….but we are all fabulous.
I teach a module on eyewitness testimony and in preparation for the lecture on eyewitness identification, I ask the students to send me a composite picture of me using the Open University’s online PhotoFit Me widget: http://www.open.edu/openlearn/body-mind/photofit-me
Here are some of last year’s and this year’s efforts:
Of course current composite systems in use by the police are significantly more sophisticated than PhotoFit. E-FIT is a computerised system which enables witnesses to build up a composite of a suspect based on the features they remember. More advanced systems such as Evo-FIT are also being developed. As the Evo-FIT site reports “witnesses and victims select from screens of complete faces and a composite is ‘evolved’ over time. The system does not require eyewitnesses to have good recall of an offender’s face, unlike the traditional ‘feature’ methods, just to have seen it clearly. The approach aims to construct the most identifiable set of internal features, the central region of the face that is important for recognition by another person later.” (http://www.evofit.co.uk/how-it-works).
The point of the exercise is to demonstrate that it is incredibly hard to compile a composite photograph of someone we see regularly let alone of a perpetrator we might only have seen briefly in a stressful situation. Even with a relatively small set of attributes to choose from, these images differ considerably from one another. However, most of them do have overlapping similarities to each other and to my actual photograph. Professor Vicki Bruce and colleagues found that by combining facial composites from several witnesses, a more accurate likeness could be obtained.
I also use the images above to illustrate issues relating to sequential (one at a time) vs. simultaneous (altogether) presentations of images. This is a perennial debate in the literature for both selecting suspects from a set of photographs or from a video or physical line-up. We are much more likely to use a ‘which picture best matches my memory’ strategy with a simultaneous line-up even when we are warned that the perpetrator might not be present rather than true recognition. Sequential line-ups however have their own problems: whilst they do reduce false identification of the wrong person, they also reduce correct identification, although together these statistics do lead to greater overall accuracy (e.g., Steblay et al., 2001).
The Innocence Project in the US (and Innocence Network http://www.innocencenetwork.org.uk/ in the UK) provide concrete evidence of why accurate eyewitness identification is so vital. As of 1st November 2014, 321 wrongly convicted men and women have been exonerated thanks to DNA. Of those 321 cases, approximately 75% were due to eyewitness testimony and misidentification.
The Innocence Project website (http://www.innocenceproject.org) which has details of individual cases and lots of interesting facts and figures has some sobering facts:
On Monday, the students will vote on which is the best likeness and the winner will receive a prize. A light-hearted exercise to illustrate a very serious issue.
Bruce, V., Ness, H., Hancock, P. J. B., Newman, C., and Rarity, J. (2002). Four heads are better than one: Combining face composites yields improvements in face likeness. Journal of Applied Psychology, 87, pp. 894–902.
Steblay, N. M., Deisert, J., Fulero, S. & Lindsay, R.C.L. (2001). Eyewitness accuracy rates in sequential and simultaneous lineup presentations: A meta-analytic comparison. Law and Human Behavior, 25, 459-474.