Eyewitness evidence

[Tony Williams]

There's a fascinating experiment which involves a member of the public, picked at random, who is walking along a path. An experimenter stops the subject and asks him/her for some directions. Part-way through the conversation, two men carrying a large board pass between the experimenter and the subject, so that the subject's view of the experimenter is completely blocked. Hidden from the subject, there is another man walking along behind the board, and he switches places with the experimenter and continues the conversation with the subject.

This experiment has been run several times and usually results in about half of the subjects not noticing the switch, despite the fact that the experimenter's appearance and dress are only generally similar to his replacement's.

Another experiment was laid on by the British police some years ago, at a magistrates' conference. As the magistrates left the building at the end of a session, the police staged a near-accident involving some cars, right in front of them. The magistrates were then asked to describe what they had seen, and the results were instructive: their accounts varied greatly in the sequence of the events, the colour of the vehicles and even the number of vehicle involved.

There are many examples from case law when witnesses have been absolutely certain that they have identified the person they clearly saw committing a crime, only to be proved wrong. Very worrying...

[Dr B]

Hi Tony

The experiment you mention first is one carried out by Dan Simons and is on an attentional effect known as change-blindness.

Other researchers that you might want to read about are Ron Rensink and his group.

It's basically nothing new. In the 1960s it was short-term memory failure, in the 1970s it was working memory failure. Now its been given a new name and a lot of people get excited about it.

I find it fascinating - but it is a re-branding of an effect with a long history in cognitive psychology.

Rensink has an interesting detailed model (indeed he is one of the few that actually gets to grips with the psychology and is not just seduced by the striking effect) that you might want to read about.

A related and even more striking phenomena is that of sustained inattentional-blindness where observers can be unaware of salient information that is clearly present in the visual field. Detailed models of this effect are even thinner on the ground - but i know of one British scientist who has proposed an important one (with colleagues )

Other reasons for poor eye-witness testimony include false-memory and you should take a look at the work of Loftus, and Johnson for some nice papers on that.

[Tony Williams]

Thanks for that...I find the continuing discoveries about the often bizarre ways in which the human mind works (or doesn't work) really fascinating.

I think that the worst case of faulty eyewitness evidence I ever heard of came from the American woman (intelligent and well-educated) who was raped in broad daylight. She had several minutes in which to study his face at a distance of a few inches. She picked him out of a line-up and was certain she had the right man. She maintained this certainty through his failed appeal. Only when incontrovertible evidence emerged that it could not possibly be him (and a known rapist eventually admitted the deed) did she realise that she had made a terrible mistake.

One factor was probably that she was white and the rapist was black - humans seem to be much more sensitive to facial differences in people of their own race ("they all look the same!"). The other was that, having made the initial error at the line-up, the person she was remembering the rest of the time wasn't the rapist, but the man she saw in the line-up.

A salutary example...

[polomint38]

One factor was probably that she was white and the rapist was black - humans seem to be much more sensitive to facial differences in people of their own race ("they all look the same!").

This is true but I have found that people who are brought up in areas with a very mixed population racially, are better at distinguishing those of races other than their own, than those brought up in primarily single race areas.

This is just by my observation, I no of no studies to back this up sorry, but what are you thoughts.

[Tony Williams]

This is true but I have found that people who are brought up in areas with a very mixed population racially, are better at distinguishing those of races other than their own, than those brought up in primarily single race areas.

This is just by my observation, I no of no studies to back this up sorry, but what are you thoughts.

That seems entirely reasonable to me. Recognition depends to a great extent on familiarity.

[farmersboy]

'tis true - I was brought up in a white community (1 black kid in 11 years of school), and 'they' all look the same to me, pretty much.

[migomigs]

Eyewitnesses in a courtroom may play the role of the storyteller. He is important not because he is telling the truth but because he may lead to the truth by telling the story that doesn't have flaws, when viewed at his perspective. It may look incorrect at other perspective but it should be consistent at least at one perspective. Who else could tell a flawless stroy than the one who has been present at that time. It would be easier to convey a point when presenting evidences in a logical manner, in this case, siting the evidences at specific points of time and space in a sequence of events told by eyewitnesses.

[Mulder]

Eyewitnesses in a courtroom may play the role of the storyteller. He is important not because he is telling the truth but because he may lead to the truth by telling the story that doesn't have flaws, when viewed at his perspective.

The word 'truth' is a tricky one with regard to witnesses. Many witnesses may tell the 'truth' in that they relate what they remember. However, it is frequently not the same as what actually happened. This is easily demonstrated by staging an 'incident' in front of witnesses and then asking them what happened. The tendency to remove 'flaws' in a memory is the cause of this problem. Witnesses will often unconsciously change a memory (confabulate) so that it 'fits' with the view they take of the incident after it is over. I think relying on witness memory alone in a court is dangerous.

[Dr B]

Hi Tony

In some senses though, recognition and Familiarity are not the same thing. You can recognise a teapot as a teapot - but feal no familiarity torwards towards it (even if it is your teapot). Recognising it and feeling it as familiar do appear to be largely separable processes.

This is also indexed by recent research on face processing as well as in autobiographical memory research.

Consider Capgras delusion - where the patient thinks that a relative (often a spouse) has been replaced by an imposter. The patient 'recognises' the individual as their wife / husband - but, due to brain lesions, feels no 'familiarity' towards them. This then is thought to underlie why such patients generate delusory explanations such as 'an imposter' - they are trying to explain why they feel no familiarity for what should be the familiar.

Fregoli delusion is another example but from a slightly different angle. Note - these are delusions of familiarity and not problems with recognition.

[StoneQuarry]

I prove this to myself often.
Sometimes I am lying in bed at night and think I see something moving across the room.
I turn my head only to realize that it was a shadow. This is usually pareidolia due to the fact that I am tired, the fan is spinning, cars are moving outside and other factors. I realized I can never trust my eyes 100% and should pay very close attention if/when I ever see a crime happening.