CoChat
Sign Up
Cited by 228Child and Animal Learning DevelopmentAutism Spectrum Disorder ResearchPsychometric Methodologies and Testing

A psychometric analysis of the reading the mind in the eyes test: toward a brief form for research and applied settings

Open access
Olderbak, Sally, Wilhelm, Oliver, Olaru, Gabriel, Geiger, Mattis, Brenneman, Meghan W., Roberts, Richard D.·2015·Frontiers in Psychology
DOI 10.3389/fpsyg.2015.01503PMC PMC4593947
Open this paper in CoChat — free accountOpen PDF

Abstract

The Reading the Mind in the Eyes Test is a popular measure of individual differences in Theory of Mind that is often applied in the assessment of particular clinical populations (primarily, individuals on the autism spectrum). However, little is known about the test's psychometric properties, including factor structure, internal consistency, and convergent validity evidence. We present a psychometric analysis of the test followed by an evaluation of other empirically proposed and statistically identified structures. We identified, and cross-validated in a second sample, an adequate short-form solution that is homogeneous with adequate internal consistency, and is moderately related to Cognitive Empathy, Emotion Perception, and strongly related to Vocabulary. We recommend the use of this short-form solution in normal adults as a more precise measure over the original version. Future revisions of the test should seek to reduce the test's reliance on one's vocabulary and evaluate the short-form structure in clinical populations.

AI snapshot

Written from the paper’s full text.

TL;DR

The full 36-item Reading the Mind in the Eyes Test showed poor homogeneity and poor single-factor fit in two adult samples, while a data-driven 10-item ACO short form had adequate omega and CFA fit and cross-validated in a second sample. The authors recommend the 10-item short form for unimpaired adults, but note it remains strongly related to vocabulary.

Key findings

  • The full 36-item Eyes Test had weak inter-item agreement (average tetrachoric *r* = .08, range −.28 to .41) and poor single-factor CFA fit [χ²(594) = 707.03, *p* < .05; RMSEA = .020; WRMR = 1.042; CFI = .728; TLI = .711]; omega was .75 but likely inflated by test length, and 20 items (56%) had weak loadings (average loading = .278).
  • EFA favored a five-factor solution, but only 27 of 36 items loaded >.30 on any factor and factors were weakly related; previously proposed affect-based short forms/subscales (Konrath, Harkness) showed poor fit, improper solutions, or poor omega (e.g., Konrath reduced ω = .58).
  • The 10-item ACO solution had adequate fit and omega in Study 1 (ω = .70; CFI = .937; TLI = .918; RMSEA = .023) and Study 2 (ω = .73; CFI = 1.000; TLI = 1.002; RMSEA = .000), with measurement invariance supported across samples.
  • Latent correlations for the ACO model were with Cognitive Empathy .21* (Study 1) and .56* (Study 2), Emotion Perception .34* and .29*, and Vocabulary .49* and .62*; the full scale correlated .21*, .35*, and .54* with those constructs in Study 1.

Methods

  • Two Amazon Mechanical Turk samples: Study 1 final N = 484 (243 female), mean age 33.4 years (SD = 11.0), 76% White non-Hispanic, 96% native English; Study 2 final N = 210 (108 female), mean age 39.9 years (SD = 22.4), 82% White non-Hispanic, 97% native English.
  • Measures included the 36-item Eyes Test, IRI Perspective Taking, ESE Cognitive Empathy, DANVA 2 Faces, and an 18-item ETS vocabulary test; analyses used tetrachoric correlations, EFA and CFA with WLSMV, omega, latent correlations, and data-driven MML and ant colony optimization (ACO) short-form selection.
  • Scale solutions were evaluated on CFA fit and omega; ACO was specified to find the shortest solution maintaining ω = .70, CFI = .95, and RMSEA = .02, followed by cross-validation and multiple-group invariance testing.

Limitations

  • The authors state the ACO solution might be somewhat overfitted in Study 1, so they cross-validated it in Study 2; they also state future revisions should reduce the test’s reliance on vocabulary and evaluate the short-form structure in clinical populations.
  • The samples were non-clinical Mechanical Turk adults, primarily White, US-based, and native English speakers, so the findings may not generalize to clinical populations or other cultural/language groups (evident from the reported sample descriptions).
  • The full-scale convergent correlations could not be meaningfully interpreted because of poor CFA fit, and the affect-based subscales could not be modeled; the short form still showed a strong relation with vocabulary, raising questions about discriminant validity from vocabulary (evident from the reported results).

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

Integrity signals

Registry

Crossref · 10.3389/fpsyg.2015.01503

Cites

88 of 88 references verified0 citation errors · 0 not checkable

Statistics

20 tests recomputed · 0 with the reported and recomputed p on opposite sides of .05

Claims the paper makes

12 of 71 cited sources checked · 8 supported · 0 flagged · accuracy not yet measured

  • Premisecitation

    Based on published estimates, the Eyes Test typically has poor internal consistency.

    “Based on published estimates, the Eyes Test typically has poor internal consistency (Voracek and Dressler, 2006; Harkness et al., 2010; Ragsdale and Foley, 2011; Vellante et al., 2012; Khorashad et al., 2015; cf. Dehning et al., 2012; Girli, 2014; Prevost et al., 2014, for an exception).”
    Quote in the paperSubstantial claim1 of 3 sources checked
    • cites Voracek and Dressler, 2006 source
    • cites Harkness et al., 2010 source
    • cites Ragsdale and Foley, 2011 source
  • Premisecitation

    The test-retest reliability of the Eyes Test is acceptable.

    “However, the test-retest reliability of the measure is acceptable (Hallerbäck et al., 2009; Yildirim et al., 2011; Vellante et al., 2012; Prevost et al., 2014; Khorashad et al., 2015).”
    Quote in the paperNot a substantial claim1 of 3 sources checked
    • cites Hallerbäck et al., 2009 source
    • cites Yildirim et al., 2011 source
    • cites Vellante et al., 2012 source
  • Premisecitation

    The Eyes Test does not meet assumptions of normality.

    “The Eyes Test also does not meet assumptions of normality (Söderstrand and Almkvist, 2012; Vellante et al., 2012).”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites Söderstrand and Almkvist, 2012 source
    • cites Vellante et al., 2012 source
  • Premisecitation

    Amazon Mechanical Turk workers appear more representative of the general population than the typical first year psychology students comprising many reviewed studies.

    “This study will re-examine these assertions based on two samples of Amazon Mechanical Turk workers, who appear more representative of the general population than are the typical first year psychology students comprising many of the studies reviewed in the preceding (see e.g., Berinsky et al., 2012).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Berinsky et al., 2012 source
  • Interpretationown data

    The latent correlation between the ACO Model and Vocabulary is similar in magnitude to the manifest correlation with Vocabulary found by Peterson and Miller (2012).

    “At the latent level, however, we found a strong correlation (0.49 in Study 1 and 0.62 in Study 2) between the ACO Model with Vocabulary, which is similar in magnitude to the manifest correlation with Vocabulary found by Peterson and Miller (2012).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Peterson and Miller, 2012 source
  • Classified as probably not a substantial claim. 31
  • May combine two claims. 51
  • The quote only partly states this claim. 3

  • Classified as probably not a substantial claim.

    Premisecitation

    The results of a confirmatory factor analysis suggest that the division of items into an affect specific factor is a poor fit to the data.

    “However, the results of a confirmatory factor analysis (CFA) suggest that the division of items into an affect specific factor is a poor fit to the data (Vellante et al., 2012).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Vellante et al., 2012 source

      The Results (Factorial analysis of the Eyes test) report poor fit for a three-factor model grouping items by positive, negative, and neutral valence. That supports the general point that an affect-based division fit poorly, but the paper did not test the singular affect-specific factor described in the claim. (LLM, conf=0.82)

      Experimental — accuracy not yet measured

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

    Premisecitation

    The Eyes Test has been shown to be moderately to strongly related with measures of emotion perception ability.

    “In particular, the Eyes Test has been shown to be moderately to strongly related with measures of emotion perception ability (Henry et al., 2009; Petroni et al., 2011).”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites Henry et al., 2009 source
    • cites Petroni et al., 2011 source

      The Results (section 3.8.3.1) report a moderate correlation (r = 0.57) between Eyes Test scores and N170 responses to facial emotional valence. The N170 is a measure of cortical processing, not a direct measure of emotion-perception ability; the paper does not establish the claimed relationship with multiple ability measures. (LLM, conf=0.86)

      Experimental — accuracy not yet measured

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

    Premisecitation

    In general, verbal abilities are associated with performance on the Eyes Test.

    “In general, verbal abilities are associated with performance on the Eyes Test (Golan et al., 2006; Ahmed and Miller, 2011; Peterson and Miller, 2012; cf. Ferguson and Austin, 2010 who found no relation).”
    Quote in the paperNot a substantial claim1 of 3 sources checked
    • cites Golan et al., 2006 source
    • cites Ahmed and Miller, 2011 source
    • cites Peterson and Miller, 2012 source

  • May combine two claims.

    Premisecitation

    The relation between the Eyes Test and self-report measures of Cognitive Empathy is mixed; the test has been shown to be both weakly negatively related and weakly positively related to self-reported cognitive empathy.

    “Likewise, the relation between the Eyes Test and self-report measures of Cognitive Empathy is mixed; the test has been shown to be both weakly negatively related (Spreng et al., 2009) and weakly positively related to self-reported cognitive empathy (Grove et al., 2014).”
    Quote in the paperNot a substantial claim2 of 2 sources checked
    • cites Spreng et al., 2009 source
    • cites Grove et al., 2014 source

      In Results, Table 3 (page unavailable), the factor combining the Eyes and KDEF tasks correlates positively with self-reported cognitive empathy in the autism group (r = 0.32) and parent group (r = 0.48); the control-group correlation is not significant. This supports a positive association in some groups, but the paper reports correlations for a combined performance-task factor, not the Eyes Test alone, and the parent-group correlation is not clearly weak. The negative finding is attributed to Spreng et al., not Grove et al. (LLM, conf=0.87)

      Experimental — accuracy not yet measured

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

    Premisecitation

    Previous research suggests that at the manifest level, the Eyes Test is moderately related to verbal IQ.

    “Previous research suggests that at the manifest level, the Eyes Test is weakly related to verbal fluency (Ahmed and Miller, 2011) and moderately related to verbal IQ (Golan et al., 2006; Peterson and Miller, 2012).”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites Golan et al., 2006 source
    • cites Peterson and Miller, 2012 source

  • May combine two claims.

    Findingown data

    The average inter-item correlation and the range of inter-item correlations are outside the range recommended by Clark and Watson (1995) for sufficient internal consistency (0.15 to 0.50).

    “The average inter-item correlation, as well as the range of inter-item correlations, are outside the range recommended by Clark and Watson (1995) for sufficient internal consistency (recommended values range from 0.15 to 0.50).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Clark and Watson, 1995 source

  • The quote only partly states this claim.

    Findingown data

    The configural invariance model for the ACO Model across Study 1 and Study 2 fit the data well according to all fit indices except WRMR [χ2(70) = 79.29, p = 0.21, RMSEA = 0.020 (0.000-0.038), WRMR = 1.082, CFI = 0.962, TLI = 0.951].

    “Overall, this structure fit the data well according to all fit indices, which the exception of the WRMR [χ2(70) = 79.29, p = 0.21, RMSEA = 0.020 (0.000–0.038), WRMR = 1.082, CFI = 0.962, TLI = 0.951].”
    Quote in the paperSubstantial claim

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

    Findingown data

    Initially, 616 individuals participated, and 127 were removed because they failed to correctly respond to at least one attention check question.

    “Initially, 616 individuals participated; 127 were removed because they failed to correctly respond to at least one attention check questions (an admittedly strict criteria, designed to ensure compliant respondents).”
    Quote in the paperNot a substantial claim
    • The authors state this as an assumption: Failing to correctly respond to at least one attention check question indicates non-compliant respondents.

  • May combine two claims.

    Findingown data

    The omega estimate of the Konrath reduced 17-item structure was poor, indicating poor factor saturation and internal consistency.

    “Similarly, the omega estimate of this short-form structure was poor indicating poor factor saturation and internal consistency.”
    Quote in the paperSubstantial claim

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

    Findingown data

    The RMSEA for the single-factor CFA model of the full Eyes Test was 0.020, which indicated good fit.

    “The absolute fit index RMSEA (Steiger and Lind, 1980), which estimates the amount of misfit per degree of freedom by adjusting the chi-square measure to account for degrees of freedom, thus ceteris paribus simple models are preferred, indicated good fit.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Only 27 of the 36 Eyes Test items moderately loaded on at least one factor, with several items unrelated to any of the five factors.

    “However, an examination of the factor pattern, identifying items with moderate loadings or higher (>0.30), indicated that only 27 of the 36 items moderately loaded on at least one factor, with several items unrelated to any of the five factors (Table 2).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The ACO Model solution in Study 2 had an adequate omega (ω = 0.73), indicating acceptable internal consistency and factor saturation.

    “The ACO Model solution in Study 2 had an adequate omega (ω = 0.73), indicating acceptable internal consistency and factor saturation.”
    Quote in the paperSubstantial claim

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

    Findingown data

    The final sample consisted of 484 participants (243 female), 33.4 years of age (SD = 11.0), primarily White, non-Hispanic, from a variety of education levels.

    “Our final sample consisted of 484 participants (243 female), 33.4 years of age (SD = 11.0), primarily White, non-Hispanic (76%; 7% Black, non-Hispanic; 8% Asian; 9% Other), from a variety of education levels (13% High School or GED equivalent, 44% Some College, 43% Bachelors Degree or higher).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The omega estimates were poor for all three Harkness subscales, indicating poor factor saturation and internal consistency.

    “The omega estimates were poor for all three subscales indicating poor factor saturation and internal consistency.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingcitation

    The average inter-item correlation of the ACO 10-item solution was 0.18, which is acceptable according to Clark and Watson (1995).

    “The average inter-item correlation (0.18) is acceptable according to Clark and Watson (1995)”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Clark and Watson (1995) source

  • May combine two claims.

    Findingown data

    When the Harkness three affect-based subscales were modeled in a CFA as correlated latent factors, the test structure was a poor fit to the data according to the fit indices, with many items loading weakly on their respective latent factor.

    “When modeled in a CFA, with the three subscales modeled as latent factors indicated by their respective items and with the factors correlated with one another, the test structure was a poor fit to the data according to the fit indices, with many items loading weakly on their respective latent factor”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Based on latent correlations, the Konrath reduced short-form solution had weak to moderate relations with Cognitive Empathy, moderate relations with Emotion Perception, and was strongly related with Vocabulary.

    “Finally, based on the latent correlations, this short-form solution had weak to moderate relations with Cognitive Empathy, moderate relations with Emotion Perception, and was strongly related with Vocabulary.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The short-form solution is homogeneous with adequate internal consistency.

    “We identified, and cross-validated in a second sample, an adequate short-form solution that is homogeneous with adequate internal consistency, and is moderately related to Cognitive Empathy, Emotion Perception, and strongly related to Vocabulary.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The measurement model structure of the ACO Model in Study 2 fit the data sufficiently according to the CFA fit statistics, with eight of the 10 items loading moderately to strongly on the latent factor.

    “The measurement model structure fit the data sufficiently according to the CFA fit statistics, with eight of the 10 items loading moderately to strongly on the latent factor.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The relations of the ACO Model with Emotion Perception, Vocabulary, and Perspective Taking were similar between Study 1 and Study 2, but the relation with ESE Cognitive Empathy was considerably higher in Study 2.

    “The relations with Emotion Perception, Vocabulary, and Perspective Taking were similar between Study 1 and Study 2, however the relation with ESE Cognitive Empathy was considerably higher.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingcitation

    The average inter-item correlation of the ACO Model in Study 2 was 0.20, which is acceptable according to Clark and Watson (1995).

    “The average inter-item correlation (0.20) was acceptable according to Clark and Watson (1995)”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Clark and Watson (1995) source

  • May combine two claims.

    Findingown data

    The Eyes Test data were normally distributed (skew = −0.51; kurtosis = −0.06), with participants on average receiving high scores (M = 27.20, SD = 3.82, average percent correct = 76%), and individual scores ranging from 16 to 36.

    “The data were normally distributed (skew = −0.51; kurtosis = −0.06) with participants, on average, receiving high scores (M = 27.20, SD = 3.82, Average percent correct = 76%), with individual level scores ranging from 16 (44% correct) to 36 (100% correct).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    In an additional 100 runs of the ACO, the same 10 items were identified over 90% of the time, with deviating solutions of lesser psychometric quality.

    “(and in an additional 100 runs of the ACO, the same 10 items were identified over 90% of the time with deviating solutions of lesser psychometric quality; please see (Olaru et al., in press), for details on the procedure).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    At the latent level in Study 2, the ACO Model was weakly to strongly related with measures of Cognitive Empathy, weakly related with Emotion Perception, and strongly related with Vocabulary.

    “At the latent level, the model was weakly to strongly related with measures of Cognitive Empathy, weakly related with Emotion Perception, and strongly related with Vocabulary.”
    Quote in the paperSubstantial claim

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

    Findingown data

    An additional three people were removed because there was no variance in their response across an entire scale.

    “An additional three people were removed because there was no variance in their response across an entire scale (i.e., the person consistently selected the same response).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    An adequate short-form solution was identified and cross-validated in a second sample.

    “We identified, and cross-validated in a second sample, an adequate short-form solution that is homogeneous with adequate internal consistency, and is moderately related to Cognitive Empathy, Emotion Perception, and strongly related to Vocabulary.”
    Quote in the paperSubstantial claim

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

    Findingown data

    Two outliers in the Eyes Test were identified and removed, and after their removal the Eyes Test composite scores were more normally distributed.

    “Finally, we identified two outliers in the Eyes Test, where participants performed substantially worse than the rest of the sample. When those two persons were removed, the Eyes Test composite scores were more normally distributed, so these two persons were removed for the remainder of the analyses.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The tetrachoric correlations between Eyes Test items ranged from −0.28 to 0.41, with an average inter-item correlation of 0.08, indicating weak agreement between items and many negative relations.

    “The tetrachoric correlations between items ranged from −0.28–0.41, with average inter-item correlation at 0.08, suggesting there is weak agreement between items, with many items negatively related to one another (interestingly, relations between those items that shared the same target word were also weak [Cautious; r = 0.08; Fantasizing: r = 0.28; Preoccupied: r = 0.16]).”
    Quote in the paperSubstantial claim

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

    Findingown data

    With the exception of item 17, the most frequently chosen response option was the correct response option.

    “With the exception of item 17, the most frequently chosen response option was the correct response option (see Table A1 in Supplementary Materials), which is supported by prior studies of normal populations (e.g., Baron-Cohen et al., 2001; Harkness et al., 2005).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Baron-Cohen et al., 2001 source
    • cites Harkness et al., 2005 source

  • May combine two claims.

    Findingown data

    The final MML model showed acceptable fit according to the CFA fit indices, with all items moderately or strongly loading on the latent factor.

    “This final model showed acceptable fit according to the CFA fit indices, with all items moderately or strongly loading on the latent factor.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The internal consistency of all measures in Study 2 was acceptable: Perspective Taking (ω = 0.88); Emotion Specific Empathy Cognitive Empathy (ω = 0.94); DANVA 2 (ω = 0.95); Vocabulary (ω = 0.90).

    “The internal consistency of all measures was acceptable: Perspective Taking (ω = 0.88); Emotion Specific Empathy Cognitive Empathy (ω = 0.94); DANVA 2 (ω = 0.95); Vocabulary (ω = 0.90).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The ordinal omega estimate for the full Eyes Test was 0.75, indicating acceptable factor saturation and internal consistency.

    “According to omega the factor saturation and internal consistency of the scale is acceptable (ω = 0.75).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The inter-item correlations of the full Eyes Test are low, with many items unrelated to one another and in many instances negatively related.

    “The inter-item correlations are low, with many items unrelated to one another, and in many instances, negatively related”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S1 (asserted), S3 (load-bearing), S4 (based).

    Findingown data

    In the sample, factor saturation for the DANVA 2 Faces subscale was excellent (ω = 0.93).

    “Factor saturation was excellent (ω = 0.93).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    At the latent level, the ACO short-form solution was weakly related to both measures of Cognitive Empathy and moderately related to both Emotion Perception and Vocabulary.

    “At the latent level, this short-form solution was weakly related to both measures of Cognitive Empathy and moderately related to both Emotion Perception and Vocabulary.”
    Quote in the paperNot a substantial claim

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

    Findingown data

    In the sample, factor saturation for the Perspective Taking subscale was excellent (ω = 0.93).

    “In our sample, the factor saturation was excellent (ω = 0.93).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Twenty items (56% of the scale) had weak loadings on the single factor, with the average loading at 0.278 (range 0.073 to 0.560).

    “Twenty items (56% of the scale) had weak loadings with the average loading at 0.278 (loadings ranged from 0.073 to 0.560).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    For all items, residual variance (average = 0.910) exceeded variance explained by the single factor (average = 0.090), and the single factor explained no more than 31% of the variance for any item.

    “For all items there was more residual variance (average = 0.910) than variance explained by the single factor (average = 0.090), with the single factor not explaining more than 31% of the variance for any item.”
    Quote in the paperSubstantial claim

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

    Findingown data

    In the sample, factor saturation for the Vocabulary test was acceptable (ω = 0.79).

    “Factor saturation was acceptable (ω = 0.79).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The interfactor correlations for the Harkness subscales were moderate between Negative Affect and Positive Affect (r = 0.399) and between Positive Affect and Neutral Affect (r = 0.490), and strong between Negative Affect and Neutral Affect (r = 0.876).

    “The interfactor correlations were moderate (Negative Affect with Positive Affect: r = 0.399, p < 0.05; Positive Affect with Neutral Affect: r = 0.490, p < 0.05), with the correlation between the Negative Affect and Neutral Affect subscales strong (r = 0.876, p < 0.05)”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    At the latent level, the MML Solution was not significantly related to Cognitive Empathy, weakly related to Emotion Perception, and moderately related to Vocabulary.

    “At the latent level, the MML Solution was not significantly related to Cognitive Empathy, weakly related to Emotion Perception, and moderately related to Vocabulary.”
    Quote in the paperSubstantial claim

  • The quote only partly states this claim.

    Findingown data

    Using criteria of omega estimate, factor saturation, and measurement model fit, the authors identified a short-form solution (ACO Model) that meets the criteria.

    “Using these criteria, we identified a short-form solution (ACO Model) that meets both criteria.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingcitation

    The response options of the Eyes Test occur less frequently in the English language than other descriptive terms; mean frequency 8556 vs 28435.

    “The response options of the Eyes Test, in general, occur less frequently in the English language [e.g., the mean frequency of the response options, based on the Corpus of Contemporary American English (Davies, 2008), is 8556] compared with other descriptive terms [e.g., the six based emotions (anger, disgust, fear, happy, sad, surprise) where the mean frequency is 28,435].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Davies, 2008

      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).

    Findingnone given

    The ACO Model has an almost equal distribution of male and female identities.

    “Regarding the design, the ACO Model has an almost equal distribution of male and female identities, like the original version.”
    Quote in the paperNot a substantial claim

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

    Interpretationreasoning

    It seems plausible that the low internal consistency occurs because the test does not have a single factor solution, but instead measures several factors.

    “It seems plausible that the low internal consistency occurs because the test does not have a single factor solution, but instead measures several factors.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    ToM is often considered conceptually similar to, or equivalent with, cognitive empathy because both constructs involve inferring the mental state of another person.

    “ToM is often considered to be conceptually similar to, or equivalent with, cognitive empathy because both constructs involve inferring the mental state of another person (e.g., Lawrence et al., 2004).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Lawrence et al., 2004 source

  • May combine two claims.

    Premisenone given

    Available research with the Eyes Test suggests that in general, the test has poor internal consistency and there is limited evidence of the test's homogeneity.

    “In summary, available research with the Eyes Test suggests that in general, the test has poor internal consistency and there is limited evidence of the test's homogeneity.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    The Eyes Test is proposed to measure a single construct.

    “While the Eyes Test is proposed to measure a single construct (Baron-Cohen et al., 2001), as noted previously others have suggested that there might be subscales composed of those items that represent positive, negative, and neutral affect (Harkness et al., 2005; Maurage et al., 2011; Konrath et al., 2014).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Baron-Cohen et al., 2001 source

  • Classified as probably not a substantial claim.

    Premisecitation

    The Reading the Mind in the Eyes Test is a popular measure of individual differences in Theory of Mind capabilities.

    “The Reading the Mind in the Eyes Test is a popular measure of individual differences in Theory of Mind capabilities (ToM; Baron-Cohen et al., 2001).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Baron-Cohen et al., 2001 source

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

    Premisecitation

    The Eyes Test distinguishes many clinical populations, most often individuals on the autistic spectrum, from unimpaired control participants in their ToM capabilities.

    “The test distinguishes many clinical populations (most often individuals on the autistic spectrum) from unimpaired control participants in their ToM capabilities.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Baron-Cohen et al., 2001 source

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

    Premisecitation

    The Eyes Test was designed to measure the first stage of ToM attribution, identifying the relevant mental state of the stimulus, as opposed to inferring the content of that mental state.

    “The Eyes Test was designed to measure the first stage of ToM attribution, which is identifying the relevant mental state of the stimulus (as opposed to the second stage; inferring the content of that mental state; see Baron-Cohen et al., 2001).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Baron-Cohen et al., 2001 source

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

    Premisecitation

    ToM is associated with language ability.

    “ToM is also associated with language ability.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    The RMSEA can result in artificially lower estimates when there are weak factor loadings, as was the case in this model.

    “However, the RMSEA will result in artificially lower estimates when there are weak factor loadings (Heene et al., 2011; Savalei, 2012), as is the case in this model, which can lead to an acceptance of a larger saturated model (Rigdon, 1996).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Heene et al., 2011 source
    • cites Savalei, 2012 source
    • cites Rigdon, 1996 source

  • May combine two claims.

    Interpretationown data

    The ACO Model is a sufficient short-form solution to the Eyes Test because it has an adequate omega estimate and adequate fit in a measurement model.

    “Because the ACO Model has an adequate omega estimate and adequate fit in a measurement model, we suggest the ACO Model is a sufficient short-form solution to the Eyes Test.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Konrath et al. (2014) proposed a reduced 17-item version of the Eyes Test.

    “The first model tested is a reduced 17-item version proposed by Konrath et al. (2014).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Konrath et al. (2014) source

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

    Interpretationcitation

    Given that all fit indices except WRMR indicated acceptable fit, the WRMR can be ignored in this instance.

    “However, given that all fit indices, with the exception of the WRMR, indicate acceptable fit, we can ignore the WRMR in this instance (Muthén, 2014).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Muthén, 2014

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

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

    Premisecitation

    Cronbach's alpha is highly biased by long tests, and Cortina recommended it not be used for tests with more than 40 items.

    “Also, Cronbach's alpha is highly biased by long tests; Cortina (1993) recommended Cronbach's alpha not be used for tests with more than 40 items and the Eyes Test, at 36 items, is close to this limit.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Cortina, 1993 source

  • May combine two claims.

    Interpretationown data

    The MML short-form solution has a poor omega estimate indicating poor internal consistency and factor saturation, so it was not sufficient.

    “However, despite an adequate measurement model structure, this short-form solution has a poor omega estimate indicating poor internal consistency and factor saturation, so this short-form solution was also not sufficient.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    In contrast to the complete Eyes Test, the ACO Model has adequate psychometric properties according to our criteria, suggesting this is a sufficient short-form solution to the complete Eyes Test.

    “In contrast to the complete Eyes Test, the ACO Model has adequate psychometric properties according to our criteria, suggesting this is a sufficient short-form solution to the complete Eyes Test.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    Schizophrenic patients and alcoholics perform worse on the Eyes Test compared to controls.

    “In addition, researchers have found schizophrenic patients (Bora et al., 2009) and alcoholics (Maurage et al., 2011) perform worse on the Eyes Test compared to controls.”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Bora et al., 2009 source
    • cites Maurage et al., 2011 source

  • May combine two claims.

    Interpretationown data

    In the second sample, the ACO Model showed sufficient internal consistency and factor saturation.

    “In a second sample, the ACO Model maintained an adequate omega estimate, indicating sufficient internal consistency and factor saturation, and demonstrable homogeneity.”
    • Also stated in: § Summary · hedged
      “This short-form solution is homogeneous with adequate internal consistency suggesting it is a more precise measure of ToM compared with the original form.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    Mechanical Turk can be used to obtain data from a diverse sample and the data is comparable to samples collected with traditional methods.

    “We chose Mechanical Turk because research suggests this platform can be used to obtain data from a diverse sample and the data is comparable to samples collected with traditional methods (Buhrmester et al., 2011; Casler et al., 2013).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Buhrmester et al., 2011 source
    • cites Casler et al., 2013 source

  • May combine two claims.

    Interpretationreasoning

    Given that the Eyes Test is frequently applied in clinical settings, its poor internal consistency is troubling because it suggests it is unclear what exactly is being measured by the current version of the test.

    “Given that this test is frequently applied in clinical settings, the poor internal consistency in particular is troubling, because it suggests it is unclear what exactly is being measured by the current version of the test.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    The Konrath reduced 17-item structure is not a sufficient solution because it has poor internal consistency according to omega and poor structural validity evidence.

    “Because this short-form solution has poor internal consistency according to omega and had poor structural validity evidence, this structure is not a sufficient solution.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    Because the single-factor solution had poor fit in CFA, the fit of the convergent validity models was poor.

    “However, given the poor fit of the single factor solution to the data in a CFA, the fit of these models is poor, thus the correlations could not be meaningfully interpreted.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Premisenone given

    Several studies identified the Eyes Test as having poor internal consistency and some have proposed that the test is not unidimensional but instead measures multiple factors.

    “Several studies identified the test as having poor internal consistency and some have proposed that the test is not unidimensional but instead measures multiple factors.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    Konrath et al. (2014) also proposed two subscales, Positive Affect and Negative Affect, with two of the 17 items from the Konrath Reduced Scale Model not used on either subscale.

    “Konrath et al. (2014) also proposed two subscales - (1) Positive Affect; and (2) Negative Affect—with two of the 17 items from the Konrath Reduced Scale Model not used on either subscale.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Konrath et al. (2014) source

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationreasoning

    If the steps taken to shorten the Eyes Test focus on increasing the test's homogeneity, the short-form version would also be a more precise measure of ToM.

    “A reduced scale would be beneficial for the assessment field because it would be quicker to administer and if the steps taken to shorten the measure focus on increasing the test's homogeneity, the short-form version would also be a more precise measure of ToM.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    In a study comparing ACO with other methods to identify a short-form solution, the ACO method was found to be an efficient procedure that identified a short-form solution with the best fitting measurement model structure, the highest internal consistency, and the highest factor saturation.

    “In a study comparing ACO with other methods to identify a short-form solution, the ACO method was found to be an efficient procedure that, when compared with the other methods, identified a short-form solution with the best fitting measurement model structure (modeled in a CFA), with the highest internal consistency and factor saturation (assessed with ω; Olaru et al., in press).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Olaru et al., in press

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

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

    Interpretationreasoning

    ToM, as measured with the Eyes Test, is part of a larger set of research that includes Emotion Perception, referred to as social intelligence, emotional intelligence, interpersonal abilities, and/or socio-emotional abilities.

    “Given these overlaps, it should be considered that ToM, as measured with the Eyes Test, is part of a larger set of research that includes Emotion Perception, referred to as social intelligence, emotional intelligence, interpersonal abilities, and/or socio-emotional abilities.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Both ToM and Emotion Perception tests are ability tests that test the extent to which participants can read the minds, facial expressions, and feelings of another person.

  • Classified as probably not a substantial claim.

    Premisecitation

    Baron-Cohen and colleagues did not discuss purposely including items with different affects.

    “First, Baron-Cohen and colleagues did not discuss purposely including items with different affects and they never proposed that the affect of the target word could be used to create subscales or should bias performance on the test, so in our analyses we operated under the assumption that there was no preference in terms of which items were the most appropriate for the measure of ToM.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Baron-Cohen et al., 2001 source

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    Differences in groups identified by the Eyes Test might be because the test identifies differences between these groups in their Vocabulary and Emotion Perception abilities.

    “Thus, the differences in groups identified by the Eyes Test might be because the test identifies differences between these groups in their Vocabulary and Emotion Perception abilities.”
    Quote in the paperSubstantial claim

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

    Premisecitation

    Previous research suggests that at the manifest level, the Eyes Test is weakly related to verbal fluency.

    “Previous research suggests that at the manifest level, the Eyes Test is weakly related to verbal fluency (Ahmed and Miller, 2011) and moderately related to verbal IQ (Golan et al., 2006; Peterson and Miller, 2012).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Ahmed and Miller, 2011 source

  • May combine two claims.

    Premisecitation

    Baron-Cohen and colleagues never proposed that the affect of the target word could be used to create subscales or should bias performance on the test.

    “First, Baron-Cohen and colleagues did not discuss purposely including items with different affects and they never proposed that the affect of the target word could be used to create subscales or should bias performance on the test, so in our analyses we operated under the assumption that there was no preference in terms of which items were the most appropriate for the measure of ToM.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Baron-Cohen et al., 2001 source

  • May combine two claims.

    Premisecitation

    Research on the extent to which individuals on the Autism Spectrum have a deficit in emotion perception is mixed, but findings generally point to a deficit.

    “Research regarding the extent to which individuals on the Autism Spectrum have a deficit in their ability to perceive emotions in the face (cf. Harms et al., 2010) and a deficit in language ability (cf. Bosseler and Massaro, 2003) is mixed, however findings generally point to a deficit in both.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Harms et al., 2010 source

  • May combine two claims.

    Premisecitation

    Research on the extent to which individuals on the Autism Spectrum have a deficit in language ability is mixed, but findings generally point to a deficit.

    “Research regarding the extent to which individuals on the Autism Spectrum have a deficit in their ability to perceive emotions in the face (cf. Harms et al., 2010) and a deficit in language ability (cf. Bosseler and Massaro, 2003) is mixed, however findings generally point to a deficit in both.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Bosseler and Massaro, 2003 source

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

    Premisecitation

    ToM is largely considered by some to be similar to, if not the same as, Cognitive Empathy.

    “ToM is largely considered by some to be similar to, if not the same as, Cognitive Empathy (e.g., Lawrence et al., 2004).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Lawrence et al., 2004 source
77 of 77

  • The authors state this as an assumption.

    Failing to correctly respond to at least one attention check question indicates non-compliant respondents.

    “127 were removed because they failed to correctly respond to at least one attention check questions (an admittedly strict criteria, designed to ensure compliant respondents).”

  • The paper relies on this without support.

    None of the five factors identified by the exploratory factor analysis had been previously postulated.

    “Given that the Eyes Test is purported to measure a single factor, and that none of these five factors was previously postulated, we instead applied CFA to examine the factor structure of the test.”

  • The authors state this as an assumption.

    All items are equal indicators of theory of mind (ToM).

    “In both attempts, we are assuming that all items are equal indicators of ToM.”

  • The authors state this as an assumption.

    In the analyses, there was no preference in terms of which items were the most appropriate for the measure of ToM.

    “so in our analyses we operated under the assumption that there was no preference in terms of which items were the most appropriate for the measure of ToM.”

  • The paper relies on this without support.

    The response option definitions in the Eyes Test instructions were designed to mitigate the test's reliance on Vocabulary.

    “the inclusion of the response option definitions in the instructions for the test was designed to mitigate this effect.”

  • The paper relies on this without support.

    Several empirically derived short-form and subscale solutions to the Eyes Test were proposed in previous work.

    “In previous work, several empirically derived short-form and subscale solutions were proposed.”

  • The paper relies on this without support.

    Both ToM and Emotion Perception tests are ability tests that test the extent to which participants can read the minds, facial expressions, and feelings of another person.

    “Both ToM and Emotion Perception tests are described as ability tests and both test the extent to which participants can read the minds, facial expressions, and feelings of another person.”

  • The paper relies on this without support.

    The test's reliance on Vocabulary was mentioned by the test designers.

    “The test's reliance on Vocabulary was mentioned by the test designers”
8 of 8

Statistics that don’t add up

Each reported p-value recomputed from its test statistic and degrees of freedom.

  • Inconsistent p-value: t(481) = 1.29, p = .202

    Minor

    The paper reports p = .202 here; recomputing from the test statistic and its degrees of freedom gives p = .198 (two-tailed). The check cannot tell whether the statistic or the p-value is the misprint.

    Reported p = .202 · recomputed p = .198

  • Inconsistent p-value: χ2(226) = 226.78, p = .15

    Minor

    The paper reports p = .15 here; recomputing from the test statistic and its degrees of freedom gives p = .473. The check cannot tell whether the statistic or the p-value is the misprint.

    Reported p = .15 · recomputed p = .473

2 of 2

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.

  • Doesn’t add upt(481) = 1.29reported p = .202 · recomputed p = .198

    “t(481) = 1.29, p = .202”

  • Doesn’t add upχ²(226) = 226.78reported p = .15 · recomputed p = .473

    “χ2(226) = 226.78, p = .15”

  • ConsistentF(4, 479) = 1.22reported p = .301 · recomputed p = .301

    “F(4, 479) = 1.22, p = .301”

  • Consistentχ²(559) = 623.02reported p < .05 · recomputed p = .031

    “χ2(559) = 623.02, p < .05”

  • Consistentχ²(525) = 564.01reported p = .12 · recomputed p = .116

    “χ2(525) = 564.01, p = .12”

  • Consistentχ²(492) = 525.3reported p = .14 · recomputed p = .145

    “χ2(492) = 525.30, p = .14”

  • Consistentχ²(460) = 488.45reported p = .17 · recomputed p = .173

    “χ2(460) = 488.45, p = .17”

  • Consistentχ²(594) = 707.03reported p < .05 · recomputed p < .001

    “χ2(594) = 707.03, p < .05”

  • Consistentχ²(119) = 126.15reported p = .31 · recomputed p = .309

    “χ2(119) = 126.15, p = .31”

  • Consistentχ²(591) = 694.77reported p < .05 · recomputed p = .002

    “χ2(591) = 694.77, p < .05”

  • Consistentχ²(14) = 21.18reported p = .1 · recomputed p = .097

    “χ2(14) = 21.18, p = .10”

  • Consistentχ²(35) = 44.25reported p = .14 · recomputed p = .136

    “χ2(35) = 44.25, p = .14”

  • Consistentχ²(35) = 34.85reported p = .48 · recomputed p = .475

    “χ2(35) = 34.85, p = .48”

  • Consistentχ²(70) = 79.29reported p = .21 · recomputed p = .209

    “χ2(70) = 79.29, p = .21”

  • Consistentχ²(79) = 85.31reported p = .29 · recomputed p = .294

    “χ2(79) = 85.31, p = .29”

  • Consistentχ²(9) = 6.02reported p = .74 · recomputed p = .738

    “χ2(9) = 6.02, p = .74”

  • Consistentχ²(226) = 241.41reported p = .23 · recomputed p = .230

    “χ2(226) = 241.41, p = .23”

  • Consistentχ²(20) = 14.63reported p = .8 · recomputed p = .797

    “χ2(20) = 14.63, p = .80”

  • Consistentχ²(227) = 264.04reported p < .05 · recomputed p = .046

    “χ2(227) = 264.04, p < .05”

  • Consistentχ²(1) = 22.63reported p < .05 · recomputed p < .001

    “χ2(1) = 22.63, p < .05”

20 of 20

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.

  • Recommended short form fails the paper's CFA fit cutoffs

    Major

    The paper defines adequate fit as CFI and TLI of at least 0.95. The Study 1 ACO model has CFI = 0.937 and TLI = 0.918, yet its supposedly adequate fit is a basis for recommending it.

  • Lower end of comparison scores is not 74% correct

    Minor

    On the stated 36-item test, a score of 26.2 is about 72.8% correct, which rounds to 73%, not the reported lower bound of 74%.

    Reported 74 · recomputed 72.777778

  • Study 2 emotion-perception relation called both weak and moderate

    Minor

    The Study 2 results call this relation weak, consistent with the reported correlation of 0.29 and the paper's weak-correlation cutoff of 0.30. The Study 2 conclusion instead calls it moderate.

  • Invariance comparison reports incompatible degrees of freedom

    Minor

    Both compared models are reported with 226 degrees of freedom, but their comparison is reported with a 20-degree-of-freedom difference. The reported model degrees of freedom imply a difference of zero.

  • Study 1 kurtosis has opposite signs

    Minor

    The Study 1 descriptive text reports kurtosis of −0.06, whereas the full-scale row of Table 5 reports 0.06.

5 of 5

Cites

Flagged first, then by citations

Research works printed but not linked: 5

  1. Ekstrom, 1976, Manual for Kit of Factor-Referenced Cognitive Tests
  2. Kirkland, 2012a, Meta-analysis reveals a moderate relationship between tests of theory of mind and the Eyes Test
  3. Muthén, 2014, Re: Model Fit Index WRMR
  4. Muthén, 2011, Mplus User's Guide, 6th Edn
  5. Nunnally, 1994, Psychometric Theory.

Cited by

Sorted by citations

Related

Authors
Olderbak, S., Wilhelm, O., Olaru, G., Geiger, M., Brenneman, M. W., & Roberts, R. D.
Affiliations
  • Universität Ulm (DE)
  • Educational Testing Service (US)
  • Professional Examination Service (US)
Corresponding author
Olderbak, Sally
ORCID
Venue
Frontiers in Psychology
Publisher
Frontiers Media SA
Published
2015-10-06
Volume
6
Pages
1503-1503
ISSN
1664-1078 (ISSN-L)
DOI
10.3389/fpsyg.2015.01503
PMC
PMC4593947
OpenAlex
W2112864572
Cited by
228
References
88
Citations by year
2026: 15 · 2025: 21 · 2024: 16 · 2023: 27 · 2022: 32 · 2021: 36 · 2020: 29 · 2019: 20 · 2018: 15 · 2017: 9 · 2016: 7
FWCI
26.07
Type
journal article
Language
en
Open access
Licence
cc-by
Locations
  • Frontiers in Psychology (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • OPen Access Repositorium der Universität Ulm (OPARU) (Ulm University) (submittedVersion, cc-by)
  • Universität Ulm, Kommunikations- und Informationszentrum (kiz) (cc-by)
Funders
U.S. Army Research Institute (W5J9CQ-12-R-0002)
Topics
  • Social Sciences › Psychology › Developmental and Educational Psychology › Child and Animal Learning Development
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Autism Spectrum Disorder Research
  • Social Sciences › Decision Sciences › Management Science and Operations Research › Psychometric Methodologies and Testing
Keywords
Reading the Mind in the Eyes Test, theory of mind, psychometric properties, factor structure, internal consistency, convergent validity, cognitive empathy, emotion perception, vocabulary, autism spectrum disorder, individual differences

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 2026-10-28.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-28.

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