Finding a Face

A flyer on the wall in Cotton Street caught my eye.

Whether it was the photo of the single blue eye, or the offer of free movie passes I’m not sure, but I went for a closer look. Interesting – it was a survey being conducted by Dr Matthew Gerrie in the School of Psychology, examining how people recognise faces. I had to have good vision and not know too much psychology – that was OK – so I signed up.

The eye tracking camera in operation

Matt met me at level 6 in Easterfield and took me to the testing room. I signed a consent form, read some instructions and did a trial set of questions on the computer. Then my chin and forehead were steadied in a rest and my eye movements tracked using an infra-red camera above me. The camera is a pretty hi-tech piece of apparatus which works at 1000 frames per second to track eye movement in response to what is being observed on a screen.

In the test, I was first shown an image of a man (always with no facial hair and a shaved head) for about two seconds. Then I had an exercise to do - deciding if words appearing on the screen were real or false. They had found some wonderful ones - frugal and kwahl, apathy and fraded! Next was to select the image I’d been shown from a group of six similar faces. The identification task was made more complicated by having been told that the person may or may not appear in the group of faces. I was also asked to rate how confident I was about the identification I made.

A subject selecting a face from a group of six

I found myself reflecting on the way my eyes moved over the faces and how sometimes it was easy to find the face and sometimes much harder. Harder still was to know for certain whether the face was actually present in the group or not, if it didn’t ‘jump out’ at me.

Matt’s research is looking into how witnesses choose faces in an identity parade. He helpfully compares the processes used to how people approach multi choice questions in a test. Is a candidate choosing the right answer from memory or are they going through a process of elimination and selecting the most likely one? The process of elimination is fine for multi choice questions, but it can be disastrous if applied in an identity parade. In fact, the leading cause of wrongful criminal conviction is eyewitness error.

By looking at the patterns of eye movement in a test such as I undertook, researchers can study the decision-making processes which go on and find out if someone is using memory or elimination to find a particular face. Along with the speed of the decision making (faster is generally more accurate), the patterns can show how confident and therefore how accurate people are in their identifications of faces.

Dr Gerrie is a Research Fellow in the School of Psychology and completed his PhD at Victoria University with Professor Maryanne Garry. This work is being carried out as part of the Innocence Project New Zealand (IPNZ), a joint venture between Victoria University and the University of Otago, with support from the Marsden Fund. 

Sarah Wilcox - Web Editor, Science Faculty