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-11-19 (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: A Meta‐Analysis and Meta‐Regression of Incidental Second Language Word Learning from Spoken Input. https://app.cochat.ai/papers/10.1111/lang.12296
The paper
de Vos, J. F., Schriefers, H., Nivard, M. G., & Lemhöfer, K. (2018). A Meta‐Analysis and Meta‐Regression of Incidental Second Language Word Learning from Spoken Input. Language Learning. https://doi.org/10.1111/lang.12296

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.