Re: Can't Anyone Here Play This Game?

1

This isn't a responsive comment, but...

Who are the schmucks used as "fillers" in a lineup? Is it Bob from shipping, Lt. Enrique from the narc division, and Jim the temp secretary? The first 10 guys in the drunk holding pen?

horizontal rule
2

You are right that it might have made more sense to make both double-blind so that they could isolate and evaluate just one effect.

It is not surprising to me that you would have greater error in the sequential identification--just as it is not surprising that there are errors in the traditional method.

In both situations, the witness is likely to feel pressure to pick someone--even if told not to pick unless they were sure. It's just human nature. At least in the simultaneous lineup, they are more likely to pick the person who looks most like the actual suspect (good if the suspect is in the lineup, bad if he or she is not). In the sequential lineup, I can see that there would be a tendency to pick the first person the witness came to who looked close.

horizontal rule
3

Who's That Guy in the Police Lineup?

horizontal rule
4

I thought I heard of a proposal to just show people one at a time. This him? This him? This him? Etc. No lineup at all.

horizontal rule
5

Thanks, eb!

horizontal rule
6

I have come up with a new lineup method that is 66% more effective against terrorists than the conventional method.

horizontal rule
7

4. That's the sequential lineup, unless I'm not understanding you.

2. Well, the study shows that people using the sequential lineup are less likely to pick anyone at all, so it appears that if the pressure to pick someone is a factor, it at least isn't more of a factor in the sequential lineup.

horizontal rule
8

At least in the simultaneous lineup, they are more likely to pick the person who looks most like the actual suspect (good if the suspect is in the lineup, bad if he or she is not).

I know this is a typo rather than a real mistake, but I can't resist. You mean 'good if the criminal is in the lineup.' We know the suspect is always going to be there -- the problem is that the criminal might not be.

horizontal rule
9

And 6. Don't laugh, people used to use your terrorist resistant lineup -- it's called a 'show up', where the cops show you one guy and ask "Is this him." Gets a lot of false IDs.

horizontal rule
10

I can see how the pressure to pick someone might be a factor. Presumably most witnesses have seen as much or more police procedurals as I have, and I *know* that the real guilty party is always included in the lineup, that there's always DNA evidence, that the eyewitness is just the icing on a cake of solid police-work...

horizontal rule
11

They should make people submit to multiple lineups, and have the suspect be in only one of them.

horizontal rule
12

You mean 'good if the criminal is in the lineup.' We know the suspect is always going to be there -- the problem is that the criminal might not be.

Of course, what you said.

horizontal rule
13

So LB, are you gonna write a letter to the Times saying what you just said (although more succinctly, of course)? 'Cause you oughta. Hell, everyone on this thread ought should.

horizontal rule
14

Thinking about it more, I can see circumstances where what the researchers did was not unreasonable. Imagine you are a researcher. You have grant money for one medium-sized study (making up the facts here, but these are not unreasonable facts to make up). The conventional wisdom is that sequential lineups are better than the standard ones and that double-blind lineups are better than the current method. All that is needed is one good, rigorous study to confirm this. So, you design a study comparing the status quo to what everyone "knows" is the better method--double-blind sequential lineups. If it results come out the way everyone assumes they will, you have acheived a great result.

On the other hand, to prove both that double-blind lineups are better than non-blind ones and that sequential lineups are better than simultaneous ones, you need two studies, or at lease one much bigger and more complex one, where you can test all of the variables. In retrospect, they obviously should have found the time and the money to take a more rigorous approach, but I can imagine how this happened.

horizontal rule
15

Except that there's a real problem with figuring out what a 'better' result is. The 'best' result is that the witness identifies the suspect only when the witness is actually certain of the identification without prompting; but the reporting of this study suggests that fewer identifications of the suspect are being considered a downside, rather than an upside, of the sequential lineup. From the reporting, at least, the study really doesn't appear to be well thought out.

horizontal rule
16

I am polyanna on Saturday mornings, and I wonder, uncharacteristically, if it's possible to imagine an environment where eyewitness testimony is not overvalued, and most people are no more swayed by it than they should be. And whether it's possible to imagine an environment where what an overwhelming majority of the public wants of the police is that they make really honest efforts to develop and follow the evidence. So that the pressures now existing to look effective, and the dark reasons you sometimes hear, that "society" cares about the show of force and action, to create the feeling we have control and are doing something, is as muted and suppressed as leadership and shared values can make it.

I have never been cynical about the police per se. Majorities under all systems not hopelessly corrupt try to do this now, yet we have the conviction that these abuses happen. Perhaps the officers in question convince themselves that the system is an impediment to the use of their best judgment, and that they are doing their best to get the right results.

Would that "we" held them, and therefore they held themselves, to a standard where they had no interest at all, acknowledged or unacknowledged, in false IDs.

horizontal rule
17

As a witness, I would imagine that I would feel a growing anxiety to pick "somebody" as I saw more and more potential suspects pass by.

16. Police are also subject to quotas, performance reports, and pressure to produce cases that conform to the department goals (never forget police departments are beurocratic organizations who spent a good part of their time justifying their budgets). And this is all outside of what's considered "coruption".

horizontal rule
18

I disagree; letting those things determine results is the essence of what I mean by corruption.

horizontal rule
19


In a journalism course once, we watched a video made at Northwestern University (or similar), where the prof was giving a lecture on eyewitness accounts when a man swooped in the middle of class and stole the lecturer's briefcase. 'Gee,' said the prof, 'who would have thought that would happen? did anyone get a good look at him? Brown hair? I think he had a beard? '

The students were then given a book of faces and told to pick out the suspect. I don't recall the details, but almost none of them -- trained investigative journalist-wannabes -- identified the correct man. Most of them picked a face that had facial hair, was white, and had brownish hair..

Of course, the 'criminal' (probably a grad student) was clean-shaven. Moral: eyewitnesses are t3h unr3liabl3 and t3h influ3nceabl3.

horizontal rule
20

That's pretty depressing, LB. I can't believe these people take themselves seriously :(

horizontal rule
21

the 'criminal' (probably a grad student)

Profiler!

horizontal rule
22

Another questionable lineup technique.

"a multiple-choice test without any wrong answers"

horizontal rule
23

That strikes me as a different issue (for anyone who didn't bother to click through, it's an article about the Duke lacrosse team incident. The witness was given a photo book of the lacrosse team to identify her assailants from, with no 'known to be innocent' fillers.)

The general problem with stranger identifications is that we don't know if the witness has ever seen the suspect before -- the suspect may be someone who has never been in the witness's presence, and the risk is that the witness may err and think they recognize someone they've never seen. In the Duke case, we know that the team-members were at the party, and the witness saw them -- while her veracity is still in question, there isn't any significant risk of an innocent mistake as to her identification of the men she alleges to have assailed her at the party. (I suppose she could be confused between two similar looking team members, but I don't see how adding fillers would reduce that possibility.)

horizontal rule