Science Buddies: "Ask an Expert"

I'm just starting a project on how criminal lineup affects memory of a crime and who committed it. the introduction to this project says:

"There is a significant amount of controversy surrounding the accuracy of eyewitness identifications used in the courtroom. Lineups are frequently used to capture witness testimony, however, they also raise a number of questions about the accuracy of the criminal identification process. This project will test one part of the identification process, the lineup, to determine the level of accuracy of eyewitness identification using four different lineup techniques."

It tells you that to conduct an experiment on the subject you are supposed to tape a "crime" and show it to a group of people (possibly a class) and see which lineup is most efficient. I wasn't sure if I could tape the crime with people that all of the survey groups knew well and still have a valid report. I also wasn't sure about how to film the crime and not completely give away who did it. Another question was how to do the lineup-with everyone in the movie or suspect look-a-likes. If anyone can offer any help it would be very appreciated. Thanks!!!

~Cate~

(i have a very small school, so everyone knows each other like family. this could be a problem.)

Hi Cate,

This sounds like a very interesting project. I've heard of a lot of controversy associated with testimonies and lineups.

You mention in your post that you're going to use four different lineup techniques. What are the four different methods? This will give us a better idea of how you may go about taping the crime and setting up the experiment.

Depending on what variables you are going to change in the lineup, using people that the subjects know (like other students) may be acceptable, or not. It all depends on what you're going to test.

I'd also suggest trying to make the "crime" as realistic as possible. Many times, when subjects know that they are being tested, they respond differently than if the circumstances mimic reality.

On a final note, I think the key to this project is going to be a large sample size. I think you're going to need the input of several people to be able to reach conclusions about your different conditions. Also, any experiments that involve people require informed consent from the SRC. This requires that each person fill out a form. If you're not sure about this, post back here and I can provide you with a link.

Best of luck, and I look forward to hearing from you.

Linh

Hi Cate,

Have you read the notes on the science buddies site about this type of project? it has some useful ideas about questionairres and so on.
https://www.sciencebuddies.org/mentorin ... ?from=Home

Are you totally set on line-ups or would photo arrays/ mug shots also be a possibility?

I have an idea that might be a way to get around your difficulty with having enough people to be in the film and still people left over to be in the experiment. You could use an extract from a movie with a crime in it. Pick a film that is little known, with actors that are not easily recognisible. You could maybe even use a foreign film as the words wouldnt be important if the action is obvious. (Remember don't pick anything that is inappropriate for the age of your test subjects, I would use a robbery, nothing violent) Then download picture of actors from the internet to make the mugshots. You could try and find one of the actor who commits the crime looking really different like different hair colour or with a beard or something.
Just something to think about...

Keep posting your ideas.
Good luck,

-Caroline

Thank you guys for all the help! The four lineups I was going to use were with everyone lined up with the crime scene in the backround, everyone lined up with no backround, mug shots of everyone with the backround, and mug shots without the backround. I think that it may be hard to make the experiment accurate if everyone knew the others well, but I'm still quite not sure about that affecting the results. It would definetely be more fun that way, but then again, it might not work.

I kind of like the idea of getting a clip from a movie, but that might take some fun out of the project. I like to make movies, and do it all the time, so I know it would be easy to make the video really cool, but there is still the actor issue. If I was to film it, how do you think it should be done? Should I make it really obvious who did the crime, or just sort of a glimpse? Thanks again for helping...

~Cate~

oh, and one more thing, who should the pictures be of? several people who look similar, everyone in the movie scene, or something else?

Hi Cate,

I think the questions you raised bring up that there are slightly different aspects of a line up that you could test in your experiment. I would say its probably best to concentrate on one type of test at least at first.

The first type of variable you mention is about whether it is clear who did the crime or not.

Another variable you mention is about who else to use in the lineups as likely false positives (not the criminal but someone people might often pick or positively ID falsly)
One of the articles linked from the science buddies project description talks about an experiment where a college student performed a synthetic crime and when they analyzed peoples lineups responses they found that those who identified the wrong person often chose a bystander who was at the crime scene.

What do you think are the benefits and downsides or difficulties to testing these variables?

-Caroline

Well, I think that I should at least test 2 of the lineups, maybe just 1 with mug shots and another with a standard lineup. That way I wouldn't have to test 4 different groups. When taping maybe it would be best to just
tape the crime like you were watching it. Also, maybe add some dialogue from before and after the crime to make it not so obvious. We still have the issue of who would be acting, etc. An option, though, would be to disguise the people to look different from their normal selves. As I mentioned before, I do movies all the time, so I could dig up quite a few things for disguises. Also, I edit videos on a program that lets you add effects, etc., so I could spice it up with that maybe. At least do some things that make the crime harder to analyze...

Hi Cate,

Thats great that you have experience with videotaping and editing and will help make it a good tape and fun to do the experiment (always a good thing).

What I am trying to get at is what kind of question are you trying to test with your experiment.
Are you trying to test how good people memories are over different time lengths after a crime happens?
or are you trying to test witnesses abilities to recognise a face?
are you trying to test whether people will make an ID even if the criminal is not in the lineup?

Do you see what i mean, there are lots of ways that this project would go, if you narrow down the question then it will help you with decideing the details of how to carry out the test.

-Caroline

I was more going at: What criminal lineup works better to trigger a reaction that would help more people recognize the real criminal. If mug shots were better or a lineup of all the suspects. I never really thought of different ways this could be approached. What do you think would be a better question in my situation? (Thanks for all the help, I probably couldn't find it anywhere else.)
~Cate~

Hi Cate,

You are welcome - thats what we are here for

I think that the type of question to choose to study is going to depend on the type and number of people that are in your test group and the people in your movie.

If the people in the movie were people not well known to the test group, or perhaps if some were known and some not familiar, then you could test the question about whether the in person line up works better than the mugshots. And you might even be able to test how it works differently if the criminal is a familiar face or an unknown.

But you mentioned that everyone in your school knows each other quite well. So if both the test group and the people in the movie were from your school then i think that the test subjects ability to recognize someone would be compromised, because they would know all of the "suspects".
In that case I would think about testing a different type of question - perhaps more about memory of the sequence of events and how that memory could be affected by time or suggestion.

-Caroline

You're right about everyone knowing each other. Maybe it would make more sense to make the question a matter of time between the crime and survey, but then the experiment could turn into "who has the best memory". I think that it might be a good idea to make a movie with familiar people, and then take a clip from an old movie-like you suggested- so the actors wouldn't be familiar. Then I could do the experiment twice, but with one familiar, and one non-familiar movie. Could that still be considered a science experiment?

Hi Cate,

If you can ask a scientific question and set out to test the answer in a scientific way then its a science experiment.

Its doesn't have to be about who has the best memory but could be about what details about the crime do the majority of people get right or wrong. For example, are people more apt to remember visual details - what was the colour of the hat the terrorist wore?, what shape was the suspicous package? what order were the 3 criminals standing in as they approached the bank teller?
or verbal details - what did the male criminal say to the bank teller? what was the suspect saying when he got off the bus?
Are males better at remembering visual details than females? those sorts of questions.

I like your idea of trying to compare recogition of familiar vs unfamiliar. So for example in your questionairre you could ask whether people can assign correctly which of the 3 robbers said "Stick 'em up", which one wore the red hat, etc and see if the group scores better on the details when they know the suspect or when the suspect is an unfamiliar face.

-Caroline