Super-recognisers' long-term face memory
Updated: Aug 3, 2020
Key outcome: Most super-recognisers of faces retain their superiority over long delays. However, at least one-quarter of all super-recognisers classified as such using short-term memory tests cannot sustain their superior skills.
Key recommendation: Super-recognisers employed in occupations in which memory for faces is a key to success (i.e. will they recognise a fugitive after a delay of a week or so?) need to be assessed on their longer-term skills. Short-term face memory tests will not suffice.
New journal article
We are very pleased to announce a new journal article published in Applied Cognitive Psychology authored by three members of University of Greenwich staff – Josh Davis, Elena Belanova and Trevor Thompson and ex University of Greenwich BSc Psychology student: Diandra Bretfelean.
Davis, J. P., Bretfelean, D., Belanova, E., & Thompson, T. (2020). Super-recognisers: face recognition performance after variable delay intervals. Applied Cognitive Psychology, DOI:10.1002/acp.3712 (read pre-view here)
Diandra helped developed the tests included in the research for her undergraduate final year project. Diandra was Undergraduate Prize Winner at the British Psychological Society: Cognitive Section Annual Conference, at the University of Northumbria, Newcastle on 30 August 2017. Diandra presented a poster at the conference.
The journal article describes two experiments (total participants n = 2018), and primarily compares super-recognisers (n = 422) and controls (n = 580) of typical-range ability (i.e. roughly ‘average’ short-term face memory ability) on their memory for 10 (Experiment 1) or 20 (Experiment 2) people shown in Phase 1 in videos for up to 1 minute each.
In Phase 2, conducted almost immediately (Experiment 2 only), or with randomly varied delays of I day, 7 days, 14 days, 28 days, or 56 days (Experiment 1 only), participants viewed 10 (Experiment 1) or 20 (Experiment 2) line-ups (identity parades). Most eyewitness identification research uses one actor and one line-up, so the task demands were very high – particularly for participants in the longest delay conditions.
Was that you?
Participants were classified as super-recognisers or controls based on their Cambridge Face Memory Test: Extended (Russell, Duchaine, & Nakayama, 2009) scores, and we checked classifications using the Glasgow Face Matching Test (Burton, White, & McNeill, 2010).
Result 1: Longer delays were associated with reduced face recognition test scores in super-recognisers and controls.
Result 2: As a group, super-recogniser’s score superiority over controls at recognising faces seen previously was about the same after 1 day as after 1 month. In other words, decay of faces in memory seemed to be about the same in both groups. Super-recognisers start out from a higher place.
Result 3: As a group, at all delay intervals, super-recognisers significantly outperformed controls at rejecting a line-up as not containing a person seen before. In other words, they were consistently less likely to make a misidentification.
Result 4: Within these groups there were large individual differences, so that some super-recognisers (about 25%) scored worse than the control mean performance at each delay interval. Some controls outperformed the super-recogniser group mean.
Result 5: A proportion of super-recognisers were significantly superior to controls at each delay interval despite the small number of trials in the two experiments. These participants were likely to be ‘real’ super-recognisers.
Trial numbers were low (10 or 20), and line-ups contained 6 faces only (plus participants could respond “none of the above”). It would be possible to ‘guess’ some answers. Alternatively, any distractions in the learning phase will have impacted later scores.
Super-recogniser and control groups were based on two tests only. Reliability of group classification will likely be higher with the use of multiple tests. Davis (2019) discusses how super-recogniser recruitment for policing and other roles is best achieved.
A substantial minority of participants did not complete Phase 2. Why they dropped out is unknown.
All faces in the project were from white ethnic groups. It is not clear whether effects would be consistent with faces from other ethnicities, although recent research suggests that super-recognisers retain superiority over controls with other-race faces (Robertson, Black, Chamberlain, Megreya, & Davis, 2020).
All was online – so control over conditions reduced. It is not possible to estimate what distracting influences on their memory for the faces in the test participants may have encountered, particularly in the longer delay conditions. It is also not possible to estimate whether participants attempted rehearsal strategies to retain faces in memory. This is clearly an avenue for future research.
Burton, A. M., White, D., & McNeill, A. (2010). The Glasgow face matching test. Behavior Research Methods, 42, 286–291. doi:10.3758/BRM.42.1.286.
Davis, J. P. (2019).The worldwide public impact of identifying super-recognisers for police and business. The Cognitive Psychology Bulletin, 4, Spring 2019, 17-21. https://shop.bps.org.uk/the-cognitive-psychology-bulletin-issue-4-spring-2019
Robertson, D., Black, J., Chamberlain, B., Megreya, A. M., & Davis, J. P. (2020). Super-recognisers show an advantage for other race face identification Applied Cognitive Psychology, 34(1), 205-216. DOI: 10.1002/acp.3608
Russell, R., Duchaine, B., & Nakayama, K., (2009). Super-recognizers: People with extraordinary face recognition ability. Psychonomic Bulletin & Review, 16, 252–257. doi:10.3758/PBR.16.2.252