Cross-Race Identification (Dan Reisberg, ch. 3)

In Chapter 3 of the textbook, we argued that people use two distinctly different mechanisms for recognizing the stimuli they encounter. One mechanism is feature based, and works by first identifying the input's parts. These parts are then assembled into larger and larger wholes until the entire input is identified. The other recognition mechanism, in contrast, is configuration based. It is less concerned with individual features, but is exquisitely sensitive to the overall arrangement in the input.

The second of these mechanisms, relying on configurations, is crucial when we are recognizing faces, but we also use it in other settings: Expert bird watchers, for example, seem to use this mechanism when they are making distinctions among different types of birds; expert dog-show judges rely on the same mechanism to recognize individual dogs. Overall, it seems that people use the configuration-based mechanism whenever they are identifying individuals within an enormously familiar category. How is this pertinent to law enforcement? Imagine that you witness a crime. From the information you provide, the police develop a hypothesis about who the perpetrator might have been. They place the suspect's photo on a page together with five other photos and show you this "photospread." Will you recognize the perpetrator within this group? If so, this provides important evidence confirming the police suspicions.

In this situation, your ability to identify the perpetrator depends on many factors, including the suspect's race. This is because people are much better at recognizing individuals from their own race than they are in recognizing individuals from other races. Indeed, if the criminal is present in the photospread, witnesses are roughly 50% more likely to miss him if the identification is across-races (e.g., a White person identifying an Asian person, or an Asian person identifying a Black person) than if it is within a race. Likewise, if the criminal is not present in the photospread, the risk of the witness falsely identifying someone who's actually innocent is roughly 50% higher in cross-race identification.

Why should this be? We have emphasized that the recognition system people use is configuration-based only when they are making an identification within an enormously familiar category - and the fact is that most people are extremely familiar with faces from their own race, less so with faces from other races. As a result, people can rely on the configuration-based mechanism when making same-race identifications, and so benefit from this system's sensitivity and sophistication. When making cross-race identifications, however, people are less sensitive to the face's configuration. They therefore have to base their identification on the face's features, and it turns out that this is an appreciably less effective means of identifying individual faces. As a result, cross-race identifications end up being less accurate. Apparently, then, courts need to be more cautious in interpreting cross-race identifications. But does every witness show this pattern? Or perhaps, is the disadvantage with cross-race faces smaller for people who live in a racially integrated environment? Questions like these continue to be the subject of research. In the meantime, it's already clear that this issue may help the courts in deciding when to put full trust in a witness's identification, and when to be wary of an identification's accuracy.

To learn more about this topic in cognitive psychology and the law:

See, in the textbook chapter, pages 94-97

*Pezdek, K., Blandon-Gitlin, I., & Moore, C. (2003). Children's face recognition memory: More evidence for the cross-race effect. Journal of Applied Psychology, 88, 760-763.

*Wright, D., & Stroud, J. (2002). Age differences in lineup identification accuracy: People are better with their own age. Law & Human Behavior, 26, 641-654.

Meissner, C. A., & Brigham, J. C. (2001). Thirty years of investigating the own-race bias in memory for faces: A meta-analytic review. Psychology, Public Policy & Law, 7, 3-35.