Naked Science Forum
General Science => General Science => Topic started by: Eric A. Taylor on 19/05/2010 09:20:34
-
I've read that mug shot line ups aren't really all that reliable. Especially if the real perpetrator is not included in the line up. Most people will assume that the person who committed the crime MUST be in the line up so they usually pick the person who most resembles the face they remember.
Here's an idea for an experiment on this subject, but I'm not sure it's entirely ethical. You take a person who is going to see the doctor. When the doctor walks in he or she is not the patient's regular doctor but a someone they haven't seen before (you would, of course need to use real medical doctors for this) The doctor performs the exam just as normal, perhaps writes a prescription and sends the person on their way.
Part 2 of the experiment takes place the next day (at least 24 hours later). "Investigators" show up at the patient's home and ask them to look at a photo line up. You would have 4 groups of investigators. Group 1 will show a line up with the doctor's photo included, group 2would NOT have the doctor's photo included but include a photo that kinda sorta looks like the doctor. Groups 3 and 4 will be the same except they will remind the patents that the doctor's photo may not be in this line up.
After the person picks it can be relieved to them that this is an experiment and that nothing untoward took place. That their doctor was in fact a real qualified doctor.
I think this would be a good way to find out because while you don't spend more than 15 minutes with your doctor you are in close contact with them. Also the subjects would be unaware that they need to remember the doctors face so they won't be trying to memorize it.
Not sure if it would be entirely ethical though. I can't think of a reason the "investigators" would need to know any details of the medical visit so no privacy issues need to come into play.
-
Eric - I like it. First off you could remove the doctor and replace with some other form of interaction that does not raise such profound ethical problems - a common 'trick' is to use a request for market research paid in vouchers. I would want to include a third (and sixth) group where there is no deliberate resemblance. And of course the investigators must not know to which group they belong!
To those who are not concerned with witness identification a simple video experiment is the easiest way to convince:
http://www.theinvisiblegorilla.com/videos.html - try the top one
For those jaded cynics who have seen it, done it, and got the t-shirt; watch the monkey business video further down.
Eric, I think the reason that these experiments are not studied by every law and justice agency is that they do not wish to compromise the holy grail of evidence; a disinterested stranger saying 'I saw him do it'. Practitioners know the fragility of this form of evidence but they also know the great persuasive power on a jury.
-
I saw a really interesting talk show on this subject. I think it was Montel Williams but I'm not sure. Don't normally watch these kinds of show but this one was on the unreliability of eye witness testimony.
The host showed a prerecorded tape of himself doing some kind of interview when, from off camera someone assaulted the host then stole his wallet and ran off.
The audience had been told before hand they were going to see some sort of crime committed but not what the crime would be. After the video the host brought on several people for the line up and the audience then voted which of the people they thought had done it.
Despite the fact that the audience had been TOLD they were about to see a crime hardly anyone voted for the correct perpetrator. Several people thought the perpetrator was black (the host was black) some thought the perpetrator was a woman, and one lady said the perpetrator was one of the camera men filming the show!
I heard of another experiment where a large group of people were told they were going to see a car crash and that they would be asked for details after the crash. A large dump truck was then crashed into the back of a small passenger car. The damage was incredible. The rear bumper of the car had been pushed forward to where the front seats were! The truck hit the car at just 35 miles per hour but most of the witnesses stated the truck was going 70 or more. No one was even close to the correct speed for the truck. The slowest speed given was 55 MPH.
They all really thought the truck was going very fast!