Think you’re good with faces? In fact, you probably don’t know much about your own face-recognition skills

Life would be awfully confusing if we weren't able to recognise familiar faces. It's a skill most of us take for granted, and we rarely stop to consider the impressive cognitive wizardry involved. But some of us are better at it than others: in the last decade or so it's become apparent that around two per cent of the population are born with a severe face-recognition impairment (known as congenital prosopagnosia), that there is a similar proportion of "super-recognisers" with unusually exceptional face-recognition skills, and that the rest of us are on a spectrum in between.

Where do you think your abilities lie? A new study in the Quarterly Journal of Experimental Psychology suggests that, unless you are severely impaired at face-recognition, you probably don't have much insight into this question. When participants were confronted with the question: "Overall, from 1-'very poor' to 9-'very good', how would you describe your general ability to recognise faces?", the research found that most participants' answers bore no relation to their performance on a range of lab-based face-recognition tests.

Romina Palermo at the University of Western Australia led a large multi-national team that tested the face-recognition skills of nearly 300 undergrads recruited in Italy, Australia, and Belgium. The research involved established face-recognition tests and specially designed new ones. These tasks required participants to look at specific faces and then pick them out from among a larger range of faces later on, including in new lighting conditions and from different angles. One newly designed task involved watching film clips of people conversing, and then later identifying the actors' faces from a display that featured several new faces.

Before they completed the face-recognition challenges, the participants answered questions about how well they thought they would do, including the single item question mentioned above or another that read "How well do you think you will perform (from 1 to 9) on studying and recognising faces from your own race". There was also an established 15-item survey that involved questions like "I can easily follow actors in a movie" but this scale has been criticised for also including a few items not related specifically to face-recognition; and finally, there was a new, mammoth in-depth survey with 77-items that tapped participants' beliefs about their basic face-recognition skills, but also what friends and family said about their skills, and there were other items covering experiences in specific contexts, such as "I have difficulty recognizing a colleague / student if I see his/her face outside my workplace / study."

Overall, there was no correlation between participants' scores on the single-item questions about their abilities and their actual face-recognition performance. There was minimal correlation between the 15-item survey and one of the face tests (the correlation was -0.14; note that higher scores on this test equal greater perceived difficulty hence the negative correlation). And there was a moderate correlation (0.30) between answers to the 77-item survey and test performance: this modest level of insight is similar to what's seen when people are asked about their general memory abilities. So we might be capable of some insight into our face-recognition skills, but only if we're led through a comprehensive series of nearly 80 questions, and even then it's likely our answers will only correlate modestly with our actual performance.

The researchers also tested 13 participants who had registered with a congenital prosopagnosia website as having everyday problems with face recognition: they rated themselves as having poorer face-recognition skills than the students on the 15-item survey, which isn't really a surprise as they complained of face-recognition problems. But crucially the prosopagnosics were no better than the students at identifying their level of ability relative to other people with prosopagnosia. In other words, they knew they had problems, but they didn't have much insight into just how bad they were. However, a larger sample is needed to confirm this result.

"In summary," the researchers said, "the ability of self-report questionnaires to measure insight into face recognition ability, even when the items concentrate on face recognition, appears limited." They added that: "People with very poor face recognition skills may be more aware of their difficulties than the typical population, but such insight is clearly not universal."

Assuming these results are accurate, why are most of us fairly useless at judging our face recognition abilities? The researchers speculated that part of the reason might be that, compared with, say, our language skills, which are tested regularly through schooling, we never actually receive any formal tests of our face recognition abilities. Also, they suggested that perhaps our subjective sense of our facial recognition skills is easily confused by our sense of our more general person-recognition abilities, which could be based on other cues, such as clothing, gait or tone of voice.

One limitation of this research was its focus on young adults. Face-recognition skills are known to mature and peak at around age 30. It will be interesting to see if middle-aged and older adults have better insight into their face identification skills than young people do.

Further reading

—Do people have insight into their face recognition abilities?

Image via Lisa Cee/Flickr