CoChat
Sign Up
Cited by 12Face Recognition and PerceptionFace recognition and analysisEvolutionary Psychology and Human Behavior

The DOI matches its Crossref record; 33 of the 33 references it cites were found in a registry or an index; 4 of the 41 sources its claims cite were checked for support.

Enhanced Matching of Children’s Faces in “Super-Recognisers” But Not High-Contact Controls

Open access
Bate, Sarah, Bennetts, Rachel, Murray, Ebony, Portch, Emma·2020·i-Perception
DOI 10.1177/2041669520944420PMC PMC7385838
Open this paper in CoChat — free account

Abstract

Face matching is notoriously error-prone, and some work suggests additional difficulty when matching the faces of children. It is possible that individuals with natural proficiencies in adult face matching (“super-recognisers” [SRs]) will also excel at the matching of children’s faces, although other work implicates facilitations in typical perceivers who have high levels of contact with young children (e.g., nursery teachers). This study compared the performance of both of these groups on adult and child face matching to a group of low-contact controls. High- and low-contact control groups performed at a remarkably similar level in both tasks, whereas facilitations for adult and child face matching were observed in some (but not all) SRs. As a group, the SRs performed better in the adult compared with the child task, demonstrating an extended own-age bias compared with controls. These findings suggest that additional exposure to children’s faces does not assist the performance in a face matching task, and the mechanisms underpinning superior recognition of adult faces can also facilitate the child face recognition. Real-world security organisations should therefore seek individuals with general facilitations in face matching for both adult and child face matching tasks.

Overview

AI snapshot

Written from the paper’s full text.

TL;DR

This study compared super-recognisers (SRs) and high- versus low-contact controls on simultaneous matching of adult and children’s faces. SRs generally performed well on both tasks but showed a stronger own-age bias than controls, while high contact with children did not improve matching.

Key findings

  • High- and low-contact controls performed similarly on adult and child face matching: mean sensitivity (A) was .74/.75 for high-contact controls and .76/.76 for low-contact controls; the Group × Version interaction was nonsignificant, F(1,38) = 0.110, p = .742, with Bayesian evidence favouring the null over an interaction model (BF01 = 38.65).
  • Among 10 SRs, 5 significantly outperformed controls on the adult matching test, but only 2 (SR07 and SR10) significantly outperformed controls on the child matching test; only SR10 also significantly outperformed controls on the adult test.
  • As a group, SRs scored higher on adult faces (M A = 0.94, SD = 0.03) than child faces (M A = 0.90, SD = 0.03), t(9) = 4.511, p = .001, d = 1.33; this adult–child difference was larger in SRs (M = 0.04, SD = 0.03) than controls (M = 0.00, SD = 0.11), t(47.55) = 2.259, p = .029, d = 0.50.
  • For the collapsed control group (N = 40), adult and child matching performance did not correlate, r = .195, p = .228.

Methods

  • Three groups of adult Caucasian females aged 18–50 participated: 20 high-contact controls (nursery/primary teachers; Mage = 33.9, SD = 8.1), 20 low-contact controls (Mage = 30.1, SD = 9.0), and 10 SRs (Mage = 39.0, SD = 7.3) reporting low contact with young children; SRs scored at least 1.96 SDs above control norms on the CFMT+ and Models Memory Test.
  • Participants completed two online simultaneous face matching tests, one with Caucasian adult faces (aged 20–35) and one with Caucasian children’s faces (aged 1–5), each with 48 trials; the tests had been calibrated to equal difficulty in prior control samples. Controls completed both tests in one randomized counterbalanced session; SRs’ adult scores were carried over from screening and they completed the child version about 1 year later.
  • Analyses used nonparametric signal detection measures A and b; mixed-factorial and Bayesian ANOVAs compared high- and low-contact controls; Crawford and Garthwaite modified t tests assessed individual SRs against controls, with paired and independent t tests comparing SR adult–child differences to controls.

Limitations

  • The paper states that the SR sample was small and power to detect effects greater than 2 SDs from the mean was estimated between 0.50 and 0.60; it also notes possible ceiling effects from control norms that may have caused some SRs to miss the cutoff for superior performance.
  • The paper states that one SR did not perform well on either matching task, and that inclusion criteria were tests of unfamiliar face memory rather than face matching, reinforcing concerns that screening should imitate the real-world task.
  • The study recruited only female participants, which the authors explain was to hold gender constant and because more women work in nursery/primary settings; this limits generalizability to males. The text also makes evident that the simultaneous matching task allowed unlimited time, so findings may not directly generalize to split-second real-world matching.

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

Integrity signals

Registry

Crossref · 10.1177/2041669520944420

Cites

33 of 33 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

4 of 41 cited sources checked · 1 supported · 0 flagged · accuracy not yet measured

  • Findingown data

    The departures from normality were primarily driven by skewness, which varied between −0.31 and −0.57 (minimal-to-moderate negative skew).

    “This was primarily driven by skewness, which varied between −0.31 and −0.57 (minimal-to-moderate negative skew).”
    Quote in the paperSubstantial claim
  • Findingown data

    In the Bayesian ANOVA for b, the model including main effects alone was preferred over the model including interactions by a factor of 3.11.

    “the model including main effects alone was again preferred over the model including interactions, by a factor of 3.11.”
    Quote in the paperSubstantial claim
  • Findingown data

    Modified t tests confirmed that five SRs outperformed controls on the adult test.

    “Modified t tests confirmed that five SRs outperformed controls on the adult test.”
    • Also stated in: § General Discussion · plain
      “only five significantly outperformed controls on the adult task”
    Quote in the paperSubstantial claim
  • Findingown data

    In the Bayesian ANOVA on A, there was moderate evidence that the null model was favoured over the model including version (BF01 = 4.28).

    “The Bayesian ANOVA on A revealed moderate evidence that the null model was favoured more than the model including version (BF01 = 4.28)”
    Quote in the paperNot a substantial claim
  • Findingown data

    The Bayesian ANOVA offered strong support for the null hypothesis over all models containing a group-based interaction (BF01 = 8.43-131.03).

    “and strong support for the null hypothesis over all models containing a group-based interaction (BF01 = 8.43–131.03).”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim. 20
  • It may state two separate claims as one. 14

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    The matching of facial identity across two simultaneously presented images is a notoriously difficult task, even under the simplest of circumstances.

    “The matching of facial identity across two simultaneously presented images is a notoriously difficult task, even under the simplest of circumstances (Burton et al., 2010; Megreya & Burton, 2006).”
    Quote in the paperNot a substantial claim2 of 2 sources checked
    • cites Burton et al., 2010 source

      The General Discussion says matching two face images is not trivially easy even under apparently optimal conditions. The introductory discussion reports 10%–25% error rates for pairs of unfamiliar faces, but says performance on familiar faces reaches ceiling levels. The claim omits that familiarity qualification. (LLM, conf=0.84)

      Experimental — accuracy not yet measured
    • cites Megreya & Burton, 2006 source

      The Abstract says that matching two images of an unfamiliar face is difficult even under good conditions, and that poor performance persists when people judge whether two images show the same person. But it contrasts this with very good performance for familiar faces; the claim omits the unfamiliar-face qualification. (LLM, conf=0.85)

      Experimental — accuracy not yet measured

  • It may state two separate claims as one.

    Premisecitation

    Only two studies to date have investigated the performance of typical adults when matching children's faces, and difficulties were apparent in both.

    “although only two studies to date have investigated the performance of typical adults when matching children’s faces, difficulties were apparent in both.”
    Quote in the paperNot a substantial claim2 of 2 sources checked
    • cites White et al., 2015 source

      Experiment 1 Results report lower accuracy for child than adult faces, and its Discussion describes error rates for children as especially poor. The paper does not establish that only two studies had investigated typical adults matching children’s faces, or that difficulties were found in both studies. (LLM, conf=0.84)

      Experimental — accuracy not yet measured
    • cites Kramer et al., 2018 source

  • It may state two separate claims as one.

    Findingown data

    Visual inspection of the hit and correct rejection data revealed departures from normality on several measures, and the Shapiro-Wilk statistic confirmed significant departures from normality for correct rejections on the adult test, and for hits on the child test, W(50) < .95, ps < .032.

    “Visual inspection of the hit and correct rejection data revealed departures from normality on several measures, and the Shapiro–Wilk statistic confirmed significant departures from normality for correct rejections on the adult test, and for hits on the child test, W(50) < .95, ps < .032.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data

    Neither the version nor group main effects were significant for A, F(1,38) = 0.017, p = .896, ηρ2 = .01, and F(1,38) = 0.463, p = .501, ηρ2 = .01, respectively.

    “Neither the version nor group main effects were significant, F(1,38) = 0.017, p = .896, ηρ2 = .01, and F(1,38) = 0.463, p = .501, ηρ2 = .01, respectively.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Findingown data

    One SR performed less well on both tests, achieving z scores of 1.56 and 0.75, respectively.

    “One individual performed less well on both tests, achieving z scores of 1.56 and 0.75, respectively (see Figure 1).”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    For bias (b), neither the interaction nor the version or group main effects were significant, F(1,38) = 0.001, p = .983, ηρ2 = .01, F(1,38) = 0.355, p = .555, ηρ2 = .01, and F(1,38) = 0.198, p = .659, ηρ2 = .01, respectively.

    “The same ANOVA was also run on b (bias): neither the interaction nor the version or group main effects were significant, F(1,38) = 0.001, p = .983, ηρ2 = .01, F(1,38) = 0.355, p = .555, ηρ2 = .01, and F(1,38) = 0.198, p = .659, ηρ2 = .01, respectively (see Table 2).”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    All but one of the SRs obtained consistently high scores when matching adult and child faces, performing within the top 7% of the population on both tasks, irrespective of whether single-case comparisons reached significance.

    “Indeed, all but one of the SRs obtained consistently high scores when matching adult and child faces, performing within the top 7% of the population on both tasks, irrespective of whether single-case comparisons reached significance.”
    • Also stated in: § General Discussion · plain
      “In sum, this study presents evidence of high matching performance for children’s faces in SRs who also excel at adult face matching.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Findingown data

    In the Bayesian ANOVA on A, there was little evidence that the null model was favoured over the model including group (BF01 = 2.87).

    “but little evidence that it was favoured more than the model including group (BF01 = 2.87; this may indicate that the data were not sensitive enough to reliably detect group differences in sensitivity).”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data

    Overall A control means were 0.75 (SD = 0.09) for the adult task and 0.75 (SD = 0.08) for the children's task.

    “This produced overall A control means of 0.75 (SD = 0.09) and 0.75 (SD = 0.08) for the adult and children’s tasks, respectively.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingcitation

    The facilitated performance of SR9 was restricted to face memory, and this individual achieved the lowest scores in the two matching tasks.

    “Consistent with our previous work (Bate et al., 2018; Bate, Frowd, et al., 2019), the facilitated performance of SR9 was restricted to face memory, and this individual achieved the lowest scores in the two matching tasks.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Bate et al., 2018

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

    • cites Bate, Frowd, et al., 2019 source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    Face matching performance declines when task demands are increased, such as when viewpoint varies between the two images or there are long time periods between image captures.

    “and when task demands are increased (e.g., viewpoint varies in the two images: Estudillo & Bindemann, 2014; or there are long time periods between the two image captures: White et al., 2014).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Estudillo & Bindemann, 2014 source
    • cites White et al., 2014 source

  • Probably not a substantial claim.

    Premisecitation

    Most individuals show poorer performance when asked to identify faces of a different ethnicity than their own.

    “It is well established that most individuals show poorer performance when asked to identify faces of a different ethnicity than their own”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Meissner & Brigham, 2001 source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    Face matching performance declines for unfamiliar compared with familiar face pairs.

    “The performance declines for unfamiliar compared with familiar face pairs (Bruce et al., 2001)”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Bruce et al., 2001 source

  • Probably not a substantial claim.

    Premisecitation

    In typical perceivers, adult face memory performance may be facilitated in individuals who frequently have contact with children, such as nursery teachers.

    “but may be facilitated in individuals who frequently have contact with children (e.g., nursery teachers: de Heering & Rossion, 2008; Harrison & Hole, 2009; Rhodes & Anastasi, 2012).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites de Heering & Rossion, 2008 source
    • cites Harrison & Hole, 2009 source
    • cites Rhodes & Anastasi, 2012 source

  • Probably not a substantial claim.

    Premisecitation

    Matching performance can be considerably impacted by the category of faces that are being matched.

    “and can be considerably impacted by the category of faces that are being matched.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Bruce et al., 2018 source

  • Probably not a substantial claim.

    Premisecitation

    The tendency for poorer performance on other-ethnicity faces can be reduced with exposure to faces of that ethnicity.

    “and that this tendency can be reduced with exposure to faces of that ethnicity (Meissner & Brigham, 2001).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Meissner & Brigham, 2001 source

  • Probably not a substantial claim.

    Premisecitation

    Matching performance can vary substantially between individuals.

    “matching performance can vary substantially between individuals (Bruce et al., 2018)”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Bruce et al., 2018 source

  • It may state two separate claims as one.

    Interpretationreasoning

    The current findings may indicate that SRs develop their face-space in a similar, but more efficient, way to typical individuals, leading to enhanced performance compared with controls even when viewing relatively unfamiliar classes of faces, but similar patterns of bias and limitation as those with typical face recognition abilities.

    “The current findings may indicate that SRs develop their face-space in a similar, but more efficient, way to typical individuals, leading to enhanced performance compared with controls even when viewing relatively unfamiliar classes of faces, but similar patterns of bias and limitation as those with typical face recognition abilities.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim. It may state two separate claims as one.

    Premisecitation

    Belanova et al. (2018) compared SR and control performance in a sequential matching task where pairs comprised two infant faces.

    “Belanova et al. (2018) compared SR and control performance in a sequential matching task where pairs comprised two infant faces.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Belanova et al., 2018 source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Study design and conductown data

    Due to these departures from normality, a nonparametric measure of sensitivity (A) and bias (b) was used.

    “Due to these departures from normality, we used a nonparametric measure of sensitivity (A) and bias (b; Zhang & Mueller, 2005).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Zhang & Mueller, 2005 source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Study design and conductnone given

    The SR sample size was calculated as having 80% power to detect large main effects (d > 1.2), for both within-subjects and between-subjects comparisons.

    “The sample size for the SR group was calculated as having 80% power to detect large main effects (d > 1.2), for both within-subjects and between-subjects comparisons.”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Interpretationreasoning

    Despite evidence of an own-age bias in these individuals, performance was nevertheless at a high level, indicating that the processes which underpin super-recognition for adult faces are also likely deployed when matching child faces.

    “Despite evidence of an own-age bias in these individuals, performance was nevertheless at a high level, indicating that the processes which underpin super-recognition for adult faces are also likely deployed when matching child faces.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim. It may state two separate claims as one.

    Study design and conductreasoning

    SRs had already completed the adult task during their initial screening session, and these scores were carried over to this study to avoid exposure effects.

    “As SRs had already completed the adult task during their initial screening session, we carried these scores over to this study to avoid exposure effects.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Premisecitation

    The face matching performance of Caucasian SRs did not extend to the faces of other ethnicities and was no better than non-SR native perceivers.

    “Bate, Bennetts, et al. (2019) found that the face matching performance of Caucasian SRs did not extend to the faces of other ethnicities and was no better than non-SR native perceivers.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Bate, Bennetts, et al., 2019

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

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Study design and conductnone given

    The low- and high-contact control groups completed both tests within the same session in a randomised counter-balanced design.

    “The low- and high-contact control groups completed both tests within the same session in a randomised counter-balanced design.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim.

    Study design and conductcitation

    The adult face matching test is sufficiently calibrated to discriminate between top-end performers.

    “The adult face matching test was developed in our laboratory for previous work (the Pairs Matching Test, Bate et al., 2018) and is sufficiently calibrated to discriminate between top-end performers.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Bate et al., 2018

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

  • Probably not a substantial claim.

    Study design and conductcitation

    Crawford and Garthwaite's (2002) modified t tests for single-case comparisons were used to determine whether each individual significantly outperformed controls on the two tests.

    “Crawford and Garthwaite’s (2002) modified t tests for single-case comparisons were used to determine whether each individual significantly outperformed controls on the two tests using the singlism.exe programme.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Crawford & Garthwaite, 2002 source

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Study design and conductown data

    The control sample size was calculated to give 80% power to detect moderate-to-large between-subject main effects (d > 0.72) and small-to-moderate within-subject main effects (d > 0.34) and interactions (d > 0.36) in the analysis of the control groups.

    “This control sample size was calculated to give 80% power to detect moderate-to-large between-subject main effects (d > 0.72) and small-to-moderate within-subject main effects (d > 0.34) and interactions (d > 0.36) in the analysis of the control groups (power calculations carried out in G*Power 3.1).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim.

    Interpretationreasoning

    If the SR or high-contact control group displayed any difference in performance between the two tasks, this would indicate a difference in the own-age bias compared with typical perceivers.

    “Thus, if the SR or high-contact control group displayed any difference in the performance between the two tasks, this would indicate a difference in the own-age bias compared with typical perceivers.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation

    Very recent work has begun to identify limitations in the deployment of SRs.

    “However, very recent work has begun to identify limitations in the deployment of SRs (Bate, Portch, et al., 2019).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Bate, Portch, et al., 2019

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

  • It may state two separate claims as one.

    Premisecitation

    SRs not only outperformed typical perceivers in the task but also showed relative decrements in performance for other-age faces.

    “SRs not only outperformed typical perceivers in the task but also showed relative decrements in the performance for other-age faces.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Belanova et al., 2018 source
31 of 31

  • The paper relies on this without support.

    High-contact controls had high levels of contact with young children.

    “One group of participants contained 20 nursery and primary school teachers (Mage = 33.9 years, standard deviation [SD] = 8.1) who had all worked without career breaks for at least 2 years immediately prior to participation in this study (these participants are subsequently referred to as “high-contact controls”).”

  • The paper relies on this without support.

    The equal difficulty calibration of the two matching tasks holds for SRs and high-contact controls.

    “In effect, this calibration removed any inequalities in task difficulty that may result from any underlying own-age bias or biases in stimuli selection and pairing.”

  • The paper relies on this without support.

    The SRs' adult task scores from initial screening are representative of their performance in this study.

    “As SRs had already completed the adult task during their initial screening session, we carried these scores over to this study to avoid exposure effects.”

  • The paper relies on this without support.

    The adult and child versions of the matching task were matched in difficulty.

    “because the two versions of the task were matched in difficulty”
    “Because the equal calibration in task difficulty accounted for an own-age bias in controls”

  • The paper relies on this without support.

    Unfamiliar face memory and matching represent quite different tasks.

    “Given unfamiliar face memory and matching represent quite different tasks”
5 of 5

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.

  • Not confirmedF(1, 38) = 0.001reported p = .983 · recomputed p = .975

    “F(1,38) = 0.001, p = .983”

  • ConsistentF(1, 38) = 0.11reported p = .742 · recomputed p = .742

    “F(1,38) = 0.110, p = .742”

  • ConsistentF(1, 38) = 0.017reported p = .896 · recomputed p = .897

    “F(1,38) = 0.017, p = .896”

  • ConsistentF(1, 38) = 0.463reported p = .501 · recomputed p = .500

    “F(1,38) = 0.463, p = .501”

  • ConsistentF(1, 38) = 0.355reported p = .555 · recomputed p = .555

    “F(1,38) = 0.355, p = .555”

  • ConsistentF(1, 38) = 0.198reported p = .659 · recomputed p = .659

    “F(1,38) = 0.198, p = .659”

  • ConsistentF(1, 38) = 0.017reported p = .898 · recomputed p = .897

    “F(1, 38) = 0.017, p = .898”

  • ConsistentF(1, 38) = 0.282reported p = .599 · recomputed p = .598

    “F(1, 38) = 0.282, p = .599”

  • ConsistentF(1, 38) = 0.139reported p = .712 · recomputed p = .711

    “F(1, 38) = 0.139, p = .712”

  • ConsistentF(1, 38) = 0.286reported p = .596 · recomputed p = .596

    “F(1, 38) = 0.286, p = .596”

  • ConsistentF(1, 38) = 3.735reported p = .061 · recomputed p = .061

    “F(1, 38) = 3.735, p = .061”

  • ConsistentF(1, 38) = 0.018reported p = .895 · recomputed p = .894

    “F(1, 38) = 0.018, p = .895”

  • ConsistentF(1, 38) = 3.471reported p = .07 · recomputed p = .070

    “F(1,38) = 3.471, p = .070”

  • Consistentt(9) = 4.511reported p = .001 · recomputed p = .001

    “t(9) = 4.511, p = .001”

  • Consistentt(8) = 4.042reported p = .004 · recomputed p = .004

    “t(8) = 4.042, p = .004”

  • Consistentt(47.55) = 2.259reported p = .029 · recomputed p = .029

    “t(47.55) = 2.259, p = .029”

  • Consistentt(9) = 0.79reported p = .45 · recomputed p = .450

    “t(9) = 0.790, p = .450”

17 of 17

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

  • High-contact controls' adult overall accuracy does not match hits and correct rejections

    Minor

    The adult test has equal numbers of matching and mismatching trials, so overall accuracy should be the mean of the hit and correct-rejection percentages. For high-contact controls, that is 63.50%, not the reported 67.08%.

    Reported 67.08 · recomputed 63.495

1 of 1

Citations

Cites

Flagged first, then by citations

Research works printed but not linked: 1

  1. Bate S., Bennetts R., Murray E., Portch E. (2020). Enhanced matching of children’s faces in “super-recognisers” but not high-contact controls. i-Perception, 11 (4), 1–12. https://doi.org/10.1177/2041669520944420

Cited by

Sorted by citations

Related

Metadata

Authors
Bate, S., Bennetts, R., Murray, E., & Portch, E.
Affiliations
  • Bournemouth University (GB)
  • Brunel University of London (GB)
Corresponding author
Bate, Sarah
ORCID
Venue
i-Perception
Publisher
SAGE Publications
Published
2020-07
Volume
11
Issue
4
Pages
2041669520944420-2041669520944420
ISSN
2041-6695 (ISSN-L)
DOI
10.1177/2041669520944420
PMC
PMC7385838
OpenAlex
W3045316290
Cited by
12
References
33
Citations by year
2026: 1 · 2025: 2 · 2022: 1 · 2021: 8
FWCI
1.08
Type
journal article
Language
en
Open access
Licence
Locations
  • i-Perception (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Bournemouth University Research Online (Bournemouth University) (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • Brunel University Research Archive (BURA) (Brunel University London) (submittedVersion)
Topics
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Face Recognition and Perception
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Face recognition and analysis
  • Social Sciences › Psychology › Experimental and Cognitive Psychology › Evolutionary Psychology and Human Behavior
Keywords
face matching, super-recognisers, own-age bias, face recognition (cognition), preschool teachers, 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.

A language model drew the claims from the paper on 2026-11-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-11-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (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