Daniel James Carragher
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Daniel James Carragher has a total of 35 co-authors. Here are the top co-authors by number of shared publications:
Peter J. B. Hancock
Psychology, Faculty of Natural Sciences, University of Stirling, Scotland, UK
7 shared publications- Mike Elmo Richard Nicholls
College of Education, Psychology and Social Work, Flinders University, Adelaide, Australia
6 shared publications Peter Hancock
5 shared publicationsDaniel Sturman
3 shared publications- Megan Louise Bartlett
School of Psychology, University of Adelaide
3 shared publications - Nicole Annette Thomas
School of Psychology, Flinders University
3 shared publications
Latest Publications
- Open Access
- 5 November 2026
- Published by Center for Open Science in PsyArXiv
Feedback Improves Automation-Assisted Unfamiliar Face Matching Performance in Novice Observers
Identity verification often involves deciding whether two images show the same person or two different people. Automated Facial Recognition (AFR) systems can improve unfamiliar face matching accuracy when used as decision aids, yet humans consistently fail to use them optimally, with average automation-assisted performance regularly below that of the system alone. This suboptimal performance may reflect participants’ difficulty in appropriately weighing their trust in the automation with self-confidence in their own ability. We investigated whether a brief training block with either trial-by-trial feedback, or simple exposure to the AFR system, would improve automation-assisted unfamiliar face matching performance in a subsequent task. Participants were randomly assigned to one of three training conditions: feedback (AFR system, trial-by-trial feedback), experience (AFR system, no feedback), or control (no AFR system, no feedback). While we found all conditions improved with assistance from the AFR system in the subsequent face matching task, the feedback condition improved significantly more than the experience and control conditions. Feedback specifically increased trust in automation, which led to greater reliance on the AFR system. However, greater reliance also led to greater acceptance of rare AFR system errors in the feedback condition. Our results show that a brief period of trial-by-trial feedback, but not mere exposure to the AFR system, can improve automation-assisted unfamiliar face matching performance, offering a promising direction for improving identification outcomes.
- 5 November 2026
- Published by Center for Open Science in PsyArXiv
Open AccessFeedback Improves Automation-Assisted Unfamiliar Face Matching Performance in Novice Observers
Identity verification often involves deciding whether two images show the same person or two different people. Automated Facial Recognition (AFR) systems can improve unfamiliar face matching accuracy when used as decision aids, yet humans consistently fail to use them optimally, with average automation-assisted performance regularly below that of the system alone. This suboptimal performance may reflect participants’ difficulty in appropriately weighing their trust in the automation with self-confidence in their own ability. We investigated whether a brief training block with either trial-by-trial feedback, or simple exposure to the AFR system, would improve automation-assisted unfamiliar face matching performance in a subsequent task. Participants were randomly assigned to one of three training conditions: feedback (AFR system, trial-by-trial feedback), experience (AFR system, no feedback), or control (no AFR system, no feedback). While we found all conditions improved with assistance from the AFR system in the subsequent face matching task, the feedback condition improved significantly more than the experience and control conditions. Feedback specifically increased trust in automation, which led to greater reliance on the AFR system. However, greater reliance also led to greater acceptance of rare AFR system errors in the feedback condition. Our results show that a brief period of trial-by-trial feedback, but not mere exposure to the AFR system, can improve automation-assisted unfamiliar face matching performance, offering a promising direction for improving identification outcomes.
- Open Access
- 3 June 2026
- Published by SAGE Publications in Quarterly Journal of Experimental Psychology
- Vol. 79 (11) , 2960-2978
Two Out of Three Ain’t Bad: Aggregating Multiple Identification Decisions from the Same Observer Improves Unfamiliar Face Matching Performance
Deciding whether two images show the same person sounds easy, but the average individual achieves accuracy of 60% to 90% on many unfamiliar face matching tasks. However, overall task performance can be improved by combining the identification decisions made by different individuals through the “wisdom of the crowd” effect. We investigated whether similar gains occur when combining multiple decisions made by the same individual, via a “wisdom of the crowd within” effect. Participants completed the same unfamiliar face matching task three times: in one session (Experiment 1), or over 3 weeks (Experiment 2). We found a wisdom of the crowd within effect in both experiments, which became significant after two responses and was larger with three. Surprisingly, the delay did not offer additional benefit. Though the crowd within improved performance, crowds of different individuals produced larger gains. Our results contribute to a growing literature showing that combined decision-making can improve unfamiliar face matching performance.
- 3 June 2026
- Published by SAGE Publications in Quarterly Journal of Experimental Psychology
- Vol. 79 (11) , 2960-2978
Open AccessTwo Out of Three Ain’t Bad: Aggregating Multiple Identification Decisions from the Same Observer Improves Unfamiliar Face Matching Performance
Deciding whether two images show the same person sounds easy, but the average individual achieves accuracy of 60% to 90% on many unfamiliar face matching tasks. However, overall task performance can be improved by combining the identification decisions made by different individuals through the “wisdom of the crowd” effect. We investigated whether similar gains occur when combining multiple decisions made by the same individual, via a “wisdom of the crowd within” effect. Participants completed the same unfamiliar face matching task three times: in one session (Experiment 1), or over 3 weeks (Experiment 2). We found a wisdom of the crowd within effect in both experiments, which became significant after two responses and was larger with three. Surprisingly, the delay did not offer additional benefit. Though the crowd within improved performance, crowds of different individuals produced larger gains. Our results contribute to a growing literature showing that combined decision-making can improve unfamiliar face matching performance.
- Open Access
- 10 May 2026
- Published by Center for Open Science in PsyArXiv
Unfamiliar Faces Attenuate the Boundary Extension Effect
Boundary Errors occur when an observer misremembers the boundaries of a scene, either remembering being farther (Boundary Extension, BE) or closer (Boundary Restriction, BR) than they actually were. These errors can be influenced by the composition of an image (e.g., stimuli salience) and by forms of emotional arousal like state and trait anxiety. Despite their salient nature, face stimuli have rarely been studied in the context of BE. Here we investigate whether the familiarity of faces is associated with BE. Participants (n = 95) viewed 40 images of room interiors, each with a single face (20 famous, 20 non-famous) superimposed into the scene centre. After a brief filler task, participants viewed the same images, and rated whether the viewing distance had changed. Overall, participants made BE errors for faces, which were greater for famous than non-famous faces. Crucially, this pattern was replicated when we re-analysed the famous faces, categorised by participants’ individual familiarity with the faces. Participants’ trait anxiety did not influence these results. Our results suggest that arousal may contribute to BE attenuation, but pleasant arousal co-occurs with these effects and does not fully explain them.
- 10 May 2026
- Published by Center for Open Science in PsyArXiv
Open AccessUnfamiliar Faces Attenuate the Boundary Extension Effect
Boundary Errors occur when an observer misremembers the boundaries of a scene, either remembering being farther (Boundary Extension, BE) or closer (Boundary Restriction, BR) than they actually were. These errors can be influenced by the composition of an image (e.g., stimuli salience) and by forms of emotional arousal like state and trait anxiety. Despite their salient nature, face stimuli have rarely been studied in the context of BE. Here we investigate whether the familiarity of faces is associated with BE. Participants (n = 95) viewed 40 images of room interiors, each with a single face (20 famous, 20 non-famous) superimposed into the scene centre. After a brief filler task, participants viewed the same images, and rated whether the viewing distance had changed. Overall, participants made BE errors for faces, which were greater for famous than non-famous faces. Crucially, this pattern was replicated when we re-analysed the famous faces, categorised by participants’ individual familiarity with the faces. Participants’ trait anxiety did not influence these results. Our results suggest that arousal may contribute to BE attenuation, but pleasant arousal co-occurs with these effects and does not fully explain them.
- Open Access
- 9 May 2026
- Published by Elsevier in SSRN Electronic Journal
Decision Criterion Masks the Positive Relationship Between Match and Mismatch Performance in Unfamiliar Face Matching
Identity verification often relies on an observer comparing two faces to determine identity matches (same person) from mismatches (different people). When presented with familiar faces, research shows a strong positive correlation between one’s performance on identity match and mismatch trials, suggesting a unitary process underlying decision-making. However, no such correlation is found for unfamiliar faces, which may indicate separable cognitive processes when familiarity is not available to draw upon. Across two experiments, we replicated this positive association between familiar face matching decisions, but report several surprising negative correlations between unfamiliar face matching decisions across multiple tasks. However, when the effect of response bias was removed from accuracy, positive associations were found in unfamiliar face matching performance. We also show that factors related to the images themselves, such as visual similarity and task difficulty, contribute to these correlations. Therefore, there is evidence that response bias obscures a positive relationship between unfamiliar face matching decisions, and there may be a singular underlying process rather than truly separable cognitive processes. Together, these findings suggest that face matching decisions are influenced by response bias, but the development of this bias can be a person-specific tendency while also being influenced by task-dependent demands.
- 9 May 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessDecision Criterion Masks the Positive Relationship Between Match and Mismatch Performance in Unfamiliar Face Matching
Identity verification often relies on an observer comparing two faces to determine identity matches (same person) from mismatches (different people). When presented with familiar faces, research shows a strong positive correlation between one’s performance on identity match and mismatch trials, suggesting a unitary process underlying decision-making. However, no such correlation is found for unfamiliar faces, which may indicate separable cognitive processes when familiarity is not available to draw upon. Across two experiments, we replicated this positive association between familiar face matching decisions, but report several surprising negative correlations between unfamiliar face matching decisions across multiple tasks. However, when the effect of response bias was removed from accuracy, positive associations were found in unfamiliar face matching performance. We also show that factors related to the images themselves, such as visual similarity and task difficulty, contribute to these correlations. Therefore, there is evidence that response bias obscures a positive relationship between unfamiliar face matching decisions, and there may be a singular underlying process rather than truly separable cognitive processes. Together, these findings suggest that face matching decisions are influenced by response bias, but the development of this bias can be a person-specific tendency while also being influenced by task-dependent demands.
- Open Access
- 7 May 2026
- Published by Center for Open Science in PsyArXiv
Identity in the Spotlight: Matching Faces Without Overlapping Features
One-to-one face matching is widely used for identity verification yet becomes difficult with unfamiliar faces or when parts of the face are covered. We investigated whether observers could accurately verify identity from different non-overlapping facial fragments. Each stimulus of our novel face matching task had the appearance of a single face image as revealed through four circular 'spotlights,' one from each quadrant of the face. Observers judged whether the designated target spotlight (from image A) depicted the same identity as the three reference spotlights sampled from a different image (all from image B). Across three experiments, we examined observers' matching performance for familiar and unfamiliar faces under Baseline (Experiment 1), Scrambled (Experiment 2), and Inverted (Experiment 3) spotlight arrangements. Accuracy reliably exceeded chance across all but one condition: familiar inverted faces. A familiar-face advantage emerged under baseline conditions, disappeared when the spatial layout was scrambled, and reversed under inversion —while unfamiliar face accuracy remained robust to these manipulations. These findings demonstrate that observers can exploit latent identity information without direct feature correspondence to achieve above chance accuracy in identification and reveal an unexpected dissociation between familiar and unfamiliar face matching when spatial orientation is disrupted in these highly challenging conditions.
- 7 May 2026
- Published by Center for Open Science in PsyArXiv
Open AccessIdentity in the Spotlight: Matching Faces Without Overlapping Features
One-to-one face matching is widely used for identity verification yet becomes difficult with unfamiliar faces or when parts of the face are covered. We investigated whether observers could accurately verify identity from different non-overlapping facial fragments. Each stimulus of our novel face matching task had the appearance of a single face image as revealed through four circular 'spotlights,' one from each quadrant of the face. Observers judged whether the designated target spotlight (from image A) depicted the same identity as the three reference spotlights sampled from a different image (all from image B). Across three experiments, we examined observers' matching performance for familiar and unfamiliar faces under Baseline (Experiment 1), Scrambled (Experiment 2), and Inverted (Experiment 3) spotlight arrangements. Accuracy reliably exceeded chance across all but one condition: familiar inverted faces. A familiar-face advantage emerged under baseline conditions, disappeared when the spatial layout was scrambled, and reversed under inversion —while unfamiliar face accuracy remained robust to these manipulations. These findings demonstrate that observers can exploit latent identity information without direct feature correspondence to achieve above chance accuracy in identification and reveal an unexpected dissociation between familiar and unfamiliar face matching when spatial orientation is disrupted in these highly challenging conditions.
- Open Access
- 30 March 2026
- Published by Center for Open Science in PsyArXiv
Stimulus Facial Hair Impairs Unfamiliar Face Matching Performance
Many identification tasks require human observers to decide whether the photograph on an identification document matches the person presenting it for inspection. Though this is a common task, average unfamiliar face matching accuracy in laboratory tests is often between 60-90%, even when the photographs are front facing, relatively high quality, and unobstructed. Unnatural obstructions such as reading glasses and surgical face masks further impair the accuracy with which two faces can be matched. Here we examined whether natural facial hair, which covers a similar area of the lower face as a surgical mask, also affects human performance in an unfamiliar face matching task. In Experiment 1, facial hair impaired unfamiliar face matching accuracy by approximately 8%, regardless of whether one or both faces in the pair wore a beard, compared to when the same identities were shown cleanshaven. Results from Experiment 2 indicated that natural facial hair does not provide identity information that participants used reliably in the matching task, with similar levels of accuracy found whether faces wore natural beards, fake beards, or surgical masks. Our findings suggest that beards are obstructions of the lower face that can pose a challenge to accurate identity verification.
- 30 March 2026
- Published by Center for Open Science in PsyArXiv
Open AccessStimulus Facial Hair Impairs Unfamiliar Face Matching Performance
Many identification tasks require human observers to decide whether the photograph on an identification document matches the person presenting it for inspection. Though this is a common task, average unfamiliar face matching accuracy in laboratory tests is often between 60-90%, even when the photographs are front facing, relatively high quality, and unobstructed. Unnatural obstructions such as reading glasses and surgical face masks further impair the accuracy with which two faces can be matched. Here we examined whether natural facial hair, which covers a similar area of the lower face as a surgical mask, also affects human performance in an unfamiliar face matching task. In Experiment 1, facial hair impaired unfamiliar face matching accuracy by approximately 8%, regardless of whether one or both faces in the pair wore a beard, compared to when the same identities were shown cleanshaven. Results from Experiment 2 indicated that natural facial hair does not provide identity information that participants used reliably in the matching task, with similar levels of accuracy found whether faces wore natural beards, fake beards, or surgical masks. Our findings suggest that beards are obstructions of the lower face that can pose a challenge to accurate identity verification.
- Open Access
- 21 November 2025
- Published by SAGE Publications in Quarterly Journal of Experimental Psychology
- Vol. 79 (7) , 1631-1643
Time Pressure Increases Automation Reliance in a Face Matching Task
Automated facial recognition (AFR) systems are commonly used to help verify the identity of individuals, often in a one-to-one face matching task. The AFR system may be a decision aid, such that a human operator must ultimately approve the final identification decision. Previous research has shown that time pressure, a common operational factor in many applied settings, impairs human face matching accuracy. We investigated whether time pressure influences human reliance on AFR decisions in a face matching task, predicting that participants would show greater automation reliance as time pressure increased and the task became harder. Each participant ( n = 129) completed a face matching task under blocks of high (2s), medium (5s), and low time pressure (10s), where the stimuli disappeared from the screen after the allocated time. For each pair of faces, participants made an initial identification decision without assistance. Participants were then shown an identification decision from a simulated AFR system (92.3% accurate), before they were asked to submit their final (assisted) decision to the same trial. As predicted, human accuracy improved with AFR assistance in all time pressure conditions, with the greatest improvements occurring under higher time pressure. But replicating previous studies, average aided human accuracy was below that of the AFR system alone. Moreover, assisted accuracy fell in all conditions when the AFR system provided an incorrect decision, suggesting participants struggled to correct system errors. Our results have implications for human oversight of AFR systems in applied face identification scenarios.
- 21 November 2025
- Published by SAGE Publications in Quarterly Journal of Experimental Psychology
- Vol. 79 (7) , 1631-1643
Open AccessTime Pressure Increases Automation Reliance in a Face Matching Task
Automated facial recognition (AFR) systems are commonly used to help verify the identity of individuals, often in a one-to-one face matching task. The AFR system may be a decision aid, such that a human operator must ultimately approve the final identification decision. Previous research has shown that time pressure, a common operational factor in many applied settings, impairs human face matching accuracy. We investigated whether time pressure influences human reliance on AFR decisions in a face matching task, predicting that participants would show greater automation reliance as time pressure increased and the task became harder. Each participant ( n = 129) completed a face matching task under blocks of high (2s), medium (5s), and low time pressure (10s), where the stimuli disappeared from the screen after the allocated time. For each pair of faces, participants made an initial identification decision without assistance. Participants were then shown an identification decision from a simulated AFR system (92.3% accurate), before they were asked to submit their final (assisted) decision to the same trial. As predicted, human accuracy improved with AFR assistance in all time pressure conditions, with the greatest improvements occurring under higher time pressure. But replicating previous studies, average aided human accuracy was below that of the AFR system alone. Moreover, assisted accuracy fell in all conditions when the AFR system provided an incorrect decision, suggesting participants struggled to correct system errors. Our results have implications for human oversight of AFR systems in applied face identification scenarios.
- Open Access
- 3 October 2025
- Published by Center for Open Science in PsyArXiv
Two out of three ain’t bad: Aggregating multiple identification decisions from the same observer improves unfamiliar face matching performance
Deciding whether two images show the same person sounds easy, but the average individual achieves accuracy of 60-90% on many unfamiliar face matching tasks. However, overall task performance can be improved by combining the identification decisions made by different individuals, through the “wisdom of the crowd” effect. We investigated whether similar gains occur by combining multiple decisions made by the same individual, via a “wisdom of the crowd within” effect. Participants completed the same unfamiliar face matching task three times: in one session (Experiment 1), or over three weeks (Experiment 2). We found a wisdom of the crowd within effect in both experiments, which was significant after two responses, and larger with three. Surprisingly, the delay did not offer additional benefit. Though the crowd within improved performance, crowds of different individuals produced larger gains. Our results add to a growing literature showing combined decision-making can improve unfamiliar face matching performance.
- 3 October 2025
- Published by Center for Open Science in PsyArXiv
Open AccessTwo out of three ain’t bad: Aggregating multiple identification decisions from the same observer improves unfamiliar face matching performance
Deciding whether two images show the same person sounds easy, but the average individual achieves accuracy of 60-90% on many unfamiliar face matching tasks. However, overall task performance can be improved by combining the identification decisions made by different individuals, through the “wisdom of the crowd” effect. We investigated whether similar gains occur by combining multiple decisions made by the same individual, via a “wisdom of the crowd within” effect. Participants completed the same unfamiliar face matching task three times: in one session (Experiment 1), or over three weeks (Experiment 2). We found a wisdom of the crowd within effect in both experiments, which was significant after two responses, and larger with three. Surprisingly, the delay did not offer additional benefit. Though the crowd within improved performance, crowds of different individuals produced larger gains. Our results add to a growing literature showing combined decision-making can improve unfamiliar face matching performance.
- Open Access
- 1 September 2024
- Published by Wiley in Applied Cognitive Psychology
- Vol. 38 (5)
The Wisdom of the Crowd Can Unmask Faces
Accurately determining whether two images show the same person is a surprisingly difficult task, which becomes even harder if one or both faces are wearing medical face masks. Attempts to improve unfamiliar face matching accuracy have generally had limited success. Although one brief training program improved masked face matching accuracy by 4.9%, this increase would not overcome the entire performance deficit caused by masks. Here, we investigate whether combining independent identification decisions from different individuals can improve masked face matching performance through the wisdom of the crowd effect. Accuracy gains emerged reliably after combining the decisions of three individuals, culminating in significant improvements of 11%–26% among the largest crowds. Despite creating crowds of up to 80 people, half of the eventual improvement had generally already occurred in crowds of 6 individuals. The wisdom of the crowd effect is a highly effective approach to improving accuracy when identifying masked faces.
- 1 September 2024
- Published by Wiley in Applied Cognitive Psychology
- Vol. 38 (5)
Open AccessThe Wisdom of the Crowd Can Unmask Faces
Accurately determining whether two images show the same person is a surprisingly difficult task, which becomes even harder if one or both faces are wearing medical face masks. Attempts to improve unfamiliar face matching accuracy have generally had limited success. Although one brief training program improved masked face matching accuracy by 4.9%, this increase would not overcome the entire performance deficit caused by masks. Here, we investigate whether combining independent identification decisions from different individuals can improve masked face matching performance through the wisdom of the crowd effect. Accuracy gains emerged reliably after combining the decisions of three individuals, culminating in significant improvements of 11%–26% among the largest crowds. Despite creating crowds of up to 80 people, half of the eventual improvement had generally already occurred in crowds of 6 individuals. The wisdom of the crowd effect is a highly effective approach to improving accuracy when identifying masked faces.
- Open Access
- 8 August 2024
- Published by Elsevier in Applied Ergonomics
- Vol. 121, 104364
Benchmarking automation-aided performance in a forensic face matching task
- 8 August 2024
- Published by Elsevier in Applied Ergonomics
- Vol. 121, 104364
Open AccessBenchmarking automation-aided performance in a forensic face matching task