CoChat
Sign Up
Cited by 12Social and Intergroup PsychologyDecision-Making and Behavioral EconomicsPsychology of Moral and Emotional Judgment

Agency and self-other asymmetries in perceived bias and shortcomings: Replications of the Bias Blind Spot and link to free will beliefs

Open access
Chandrashekar, Subramanya Prasad, Yeung, Siu Kit, Yau, Ka Chai, Cheung, Chung Yee, Agarwal, Tanay Kulbhushan, Wong, Cho Yan Joan, Pillai, Tanishka, Thirlwell, Thea Natasha +6 more·2021·Judgment and Decision Making
DOI 10.1017/s1930297500008470
Open this paper in CoChat — free accountOpen PDF

Abstract

Bias Blind Spot (BBS) is the phenomenon that people tend to perceive themselves as less susceptible to biases than others. In three pre-registered experiments (overall N = 969), we replicated two experiments of the first demonstration of the phenomenon by Pronin et al. (2002). We found support of the BBS hypotheses, with effects in line with findings in the original study: Participants rated themselves as less susceptible to biases than others (d= –1.00 [–1.33, –0.67]). Deviating from the original, we found an unexpected effect that participants rated themselves as having fewer shortcomings (d= –0.34 [–0.46, –0.23]), though there was support for the target’s main premise that BBS was stronger for biases than for shortcomings (d= –0.43 [–0.56, –0.29]). Extending the replications, we found that beliefs in own free will were positively associated with BBS (r ∼ 0.17–0.22) and that beliefs in both own and general free will were positively associated with self-other asymmetry related to personal shortcomings (r∼ 0.16–0.24). Materials, datasets, and code are available on https://osf.io/3df5s/.

AI snapshot

Written from the paper’s full text.

TL;DR

In three preregistered replications of Pronin et al. (2002), participants rated themselves as less susceptible to biases than others, and also—contrary to the original null—as having fewer personal shortcomings, with the bias gap larger than the shortcomings gap. Free will beliefs were associated with stronger self-other asymmetries, especially for personal shortcomings, though these extension findings are correlational.

Key findings

  • Bias blind spot replicated: mini meta-analytic self-other asymmetry in susceptibility to biases was d = −1.00 [−1.33, −0.67]; self-other asymmetry in personal shortcomings was d = −0.34 [−0.46, −0.23]; the difference between bias and shortcomings asymmetries was d = −0.43 [−0.56, −0.29].
  • Better-than-average effects replicated: positive personality dimensions d = 1.05 [0.57, 1.54], negative personality dimensions d = −0.98 [−1.49, −0.47], and denial of the better-than-average effect d = 1.69 [1.49, 1.88]; after being made aware of the bias, 79% in Study 1b and 82% in Study 3 still denied it.
  • Free will beliefs were negatively correlated with perceived personal shortcomings in the self (r ∼ −0.22 to −0.09) and with self-other differences in personal shortcomings (r ∼ −0.24 to −0.16); self-referential free will beliefs were associated with BBS in susceptibility to bias (r ∼ 0.17–0.22), with weaker associations for generalized free will beliefs.
  • Effects were supported across Hong Kong and U.S. samples.

Methods

  • Three preregistered studies; the abstract reports overall N = 969, with analyzed samples reported as Study 1a N = 45, Study 1b N = 43, Study 2 N = 303, and Study 3 N = 621. Study 1a/1b used University of Hong Kong undergraduates; Studies 2/3 used U.S. Amazon Mechanical Turk participants.
  • Within-subject self-versus-others ratings of eight biases and three personal shortcomings, plus better-than-average personality ratings; Studies 2 and 3 added free will belief scales (FWI, FWD, and in Study 3 FAD+).
  • Analyses used one-tailed paired-samples t-tests, chi-squared tests, correlations, and random-effects mini meta-analysis; replication results were evaluated with LeBel et al. criteria.

Limitations

  • The paper states that evidence for the relation between belief in free will and BBS effects is correlational only, so causal inferences are not possible.
  • The paper advises caution in over-interpreting the cross-cultural findings given the small sample and the singular bias under examination.
  • Evident from the text: Studies 1a and 1b had small undergraduate samples (N = 45 and 43), Study 1a was described as slightly underpowered, and the free will extension was tested only in Studies 2 and 3.

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

Integrity signals

Registry

Crossref · 10.1017/s1930297500008470

Cites

55 of 55 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

4 of 21 cited sources checked · 3 supported · 0 flagged · accuracy not yet measured

  • Interpretationreasoning

    Free will beliefs in reference to the self are related to the nature of introspection accessed.

    “This indicates that free will beliefs in reference to the self are related to the nature of introspection accessed, complementing previous work that argued that free will beliefs impact fundamental socio-cognitive processes (Genschow et al., 2017).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Genschow et al., 2017 source
  • Premisecitation

    Free will beliefs were found to be associated with an overestimation of internal factors compared to external factors when evaluating others' behavior.

    “Fourth, free will beliefs were found to be associated with an overestimation of the influence of internal (dispositional and fixed) factors as compared to external (varying and adaptable) factors when evaluating others' behavior (Genschow et al., 2017).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Genschow et al., 2017 source

      The Results report that stronger free-will belief correlated with attributing others’ behavior more to internal than external factors. The paper describes these as person and situation attributions, but the cited passages do not establish that internal factors are fixed or external factors are adaptable. (LLM, conf=0.84)

      Experimental — accuracy not yet measured
  • Findingown data

    Beliefs in own free will were positively associated with BBS (r ∼ 0.17-0.22).

    “Extending the replications, we found that beliefs in own free will were positively associated with BBS (r ∼ 0.17-0.22)”
    Quote in the paperSubstantial claim
  • Findingown data

    Participants rated themselves as less susceptible to biases than others across the three samples.

    “Participants rated themselves as less susceptible to biases than others (d = -1.00[-1.33, -0.67]).”
    • Also stated in: § 3.1 Susceptibility to biases and personal shortcomings · plain
      “We conducted a series of one-tailed paired samples t-tests, and found that participants showed a general tendency to view themselves as less susceptible to biases across the three samples”
    Quote in the paperSubstantial claim
  • Findingown data

    The mini meta-analysis found support for self-other differences in perceived personal shortcomings (d = -0.34 [-0.46, -0.23]).

    “We found support for self-other differences in perceived susceptibility to bias (d = -1.00[-1.33, -0.67]), self-other differences in perceived personal shortcomings (d = -0.34[-0.46, -0.23]), and the effects were stronger for perceived susceptibility to bias compared to perceived personal shortcomings (d = -0.43[-0.56, -0.29]).”
    • Also stated in: § 4.1 Replications of the bias blind spot · plain
      “Using a larger sample of participants, we found support for self-other differences for personal shortcomings, whereas the original article hypothesized and argued findings were in support of a null effect.”
    Quote in the paperSubstantial claim
  • Classified as probably not a substantial claim. 18
  • May combine two claims. 18
  • The quote only partly states this claim. 1

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

    Premisecitation

    Belief in free will is the generalized belief that human behavior is free from internal and external constraints across situations.

    “Belief in free will is the generalized belief that human behavior is free from internal and external constraints across situations (Feldman, 2017;Monroe & Malle, 2014).”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites Feldman, 2017 source
    • cites Monroe & Malle, 2014 source

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    Free will beliefs were measured using the Free Will Inventory (FWI) and the Free Will and Determinism (FWD) personal agency subscale.

    “For the extension, free will beliefs were measured before the replication using two free will belief subscales: Free Will Inventory (FWI) (Nadelhoffer et al., 2014) (5 items; 1 = Strongly disagree, 7 = Strongly agree; = 0.91) and Free Will and Determinism (FWD) personal agency subscale (Rakos et al., 2008) (4 items; 1 = Not true at all, 5 = Almost always true; = 0.92).”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites Nadelhoffer et al., 2014 source
    • cites Rakos et al., 2008 source

  • May combine two claims.

    Findingown data

    In Study 1b, only 9 of 43 participants (21%) acknowledged any potential bias, leaving 79% claiming to be better than their average peers.

    “In Study 1b, we found that despite being made aware of their potential bias, only 9 of the 43 participants (21%) acknowledged any potential bias, leaving 79% of participants still claiming to be better than their average peers”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Stronger perceptions of personal free will were associated with a stronger bias blind spot in susceptibility to bias (r ∼ 0.17-0.22), but associations were much weaker for generalized free will beliefs.

    “These suggested that stronger perceptions of personal free will were associated with stronger bias blind spot, but associations were much weaker for generalized free will beliefs.”
    • Also stated in: § 4.1 Replications of the bias blind spot · plain
      “We also found support for an association between free will beliefs and the BBS in susceptibility to bias (r ∼ 0.17-0.22),but only with for the free will beliefs scale that concerned the self, rather than in general.”
    Quote in the paperSubstantial claim

  • The quote only partly states this claim.

    Findingown data

    In Study 2 and Study 3, participants rated themselves as less prone to personal shortcomings than others.

    “we found an unexpected effect that participants rated themselves as having fewer shortcomings (d = -0.34[-0.46, -0.23])”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Bias blind spot (BBS) is the phenomenon that people perceive themselves to be less susceptible to biases than others.

    “Bias blind spot (BBS) is the phenomenon that people perceive themselves to be less susceptible to biases than others.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Pronin et al. (2002) source

  • May combine two claims.

    Interpretationown data

    The replications provide strong support for the main hypotheses of the target article, with effects comparable to and at times larger than the original's.

    “We conclude strong support for the main hypotheses of the target article, with effects comparable and at times larger than the original's (see Figure 1).”
    • Also stated in: § 4.1 Replications of the bias blind spot · plain
      “Our findings were in support of the main predictions of the BBS effect.”
    Quote in the paperSubstantial claim

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

    Study design and conductnone given

    Study 2 was a separate replication of the bias blind spot effect, designed by a different replication team, with a larger, well-powered sample of different demographics.

    “Study 2 was a separate replication of the BBS effect, designed by a different replication team, and with a larger, well-powered sample of different demographics.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    Across the four studies in Pronin et al. (2002), sample size ranged from 24 to 91, with an average of 55 per study.

    “Across the four studies reported in the original paper, the sample size ranged from 24 to 91, with an average sample size of 55 per study.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Pronin et al. (2002) source

  • May combine two claims.

    Premisecitation

    Study 1 Survey 2 of Pronin et al. (2002) found support that people rate themselves as less susceptible to biases than others, but no support for self-other asymmetry in personal shortcomings.

    “Study 1 Survey 2 of Pronin et al. (2002) found empirical support for the view that people rate themselves, compared with others, to be less susceptible to biases, whereas there was no support for self-other asymmetry in ratings of personal shortcomings.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Pronin et al. (2002) source

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

    Study design and conductnone given

    In each replication study, the experiment was pre-registered on OSF, and data collection was launched within one week of registration.

    “In each of the replication studies, we first pre-registered the experiment on the Open Science Framework (OSF), and data collection was launched within one week of registration.”
    Quote in the paperNot a substantial claim

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

    Study design and conductnone given

    All measures, manipulations, and exclusions for this investigation are reported, and data collection was completed before analyses.

    “All measures, manipulations, and exclusions for this investigation are reported, and data collection was completed before analyses.”
    Quote in the paperNot a substantial claim

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

    Study design and conductnone given

    Eight participants were excluded based on pre-registered criteria: two self-reported not being serious about filling in the survey and six were not from the USA.

    “We excluded 8 participants based on pre-registered criteria: self-reported not being serious about filling in the survey (2), and not from the USA (6).”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    Display of conditions was counterbalanced in a 2 (reference: self vs. others) by 2 (attributes: biases and personal shortcomings) within-subject design.

    “Display of conditions was counterbalanced in a 2 (reference: self vs. others) by 2 (attributes: biases and personal shortcomings) within-subject design (see the Supplementary Materials for more details and full measures).”
    Quote in the paperSubstantial claim

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

    Premisecitation

    Free will beliefs have been found to be associated with academic and job performance, job satisfaction, cooperation, and well-being and meaning in life.

    “Free will beliefs have been found to be associated with a range of behavioral and psychological factors, such as academic and job performance (Feldman et al., 2016;Stillman et al., 2010), job satisfaction (Feldman et al., 2018), cooperation (Protzko et al., 2016), and well-being and meaning in life (Crescioni et al., 2016).”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Feldman et al., 2016

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Stillman et al., 2010 source
    • cites Feldman et al., 2018

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

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

    Study design and conductown data

    In Study 3, eight participants were excluded based on pre-registered criteria, leaving 621 participants in the analysis.

    “Of those, eight participants were excluded based on pre-registered exclusion criteria , leaving 621 participants included in the analysis”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    Free will belief associations are more about viewing oneself as free of biases and shortcomings, and less about viewing others as more susceptible to biases and shortcomings.

    “The results suggest that free will belief associations are more about viewing oneself to be free of biases and shortcomings, and less about viewing others as more susceptible to biases and shortcomings.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductown data

    The evidence for the relation between belief in free will and bias blind spot effects is correlational only, so the current findings do not allow for causal inferences.

    “It is important to note that our evidence for the relation between belief in free will and BBS effects is correlational only.Therefore, current findings do not allow for causal inferences.”
    Quote in the paperSubstantial claim

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

    Interpretationreasoning

    The same power issues may explain why the original study detected the hypothesized effect in only four of the eight self-serving bias items tested.

    “Similarly, the same issues with power may have been the reason why they were able to detect the hypothesized effect in only four of the eight self-serving bias items tested.”
    Quote in the paperNot a substantial claim

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

    Study design and conductown data

    In Study 2, seven participants were excluded based on pre-registered criteria, leaving 303 participants in the analysis.

    “Of those, seven participants were excluded based on pre-registered exclusion criteria , leaving 303 participants included in the analysis”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Pronin et al. (2002) provided the first demonstration of BBS by showing that people tend to perceive themselves as less biased, even when made aware of their potential bias.

    “Pronin et al. (2002) provided the first demonstration of the effect by examining people's ratings of susceptibility to common biases in themselves and in others, and showed that people tend to perceive themselves as being less biased, even when being made aware of their potential bias.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Pronin et al. (2002) source

  • Classified as probably not a substantial claim.

    Premisereasoning

    Pronin et al. (2002) predicted that the self-versus-others asymmetry is larger for biases than for personal shortcomings.

    “Based on this argument, the authors predicted that the self-versus-others asymmetry is larger for biases than for personal shortcomings.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Pronin et al. (2002) source

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

    Premisecitation

    80% of study participants in psychology come from Western, educated, industrialized, rich and democratic (WEIRD) societies.

    “Scholars note that the psychology literature is lopsided, with 80% of study participants in psychology coming from Western, educated, industrialized, rich and democratic (WEIRD) societies (Henrich et al., 2010).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Henrich et al., 2010 source

  • May combine two claims.

    Study design and conductown data

    Study 1's sample was slightly underpowered, yet the effects reported in the classic article were still detected.

    “The sample was slightly underpowered, and yet we were able to detect the effects reported in the classic article.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    The findings support the BBS hypothesis and the original findings.

    “We conclude these findings to be in support of the BBS hypothesis and the original findings.”
    Quote in the paperSubstantial claim

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

    Study design and conductnone given

    Participants who took part in Study 2 were prevented from participating in Study 3.

    “We prevented participants who took part in Study 2 from participating in Study 3.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductown data

    In Study 3 the predictions were pre-registered, and the effects found were very consistent.

    “In Study 2, these findings were discovered in exploratory analyses, and in Study 3 we pre-registered the predictions and found very consistent effects, suggesting that these were not accidental.”
    Quote in the paperSubstantial claim

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

    Premisecitation

    BBS has been theorized to result from three possible sources: introspection illusion, naïve realism, and self-enhancement motives.

    “Bias blind spot has been theorized to be a result of the three possible sources: introspection illusion, naïve realism, and self-enhancement motives (Pronin & Kugler, 2007).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Pronin & Kugler, 2007 source

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    Study 3 used an even larger sample that would allow for detecting smaller effects.

    “Study 3 aimed to replicate and extend the findings, especially regarding the free will beliefs extensions proposed and tested in Study 2, using an even larger sample that would allow for detecting smaller effects.”
    Quote in the paperSubstantial claim
29 of 29

  • The paper relies on this without support.

    There are no published direct replications of Pronin et al. (2002).

    “To the best of our knowledge, there are no published direct replications of Pronin et al. (2002).”

  • The authors state this as an assumption.

    People have limited to no control over their own biases, and biases occur regardless of choice and without awareness.

    “First, by definition, people have limited to no control over their own biases, and these seem to mostly occur to regardless of one's own choice, affecting individuals without their awareness.”
2 of 2

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?

  • Prior research has documented individuals' bias blind spots in non-U.S. populations.

    “prior research has documented individuals’ bias blind spots (Chandrashekar et al., 2021; Niszczota et al., 2023) and reliance on an affect heuristic (Finucane et al., 2000; Keller et al., 2006) in non-U.S. populations.”

    From Everyone I Don’t Like Is Biased: Affective Evaluations and the Bias Blind Spot (2026) · cites this paper as Chandrashekar et al., 2021

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.

  • Consistentt(44) = -4.54reported p < .001 · recomputed p < .001

    “t(44) = -4.54,p < .001”

  • Consistentt(302) = -16.16reported p < .001 · recomputed p < .001

    “t(302) = -16.16,p < .001”

  • Consistentt(620) = -32.04reported p < .001 · recomputed p < .001

    “t(620) = -32.04,p < .001”

  • Consistentt(302) = -5.22reported p < .001 · recomputed p < .001

    “t(302) = -5.22,p < .001”

  • Consistentt(620) = -10.54reported p < .001 · recomputed p < .001

    “t(620) = -10.54,p < .001”

  • Consistentt(44) = -1.13reported p = .265 · recomputed p = .265

    “t(44) = -1.13,p = .265”

  • Consistentt(44) = -1.97reported p = .055 · recomputed p = .055

    “t(44) = -1.97,p = .055”

  • Consistentt(302) = -6.39reported p < .001 · recomputed p < .001

    “t (302) = -6.39,p < .001”

  • Consistentt(620) = -13.01reported p < .001 · recomputed p < .001

    “t (620) = -13.01,p < .001”

  • Consistentt(42) = 5.09reported p < .001 · recomputed p < .001

    “t(42) = 5.09, p < .001”

  • Consistentt(620) = 31.74reported p < .001 · recomputed p < .001

    “t(620) = 31.74,p < .001”

  • Consistentt(42) = -4.55reported p < .001 · recomputed p < .001

    “t(42) = -4.55,p < .001”

  • Consistentt(620) = -30.38reported p < .001 · recomputed p < .001

    “t (620) = -30.38,p < .001”

13 of 13

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.

  • Study 3 denial-of-bias effect size is misquoted in the conclusion

    Minor

    The concluding comparison gives Study 3's denial-of-bias effect size as 1.27, but the preceding result gives it as 1.71. The value 1.27 belongs to Study 3's positive-personality ratings.

  • One-tailed tests report two-tailed p-values

    Minor

    The reported Study 1a p-values match two-tailed tests, despite the stated use of one-tailed tests. For t(44) = -1.97, the one-tailed p-value is about .028, not .055; for t(44) = -1.13, it is about .13, not .265.

2 of 2

Cites

Flagged first, then by citations

Research works printed but not linked: 6

  1. Nisbett, 1980, Human inference: Strategies and shortcomings of social judgment
  2. Baumeister, 2014, Advances in Experimental Social Psychology, 1
  3. Ross, 1996, Values and knowledge. The Jean Piaget symposium series, 103
  4. Jones, 1972, Attribution: Perceiving the causes of behavior
  5. Ross, 1995, Advances in Experimental Social Psychology, 255
  6. Jones, 1990, Interpersonal perception

Cited by

Sorted by citations

Related

Authors
Chandrashekar, S. P., Yeung, S. K., Yau, K. C., Cheung, C. Y., Agarwal, T. K., Wong, C. Y. J., Pillai, T., Thirlwell, T. N., Leung, W. N., Tse, C., Li, Y. T., Cheng, B. L., Chan, H. Y. C., & Feldman, G.
Affiliations
  • Hong Kong Metropolitan University (HK)
  • University of Hong Kong (HK)
Corresponding author
Feldman, Gilad
ORCID
Venue
Judgment and Decision Making
Publisher
Cambridge University Press (CUP)
Published
2021-11
Volume
16
Issue
6
Pages
1392-1412
ISSN
1930-2975 (ISSN-L)
DOI
10.1017/s1930297500008470
OpenAlex
W4286870253
Semantic Scholar
d417fda35681fec80169044e772bbf9ee7fd7af6
Cited by
11
References
47
Citations by year
2026: 2 · 2025: 3 · 2024: 2 · 2023: 2 · 2022: 1 · 2021: 1
FWCI
2.08
Type
journal article
Language
en
Open access
Licence
Locations
  • Judgment and Decision Making (publishedVersion, cc-by)
  • RePEc: Research Papers in Economics (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Topics
  • Social Sciences › Sociology and Political Science › Social and Intergroup Psychology
  • Social Sciences › Decision Sciences › General Decision Sciences › Decision-Making and Behavioral Economics
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Psychology of Moral and Emotional Judgment
Keywords
bias blind spot, self-other asymmetry, perceived bias, belief in free will, self-perception, cognitive biases, replication

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 2026-10-23.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-23.

Cited sources are checked by 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