POW Memory

POW Memory

Most people think eyewitness testimony is the best possible evidence against an alleged criminal—especially when that testimony comes from the victim. But people who survive terrifying situations may actually have surprisingly unreliable memories of who or what caused them.


How stress can undermine memory. I'm Bob Hirshon and this is Science Update.

You might think you could never forget the face of someone who tortured or assaulted you. But according to Yale University psychiatrist Andy Morgan, don't be too sure.

He and his colleagues studied more than five hundred soldiers at Fort Bragg, North Carolina, while they were being trained for prisoner-of-war situations. After half an hour of frightening interrogations, many of the soldiers couldn't pick their tormentors out of a lineup, a photo spread, or a series of photos.


Even under the best circumstances, for a very highly stressful event, we found that at least one in three people is wrong. And if you don't have a picture from the scene of the crime, it's probably no better than a coin toss.

In fact, some of the soldiers even got the interrogator's gender wrong. Yet they were sure they were right.


You have to remember these are special operations personnel and maybe pilots and Marines, so not real timid, anxious people. They were very confident in their responses to the test, but their confidence had no relationship whatsoever to accuracy.

He says soldiers in low-stress interrogations did better, but were still wrong about a quarter of the time. Morgan's team plans to further study the phenomenon and the brain chemistry behind it. I'm Bob Hirshon for AAAS, the Science Society.

Making Sense of the Research

Most people believe that there's nothing better than eyewitness testimony. In most criminal trials, eyewitnesses have a strong impact on a jury, unless the defense can convince them that the witness is unreliable or just plain lying.

But the fact is, eyewitness testimony isn't actually all that reliable. Many studies have shown that people who watch videos of a staged crime have a hard time remembering who did it or what took place, even though they're often sure they remember it perfectly. Often, people unconsciously fill in missing facts, and remember them just as clearly as the events they actually saw. Others confuse the people, locations, or other details of the event in question.

Despite this evidence, some scientists have argued that your memory would be clearer if you were the victim of a crime, rather than a witness. There's some indirect evidence for this: animals injected with the "fight-or-flight" hormone adrenaline while they're learning how to do something tend to remember it better later on. And humans who suffer from post-traumatic stress disorder (for example, crime victims or soldiers who fought in wars) can't stop thinking about a terrifying experience, no matter how hard they try. If the memory is that strong, wouldn't it at least be accurate?

Dr. Morgan's team tested this by studying soldiers in prisoner-of-war training simulations. These soldiers were subjected to interrogation techniques that are already known to increase the body's stress response. The results were striking: Despite being alone in a room with the interrogator who tormented them for over half an hour, the soldiers were not very good at picking the interrogator out of a live lineup (26 percent accuracy), a photo spread (33 percent accuracy), or a sequence of photos (49 percent accuracy). Showing the victim a picture of the interrogator taken in the actual interrogation room boosted the accuracy to just over 60 percent. For non-stressful interrogations, the accuracy rate was about 76 percent.

This study not only showed that the testimony of crime victims and former POW's may be very unreliable, but also confirmed that some identification techniques worked better than others. The "photo sequence" technique, in which the victim is shown a photo of a suspect and asked "Is this the person who attacked you?" before moving on to another photo, worked much better than a traditional police lineup. Past studies have suggested that this technique helps victims judge each suspect individually. In this study, victims who were shown a photo sequence were more likely to correctly identify the interrogator, and more likely to know when the interrogator wasn't in the lineup at all. However, false positives (when the victim fingered the wrong person) were just as common among all the victims.

So what does this mean? Dr. Morgan isn't suggesting that we do away with eyewitness testimony. But he says that it's important that judges and juries understand it in its proper context: that it isn't always reliable, that it can be influenced by the way in which the suspect was identified, and that the confidence of the witness isn't necessarily related to his or her accuracy. This information can be integrated with other forms of evidence to produce a fair verdict.

Now try and answer these questions:

  1. How did this study differ from previous studies of eyewitness testimony?
  2. What did this study suggest about certain kinds of criminal testimony?
  3. What was the most effective method for getting victims to correctly identify their interrogators?
  4. If you were a juror in a trial that hinged on eyewitness testimony, what factors would you take into account in weighing the testimony?

For Educators

The University of Northern Iowa's Eyewitness Identification Laboratory conducts research and promotes guidelines on the use and accuracy of eyewitness memory.

This recent article from the American Psychological Association discusses the role of psychological science in courtroom situations.

The homepage of Gary Wells, Ph.D., a psychologist at the University of Iowa, is a resource for research on eyewitness memory.

Related Resources

Giant Snake
6-12 | Audio
Dung Beetle Navigation
6-12 | Audio
West Nile Weather
6-12 | Audio

Did you find this resource helpful?

Science Update Details

Grades Themes Project 2061 Benchmarks