Researchers in London recently completed a new study examining the effects of high levels of stress on the ability of eyewitnesses to accurately identify a perpetrator. Drs. Tin Valentine and Jan Mesout of Goldsmiths, University of London, homed in specifically on the distinction between the low levels of stress, more akin to “heightened awareness,” and the more severe “cognitive anxiety” that most individuals experience when faced with an unexpected physical threat.
The new research further confirms earlier findings by social scientists examining the effects of stress levels approximating those experienced in a typical crime scenario: High levels of stress have a “catastrophic effect on the accuracy of eyewitness memory.” The findings were the same with respect to a witness’s ability to accurately describe a perpetrator — the higher the stress level, the more errors were recorded in descriptive details relating to the perpetrator.
The difficulty of testing the effects of high levels of stress approximating a crime scenario have been noted in the literature, given the difficulty of creating that level of anxiety within the ethical guidelines. The researchers in this case found a context in which subjects have essentially volunteered to be subjected to “cognitive anxiety” in advance, by partnering with the London Dungeon — a “horror labyrinth” involving live actors who often startle visitors.
Forty-five minutes after guests experienced a startling encounter with an individual in the dungeon, they were asked to attempt to identify the individual from either a live lineup or a photo array. The researchers measured the anxiety levels of the participants by two metrics — subjective self-reports of their level of anxiety at the time, and a heart rate monitor (and found the former to be a reliable predictor of the latter). For individuals who scored above the median in anxiety level, only 17% accurately identified the individual; for individuals who scored below the median, 75% accurately identified the individual. In short, the researchers found “a strong negative association between state anxiety and the ability to correctly report the appearance of a person encountered under stressful conditions.” More succinctly: “Eyewitness identification was dramatically impaired by high state anxiety.”
The researchers also compared anxiety levels associated with gender, and found that females reliably experienced higher levels of anxiety than males, which suggests that on average, a female’s memory of a perpetrator encountered under high stress can be expected to be less reliable than a male’s under similar circumstances — though both will become increasingly unreliable as the level of stress increases.
Valentine and Mesout also conclude that their research suggests that lab studies of stress have underestimated the effect of stress on eyewitness reliability, and that the closer the approximation to the level of stress experienced in a real crime scenario, the more dramatically the stress impairs the ability of witnesses to identify the perpetrator.
Reference: I don’t have the full cite to the study itself, since it looks like it’s still in pre-publication, but here’s an approximation: Valentine, Tin & Mesout, Jan, Eyewitness Identification under Stress in the London Dungeon, Applied Cognitive Psychol. (forthcoming 2008). I believe it will appear in the next edition, which I think should be Volume 22, Issue 6.