CoChat
Sign Up
Cited by 135Anxiety, Depression, Psychometrics, Treatment, Cognitive ProcessesVirtual Reality Applications and ImpactsDigital Mental Health Interventions

Meta-analysis of the efficacy of virtual reality exposure therapy for social anxiety

Open access
Chesham, Rachel K., Malouff, John M., Schutte, Nicola S.·2018·Behaviour Change
DOI 10.1017/bec.2018.15
Open this paper in CoChat — free accountOpen PDF

Abstract

Social anxiety is a common, debilitating psychological problem. In the present study, two meta-analyses examined the efficacy of virtual reality exposure therapy for social anxiety. The first meta-analysis tested whether virtual reality exposure therapy reduces social anxiety more than a waitlist control condition. The results of the first meta-analysis, consisting of six studies and 233 participants, showed a significant overall effect size, indicating that virtual reality exposure therapy was effective in reducing social anxiety. The second meta-analysis tested whether the standard treatment for social anxiety, which includes in vivo or imaginal exposure, leads to greater effects than virtual reality exposure therapy. The second meta-analysis, consisting of seven studies and 340 total participants, showed essentially no difference in effect sizes between virtual reality exposure and in vivo or imaginal exposure. The results of the two meta-analyses support the use of virtual reality in the treatment of social anxiety.

AI snapshot

Written from the paper’s full text.

TL;DR

Two meta-analyses found that virtual reality exposure therapy (VRET) reduced social anxiety more than waitlist control, but VRET did not differ significantly from standard in vivo or imaginal exposure. The authors conclude the results support VRET as an efficacious treatment for social anxiety.

Key findings

  • VRET versus waitlist: across six studies and 233 participants, the overall effect was significant, Hedges’ g = 0.82, p < .001, 95% CI [0.49, 1.15]. In the random-assignment-only analysis, four studies and 161 participants yielded g = 0.92, p < .001, 95% CI [0.44, 1.41].
  • Publication-bias adjustment for VRET versus waitlist: trim-and-fill imputed one missing study, reducing the effect to g = 0.71, 95% CI [0.34, 1.09]; for random-assignment studies, the adjusted effect was g = 0.76, 95% CI [0.25, 1.27]. Orwin’s fail-safe N indicated 18 missing studies with g = 0 would be needed to reduce the overall effect to a trivial level (g = 0.2), and 14 for the random-assignment analysis.
  • VRET versus in vivo or imaginal exposure: across seven studies and 340 participants, the effect was extremely small and non-significant, g = -0.01, p = .955, 95% CI [-0.30, 0.28]. In the random-assignment-only analysis, five studies and 255 participants yielded g = 0.06, p = .770, 95% CI [-0.32, 0.43].
  • Heterogeneity was not significant in either main comparison: VRET versus waitlist Q = 7.40, df(5), p = .193, I² = 32.40; VRET versus standard treatments Q = 10.68, df(6), p = .099, I² = 43.83.

Methods

  • Two separate meta-analyses were conducted. The authors searched PubMed, EBSCOhost, Proquest Central, PsychINFO, and Proquest Dissertations and Theses in July 2017, with no date limits, plus Google Scholar, reference lists, and author contact; studies had to be in English.
  • Inclusion required a well-controlled trial with random, quasi-random, or matched assignment; a VRET condition and at least one waitlist or in vivo/imaginal exposure comparison; social anxiety as the main target; and usable social anxiety outcome data. The search yielded nine relevant studies with 573 total participants.
  • Effect sizes used Hedges’ g in random-effects models. Heterogeneity was assessed with Q and I². Publication bias was assessed with Orwin’s fail-safe N, Duval and Tweedie’s trim-and-fill, and funnel plots. A second set of analyses included only trials using random assignment.

Limitations

  • The paper states that the analysis was still based on a small number of studies and that additional relevant studies may exist despite the comprehensive search; including them might lead to different results. It also notes that Orwin’s fail-safe N indicated many null-effect studies would be needed to change the conclusion about VRET versus no treatment.
  • The paper states that not every included study used random assignment, making those studies more susceptible to confounding bias; it addressed this by running separate analyses restricted to random-assignment trials.
  • Evident from the text: the VRET-versus-waitlist funnel plot showed moderate asymmetry and trim-and-fill indicated publication bias, with the adjusted effect smaller than the unadjusted effect. The paper also notes that a non-significant difference between VRET and standard treatments does not by itself prove a lack of difference.

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

Integrity signals

Registry

Crossref · 10.1017/bec.2018.15

Cites

69 of 69 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 36 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    In the first meta-analysis restricted to random-assignment studies, the set of studies was not significantly heterogeneous, Q = 6.51, df(3), p = .089.

    “A Q-value of 6.51, df(3), p = .089,indicated that the set of studies were not significantly heterogeneous, I 2 = 53.90.”
    Quote in the paperNot a substantial claim
  • Findingown data

    In the first meta-analysis including all six studies, there was no significant heterogeneity across studies, Q = 7.40, df(5), p = .193.

    “A Q-value of 7.40, df(5), p = .193,indicated no significant heterogeneity across the studies.”
    Quote in the paperSubstantial claim
  • Findingown data

    In the first meta-analysis restricted to random-assignment studies, I2 was 53.90.

    “A Q-value of 6.51, df(3), p = .089,indicated that the set of studies were not significantly heterogeneous, I 2 = 53.90.”
    Quote in the paperNot a substantial claim
  • Findingown data

    In the first meta-analysis including all six studies, I2 was 32.40, suggesting that 32% of variance was due to heterogeneity across studies rather than to chance.

    “The I 2 value of 32.40 suggested that 32% of variance was due to heterogeneity across studies rather than to chance.”
    Quote in the paperNot a substantial claim
  • Findingown data

    In the second meta-analysis including all seven studies, there was no significant heterogeneity across studies, Q = 10.68, df(6), p = .099.

    “A Q-value of 10.68, df(6), p = .099,I 2 = 43.83,indicated no significant heterogeneity across the studies.”
    Quote in the paperSubstantial claim
  • May combine two claims. 23
  • The quote only partly states this claim. 2
  • Classified as probably not a substantial claim. 12

  • May combine two claims.

    Findingown data

    In the second meta-analysis, there was essentially no difference in effect sizes between virtual reality exposure therapy and in vivo or imaginal exposure, with a non-significant overall effect of g = -0.01, p = .955, 95% CI [-0.30, 0.28], across seven studies and 340 participants.

    “The second meta-analysis, consisting of seven studies and 340 total participants, showed essentially no difference in effect sizes between virtual reality exposure and in vivo or imaginal exposure.”
    • Also stated in: § Virtual Reality Exposure Therapy Compared to Standard Treatments · plain
      “There were seven studies and a total of 340 participants included in the analysis of all studies comparing VRET to in vivo or imaginal exposure.The overall effect was in the direction of standard treatments; however, the effect size was extremely small and non-significant, with g = -0.01,p = .955,95% CI [-0.30, 0.28].”
    • Also stated in: § Discussion · plain
      “Results of the second meta-analysis, consisting of seven studies and 340 participants, showed no significant differences in effect sizes when comparing VRET to in vivo or imaginal exposure.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    In the second meta-analysis restricted to the five studies with random assignment, the overall effect was in favour of virtual reality exposure therapy but was non-significant, with g = 0.06, p = .770, 95% CI [-0.32, 0.43], across 255 participants.

    “When only considering studies with random assignment, the overall effect across five studies and 255 participants was in favour of VRET, with g = 0.06, p = .770,95% CI [-0.32, 0.43].”
    Quote in the paperNot a substantial claim

  • May combine two claims. The quote only partly states this claim.

    Findingown data

    In the first meta-analysis, virtual reality exposure therapy reduced social anxiety more than a waitlist control, with a significant overall effect size of g = 0.82, p < .001, 95% CI [0.49, 1.15], across six studies and 233 participants.

    “The results of the first meta-analysis, consisting of six studies and 233 participants, showed a significant overall effect size, indicating that virtual reality exposure therapy was effective in reducing social anxiety.”
    • Also stated in: § Results · plain
      “The overall effect size across the six studies and their 233 participants was significant, with g = 0.82, p < .001,95% CI [0.49, 1.15].”
    • Also stated in: § Discussion · plain
      “The results of the first meta-analysis, consisting of six studies and 233 participants, showed a medium-to-large, significant overall effect size, indicating that VRET was more efficacious than waitlist.”
    Quote in the paperSubstantial claim

  • The quote only partly states this claim.

    Findingown data

    In the first meta-analysis restricted to the four studies with random assignment, virtual reality exposure therapy was significantly more efficacious than waitlist, with g = 0.92, p < .001, 95% CI [0.44, 1.41], across 161 participants.

    “With the analysis limited to the four studies with random assignment, the overall effect size across a total of 161 participants was significant, with g = 0.92, p < .001,95% CI [0.44, 1.41].”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingcitation

    In the first meta-analysis restricted to random-assignment studies, trim and fill indicated bias and imputed one missing study, yielding an adjusted effect size of g = 0.76, 95% CI [0.25, 1.27].

    “Duvall and Tweedie's (1998) trim and fill indicated bias and imputed one missing study into the funnel plot.Taking this bias into account, the resulting effect size fell to g = 0.76, 95% CI [0.25, 1.27].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Duvall and Tweedie's (1998)

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    In the second meta-analysis, for both the seven-study and five-study random-assignment analyses, no asymmetry was detected in the funnel plots of effect sizes.

    “There was no asymmetry detected in the funnel plots of effect sizes.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    In the second meta-analysis, for both the seven-study and five-study random-assignment analyses, Duvall and Tweedie's trim and fill did not impute any missing studies.

    “Further, Duvall and Tweedie's (1998) trim and fill did not impute any missing studies.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Duvall and Tweedie's (1998)

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    In the first meta-analysis including all six studies, the funnel plot showed moderate asymmetry of effect sizes, suggesting potential publication bias.

    “The funnel plot (Figure 3) showed moderate asymmetry of effect sizes for the six studies, suggesting potential publication bias.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingcitation

    In the first meta-analysis including all six studies, trim and fill indicated publication bias and added one missing study, yielding an adjusted effect size of g = 0.71, 95% CI [0.34, 1.09].

    “Duvall and Tweedie's (1998) trim and fill also indicated publication bias and corrected for this by adding one missing study to the plot, leading to an adjusted effect size of g = 0.71, 95% CI [0.34, 1.09].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Duvall and Tweedie's (1998)

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    In the first meta-analysis restricted to random-assignment studies, the funnel plot showed minor to moderate asymmetry, suggesting some publication bias.

    “With only studies that used random assignment, the funnel plot showed minor to moderate asymmetry, suggesting some publication bias (Figure 4).”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Findingown data

    The search strategy resulted in nine relevant studies, with a total of 573 participants.

    “The search strategy resulted in nine relevant studies, with a total of 573 participants.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    In the second meta-analysis, Orwin's fail-safe N analysis was not useful because the Hedges' g values were already in the trivial category.

    “Orwin's failsafe N analysis was not useful because the Hedges' g values were already in the trivial category.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Orwin's (1983)

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Previous meta-analytic studies of the effects of virtual reality exposure therapy for social anxiety had two main deficiencies: a very low number of studies and a lack of evaluation of publication bias.

    “In sum, previous meta-analytic studies of the effects of VRET for social anxiety have had two main deficiencies: a very low number of studies and a lack of evaluation of publication bias.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Valentine, Pigott, & Rothstein, 2010

      Not checked in the free run.

    • cites Schmidt & Hunter, 2014

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Opriş et al. (2012) reported a non-significant effect size in favour of in vivo and imaginal exposure (d = 0.13, 95% CI [-0.11, 0.38] for three social phobia studies), suggesting that VRET and standard treatments had comparable efficacy.

    “Opriş et al. (2012) reported a non-significant effect size in favour of in vivo and imaginal exposure (d = 0.13, 95% CI [-0.11, 0.38] for three social phobia studies, suggesting that VRET and standard treatments had comparable efficacy.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Opriş et al. (2012)

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    VRET may be useful in addressing shortcomings of established exposure methods, particularly in relation to cost-effectiveness, convenience, treatment acceptability, treatment availability, and difficulties with clients visualising scenes during imaginal exposure.

    “VRET may be useful in addressing the shortcomings of established methods of exposure, particularly in relation to cost-effectiveness, convenience, treatment acceptability, treatment availability, and difficulties with clients visualising scenes during imaginal exposure (Anderson et al., 2013;Choy, Fyer, & Lipsitz, 2007;Krijn et al., 2004).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Anderson et al., 2013

      Not checked in the free run.

    • cites Choy, Fyer, & Lipsitz, 2007

      Not checked in the free run.

    • cites Krijn et al., 2004

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    A large body of research supports the efficacy of in vivo and imaginal exposure treatments for reducing symptoms of social anxiety, although the methods do not work for all individuals with social anxiety.

    “A large body of research supports the efficacy of in vivo and imaginal exposure treatments for reducing symptoms of social anxiety, although the methods do not work for all individuals with social anxiety (e.g., Acarturk, Cuijpers, van Straten, & de Graaf, 2009;Feske & Chambless, 1995;Gould, Buckminster, Pollack, Otto, & Massachusetts, 1997;Powers, Sigmarsson, & Emmelkamp, 2008;Turner, Beidel, &”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Acarturk, Cuijpers, van Straten, & de Graaf, 2009

      Not checked in the free run.

    • cites Feske & Chambless, 1995

      Not checked in the free run.

    • cites Gould, Buckminster, Pollack, Otto, & Massachusetts, 1997

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Parsons and Rizzo (2008) reported a meta-analytic pre-post effect size of 0.96, 95% CI [0.34, 1.59], indicating a reduction in anxiety symptoms following VRET for social phobia.

    “Parsons and Rizzo (2008) investigated the pre-post effects of VRET for social phobia and reported a meta-analytic effect size of 0.96, 95% CI [0.34, 1.59], indicating a reduction in anxiety symptoms following VRET.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Parsons and Rizzo (2008)

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Powers and Emmelkamp (2008) reported a between-groups meta-analytic Hedges' g of 0.73, indicating that VRET was more efficacious than control conditions in the treatment of social anxiety.

    “They reported a between-groups, meta-analytic Hedges' g of 0.73 (no reported confidence interval), indicating that VRET was more efficacious than control conditions in the treatment of social anxiety.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Powers and Emmelkamp (2008)

      Not checked in the free run.

  • May combine two claims.

    Premiseown data

    In prior meta-analyses, the number of social anxiety studies in subanalyses ranged from two to four, and the total number of participants ranged from 50 to 216.

    “Within these subanalyses, the number of studies on social anxiety ranged from two to four, and the total number of participants ranged from 50 to 216.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    Kampmann, Emmelkamp, and Morina (2016) reported a large and significant effect size across three studies, with g = 0.82, 95% CI [0.13, 1.51], indicating that VRET outperformed waitlist.

    “Kampmann, Emmelkamp, and Morina (2016) reported a similarly large and significant effect size across three studies, with g = 0.82, 95% CI [0.13, 1.51].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Kampmann, Emmelkamp, and Morina (2016)

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    The meta-analysis included a second set of analyses restricted to trials using random assignment.

    “So, we conducted a second set of analyses including only trials using random assignment.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data

    The third author independently checked the coding by randomly selecting four studies, which included 25 lines of data and 275 entries, and the agreement rate was 96%.

    “The third author independently checked the coding by randomly selecting four studies, which included 25 lines of data (out of a total of 50 lines) and 275 entries.For continuous data entries, we considered an agreement to be a variance of anything less than 5% of the original coded value.The agreement rate was 96%.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Opriş et al. (2012) reported d = 1.01, 95% CI [0.69, 1.33] across three social phobia studies, indicating that VRET outperformed waitlist on outcome measures of social anxiety.

    “Opriş et al. (2012) reported a d of 1.01, 95% CI [0.69, 1.33] across three social phobia studies, indicating VRET outperformed waitlist on outcome measures of social anxiety.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Opriş et al. (2012)

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationnone given

    These potential advantages include greater acceptability, along with easier provision for some clients of varied social situations than is possible in vivo or in imagery.

    “These potential advantages include greater acceptability, along with easier provision for some clients of varied social situations than is possible in vivo or in imagery.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationcitation

    This finding is in line with results of the meta-analyses of Kampmann, Emmelkamp, and Morina (2016) and Opriş et al. (2012).

    “This finding is in line with results of the meta-analyses of Kampmann, Emmelkamp, and Morina (2016) and Opriş et al. (2012).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Kampmann, Emmelkamp, and Morina (2016)

      Not checked in the free run.

    • cites Opriş et al. (2012)

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationreasoning

    VR exposure treatment has potential advantages over the usual exposure in vivo or in imagery.

    “VR exposure treatment has potential advantages over the usual exposure in vivo or in imagery.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Clinicians frequently use meta-analyses as a resource to determine best treatment options for clients.

    “Clinicians frequently use meta-analyses as a resource to determine best treatment options for clients (Petitti, 2000)”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Petitti, 2000

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    The present study's findings may provide a more methodologically sound basis for informing treatment decisions.

    “the present study's findings may provide a more methodologically sound basis for informing treatment decisions.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Interpretationown data

    These results may be useful for clients and clinicians seeking evidence to inform treatment decisions.

    “These results may be useful for clients and clinicians seeking evidence to inform treatment decisions.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationcitation

    The fact that the second hypothesis was not supported might be explained by the 'dodo bird conjecture'.

    “The fact that the second hypothesis was not supported might be explained by the 'dodo bird conjecture', which postulates that when bona fide psychotherapies are compared, they produce equivalent outcomes (Wampold et al., 1997).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Wampold et al., 1997

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    The findings provide support for the extinction aspect of learning theory, the corrective learning element of emotional processing theory, and the inhibitory learning model.

    “Further, the findings provide support for the extinction aspect of learning theory, for the corrective learning element of emotional processing theory (Foa & Kozak, 1986), and for the inhibitory learning model (Craske et al., 2008).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Foa & Kozak, 1986

      Not checked in the free run.

    • cites Craske et al., 2008

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationcitation

    The finding that VRET is superior to waitlist is consistent with smaller past meta-analyses by Opriş et al. (2012) and Kampmann, Emmelkamp, and Morina (2016).

    “This finding was consistent with smaller past meta-analyses by Opriş et al. (2012) and Kampmann, Emmelkamp, and Morina (2016).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Opriş et al. (2012)

      Not checked in the free run.

    • cites Kampmann, Emmelkamp, and Morina (2016)

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    These advantages make the results more reliable and generalisable.

    “These advantages make the results more reliable and generalisable.”
    Quote in the paperNot a substantial claim
33 of 33

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?

Cites

Flagged first, then by citations

Research works printed but not linked: 2

  1. Oskam P (2005) Virtual reality exposure therapy (VRET) effectiveness and improvement. Paper presented at the 2nd Twente University Student Conference on IT, Enschede, The Netherlands.
  2. 2013, Diagnostic and statistical manual of mental disorders

Cited by

Sorted by citations

Related

Authors
Chesham, R. K., Malouff, J. M., & Schutte, N. S.
Affiliations
University of New England (AU)
Corresponding author
Malouff, John M.
ORCID
Venue
Behaviour Change
Publisher
Cambridge University Press (CUP)
Published
2018-06-18
Volume
35
Issue
3
Pages
152-166
ISSN
  • 0813-4839 (ISSN-L)
  • 2049-7768
DOI
10.1017/bec.2018.15
OpenAlex
W2809485580
Semantic Scholar
bff2e781c1b54f6d490fccf05b303291cf4655ec
Cited by
135
References
66
Citations by year
2026: 18 · 2025: 16 · 2024: 20 · 2023: 16 · 2022: 20 · 2021: 22 · 2020: 15 · 2019: 7
FWCI
1.75
Type
journal article
Language
en
Open access
Licence
www.cambridge.org/core/terms (unspecified)
Locations
  • Behaviour Change (publishedVersion)
  • RUNE (Research UNE) (submittedVersion)
Topics
  • Social Sciences › Psychology › Experimental and Cognitive Psychology › Anxiety, Depression, Psychometrics, Treatment, Cognitive Processes
  • Physical Sciences › Computer Science › Human-Computer Interaction › Virtual Reality Applications and Impacts
  • Social Sciences › Psychology › Applied Psychology › Digital Mental Health Interventions
Keywords
virtual reality exposure therapy, social anxiety, in vivo exposure, imaginal exposure, treatment efficacy, waitlist control, psychotherapy comparison

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.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-10.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-10.

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