You can believe your eyes: measuring implicit recognition in a lineup with pupillometry

ABSTRACT

As pupil size is affected by cognitive processes, we investigated whether it could serve as an independent indicator of target recognition in lineups. Participants saw a simulated crime video, followed by two viewings of either a target-present or target-absent video lineup while pupil size was measured with an eye-tracker. Participants who made correct identifications showed significantly larger pupil sizes when viewing the target compared with distractors. Some participants were uncertain about their choice of face from the lineup, but nevertheless showed pupillary changes when viewing the target, suggesting covert recognition of the target face had occurred. The results suggest that pupillometry might be a useful aid in assessing the accuracy of an eyewitness’ identification.

KEYWORDS:

• Pupillometry
• eyewitness identification
• covert recognition
• face processing
• eyewitness credibility

Acknowledgements

We thank Sam Hutton and SR Research for their excellent hardware and software support.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Camilla E. J. Elphick http://orcid.org/0000-0001-8482-0863

Graham E. Pike http://orcid.org/0000-0003-1013-8698

Graham J. Hole http://orcid.org/0000-0002-7073-2982

Notes

1 Ordinarily, the number of participants per condition is under the control of the experimenter. However, in this case, allocation to a condition depended entirely on the participants' identification decision. Our strategy was therefore adopted merely to ensure that we had a reasonable sample size for each type of decision, even those that occurred relatively infrequently. A pilot study showed that we were able to obtain reasonably reliable effects with only eight participants per condition, so our choice to keep testing until we had at least ten participants for each type of identification response was motivated by this consideration.

2 We were faced with two options, either to introduce an additional face to replace the target, or to delete the target. We thought that deleting the target was the lesser of two evils, as it meant that the only difference between the TP and TA lineups was that the TA lineup had one less face, and everything else was kept constant. Introducing a novel face runs the risk of changing the nature of the lineup. In experiment three, for practical reasons (increased speed of testing in a real-life situation) we used a different lineup size from the previous two experiments (and a different target person) and obtained similar effects, suggesting that lineup size (within the limits of those we used) was not a major influence on our results, although future studies should attempt to keep lineup size constant.