08/30/2011 05:21 pm ET Updated Dec 06, 2017

Why Eyewitness Memory Isn't All It's Cracked Up to Be (Part I)

Few moments in a courtroom are as compelling to a jury as seeing a confident witness point out a defendant and say, "that's the man I saw at the crime scene." The persuasiveness of such identifications is what makes the fallibility of eyewitness memory so surprising to most people when they learn about it.

In the past week, the New York Times has run a series of pieces on eyewitness memory and how police departments and the courts are wrestling with the best way to handle such information (here and here, for example). With this topic returning to the forefront of debate, I thought I'd share the story of my own participation as an expert in a case that hinged on questionable eyewitness identification. As the Times articles spell out, cases like this one go to trial everyday in hundreds of courtrooms across the country. So here goes:

Not long ago, I got to change up the ole' routine for a day. My regular schedule called for a morning lecture, an afternoon seminar, and some writing in my office sandwiched in between. Instead, I woke up and dressed like a grown-up for once, got in the car, and drove in the opposite direction of my campus to a Superior Courthouse in another part of the state to testify as an expert witness.

On paper, the case seemed fairly unremarkable. An armed robber had approached two men on a city sidewalk, brandishing a gun and making off with some cash and an iPod. Understandably panicked, the victims called 911 and proceeded to give the operator a detailed description of the red sports car into which they saw the culprit climb and then flee the scene.

Less detail (and time) on the 911 tape was devoted to describing the mugger himself. But the caller was able to describe his assailant as a Black man in his 20's wearing a hoodie, with dark complexion and smooth skin, 6'5" tall, and with multiple jagged or missing teeth.

Two weeks later, police obtained a security camera image of a young Black man in a hoodie from a convenience store located just a few blocks from the crime. The image had been taken less than an hour before the robbery. The man depicted didn't have much of an alibi for the night in question; in fact, he was anything but cooperative when interviewed by police. And when the two victims were (independently) shown an array of 9 photos and asked to identify the man who had robbed them, they each picked out the same guy -- the very guy from the convenience store security camera footage.

Open and shut case, right?

Well, much like witness memory itself, it turns out that this case had more to it than meets than eye. The crime narrative weaved by the police and the state began to fray at the edges the more closely you examined it.

First of all, the 18-year-old charged with the crime didn't really match the description the victims had originally given. The defendant didn't have a car of any type, much less a red sports car. He had obvious facial hair. He wasn't 6'5", but rather barely 5'10". And his teeth? As the defense attorney eloquently detailed in the first letter he sent to me regarding the case, his client "had teeth that would make a dental hygienist smile."

What about the identifications that each victim made? Well, despite the fact that they were now testifying at trial that they were 80-90% sure that they had picked the right photo, the actual identifications had been far more tentative. One victim spent a few minutes looking over the 9 images before choosing the defendant and estimating his confidence at 60%. The other victim was only able to narrow the array down to two photos, estimating a 50.1% chance that the defendant had been the mugger and a 49.9% chance for another guy.

So the defense's case rested on the premise that these victims -- while well-intentioned -- were mistaken in their identifications. Indeed, this crime included several of the specific risk factors that an extensive body of research has identified as rendering eyewitness memory less reliable:

  • The victims had limited exposure to their assailant, only seeing him briefly and under suboptimal lighting conditions.
  • The culprit's hood had covered his hairline and eyes, obscuring physical features that play a major role in accurate face perception and memory.
  • In addition to its brevity, the crime was an extremely stressful event, as demonstrated by the (understandable) difficulty that the 911 caller had catching his breath during the call.
  • The mugger flashed a gun, and research on the weapons focus effect has indicated that guns often draw a witness's attention away from the culprit (and the culprit's face): in this case the victims had given detailed descriptions of the gun to police, while admitting to having spent most of the rest of the robbery staring down at the sidewalk in fear.
  • The victims were White and the culprit was Black. Now, by no means are accurate cross-race eyewitness identifications impossible. But statistical review of dozens of published studies indicates that we're more than twice as likely to make identification errors when trying to differentiate between faces of racial groups other than our own.

These are the types of issues about which I was asked to testify as an expert. My task would be to explain to the court the basic findings of hundreds of empirical studies of eyewitness memory that psychologists have conducted over the years (for some examples, click here). To talk about the scientific consensus regarding the effects of stress on memory, cross-race facial recognition, the malleability of eyewitness confidence, and so on. And to trade in my usual afternoon of answering curious questions in front of a classroom of eager students for an hour of aggressive cross-examination by a skeptical district attorney.


Like this post? Then check out the website for "Situations Matter: Understanding How Context Transforms Your World" (available now for pre-order). You can also follow Sam on Facebook here and on Twitter here. Book trailer video below: