More than 20 years ago, acclaimed novelist Alice Sebold wrote a memoir that chronicled in chilling detail how she, a white woman, was raped by a Black man while in college in the 1980s. Recently, the person she identified as her rapist, Anthony Broadwater, was declared innocent. Daniel Medwed, GBH News legal analyst and Northeastern University law professor, joined host Aaron Schachter on Morning Edition today to discuss the phenomenon of eyewitness misidentification. This transcript has been edited lightly for clarity and length.
Aaron Schachter: This case has so many components to it. There is no doubt that a rape occurred, but it seems clear that Alice Sebold identified the wrong man. I know your research focuses on wrongful convictions. How often does this kind of thing happen and why?
Daniel Medwed: Well, sadly, Aaron, it happens quite often. There have been approximately 307 documented DNA exonerations in this country since 1989. Those are cases where someone's convicted and while they're in prison, or later, post-conviction DNA testing of the biological evidence retained from the crime scene clears them, proves they're not the guy. A deep dive into those cases reveals that eyewitness misidentification appears in about 70 plus percent of them. Oftentimes, multiple eyewitnesses get it wrong in the same case, and similar studies of other non-DNA exonerations buttress these findings.
So the question is, why? Scholars tend to identify two major factors — what they call estimator variables and system variables. Those are fancy academic phrases. I'll break them down. First: By their nature, crimes happen typically unexpectedly and often under poor visibility at night, and then it's magnified by the stress of the event for the victim. That creates a perfect storm for a failure to estimate the features of the perpetrator to get it wrong initially. And then over time, your memory shifts and evolves as you experience post-event data points as you see other people. Memory isn't static, it's elastic. It shapes over time. And so what happened in this case, as in many cases, is there was a delayed identification.
Alice Sebold didn't identify Anthony Broadwater until five months after the crime, when she saw someone on the streets of Syracuse who in her mind resembled the perpetrator. So that's the first problem, the problem with estimation. The second problem is how the system generates official eyewitness identification evidence, how the police interact with eyewitnesses and develop this evidence.
Schachter: So they're sort of pointing a victim to in a certain direction, is what you're saying.
Medwed: Yeah, in a sense. The way they do it is after they start investigating a case and they find a prime suspect, they put that prime suspect either in a physical lineup with five fillers or in a photo array with five other photos. And what they do is they present it as a group. It's often called a “six pack” to the eyewitness. And what happens then is the witness engages in a comparative or relative judgment — which of the six most closely resembles their recollection of the perpetrator — as opposed to what's called an absolute judgment. Did number one do it? Yes or no? Did number two do it? Yes or no?
It's a form of comparison shopping, and the witness is probably going to pick someone regardless of whether it's the right person. In addition, this is aggravated by the police administrator who often oversees these procedures because that person typically knows who the prime suspect is. And as you alluded to, Aaron, may subtly or not so subtly steer the eyewitness in the direction of an ID.
"Memory isn't static, it's elastic. It shapes over time."-Daniel Medwed
Schachter: Daniel, are the risks of misidentification higher in a case like this where the victim and perpetrator belong to different races?
Medwed: Absolutely. And those are called cross-racial misidentifications, and the annals of wrongful convictions are full. They're replete of instances where someone of one race misidentified someone of another race. There are lots of explanations for this. One, of course, is that we still live in a very segregated society, and we're just not very good at identifying people from different races.
Another explanation is called ethnocentric heterogeneity, and it holds that people within a racial group use very specific identifying characteristics to distinguish people within the group, and those characteristics don't always translate well to the interracial identification context. And lastly, of course, at least in some cases, bigotry or intolerance or at least indifference can play a role.
Schachter: OK, next time we speak. Daniel, you're going to tell us how to solve all this.
Medwed: You can help me with that, please.
