In college I spent a semester as a TA for an introductory course. My duties were to assist a student who was blind, taking notes for him on the visual aids shown during lecture, meeting with him to review textbook readings, and so on. I also administered his exams, asking him the multiple-choice questions aloud and then recording his responses.
I spent a lot of time with this student, and I grew to like the guy. So I'll admit that it pained me when, on occasion, he clearly hadn't studied enough for the test I was reading him. It was tempting to throw him a bone once in a while. No, nothing obvious like circling the right response instead of the wrong one he had chosen. But in the midst of the awkward silences that would fill the air as he pondered which guess to go with, I'd often mentally catalog the subtle ways I could conceivably guide his answers.
• Perhaps changing my tone of voice, emphasizing one answer choice as opposed to the other options on the table.
• Or if he was pondering two possible answers through a statement like, "well, it could be a), but..." I could jump right in if a) was the right answer, record that response, and suggest that we move on to the next question...
• ...but when was a) was wrong, I could keep quiet and give him more time to deliberate further.
Recently, I started thinking about these test sessions as I've begun serving as an expert witness in criminal trials on eyewitness memory. What's the connection? Well, as I alluded to in a previous post, the experience of being an eyewitness is a lot like taking a multiple-choice test.
Time and time again, research studies of eyewitnesses (for one bibliography, click here) indicate that remembering and then accurately picking out a face you've seen previously is not nearly as easy as we think it is. Memory doesn't work like a videotape, so eyewitnesses can't approach a lineup or photo array the same way a CSI analyst might run a computer match on a forensic sample.
In other words, witnesses don't have a perfectly stored memory trace that they're able to map on to faces they see later. Instead, the eyewitness in front of a lineup or photo array is a lot like the student taking a multiple-choice test.
The question that witnesses are most equipped to answer isn't Which of these guys is the one you saw at the scene?
It's Which of these guys looks most like the one you saw at scene, relative to all these other guys?
This means that when it comes to eyewitness identifications, there are influential factors that we often overlook–just as with multiple-choice tests. Such as the other answer choices available. The instructions given. And who it is that's administering the test to begin with.
The other answer choices make a difference in lineups precisely because eyewitnesses rely on the process of elimination. For a lineup or photo array to be fair, all of the people in it should be reasonably good matches to the eyewitness' description of the culprit. Otherwise, the suspect may stand out simply because he's the only guy who fits that description.
This seems like an obvious conclusion, but try telling that to some of the defendants who have asked me (or other experts) to testify on their behalf. As just one example, last summer I worked on a case in which the victim described his assailant as Black, 5'1", 140 pounds, and in his early 30's. The photo array police used included seven faces. Three were of adolescents that couldn't have been out of their teens. Two were of men large enough to have been football linemen. That left only two photos that even remotely fit the description the witness had given, one of which belonged to the defendant. While there were seven faces in this lineup, in terms of functional size, there were really only two faces to choose from.
The instructions given in a lineup or photo array also make a difference. When police bring a witness to the station to show them photos, there's an implied message in that invitation: We found someone that we think is the guy. Otherwise, why would they take the time and effort to bring the witness in? This implied message only reinforces witness tendencies to use the process of elimination–to pick the person who looks most like what they remember, relative to the other options available.
To remedy this issue, it becomes important for police to instruct an eyewitness that the suspect may or may not be in the lineup. This frees the witness from feeling obligated to pick someone out, lest she wind up wasting everybody's time. Indeed, this is one of the specific recommendations made by the Department of Justice's research-based guide for handling eyewitness evidence, published in 1999.
And, of course, the person administering the lineup can make a big difference as well. Much like the TA giving an oral exam, the lineup administrator has the power to shape a witness' responses through actions both overt and subtle, both intentional and inadvertent.
Consider a study at Iowa State University by Gary Wells and Amy Bradfield Douglass. Research participants were shown a security camera video and then asked to pick the "culprit" out of a photo array. Those participants whose ID was followed by the simple statement, "Good, you identified the suspect," reported being markedly more confident in their selection than did participants who didn't receive such a response. Not only that, but the reaction to their identification also influenced participants' recollections of how good a look they had gotten at the culprit, how long the video had been, etc.
Thinking back to my TA days, it's easy to see how lineup administrators can shape eyewitness identification, even when they aren't trying to. How to solve this problem? Have the officer who assembles the lineup–who knows which person the suspect is–be different than the officer who actually administers it to the witness.
Or just do what my department did when I was in college. To make sure all was kosher with the test I was giving, they set up a tape recorder to document everything that went on in that closed room. This was enough to remind me to be on my best behavior, to make sure that I administered that exam in as fair and neutral a manner as possible. It doesn't seem like that much to ask for the same type of safeguard to be put in place for lineups and photo arrays. After all, those are the real examples of high-stakes testing.
