A family is robbed at night. The robber tells them that they will not be harmed if they stay in bed and cooperate. He asks them where the jewelry and other valuables are kept. They follow orders, and all but have an extended conversation with the robber as he returns several times to make additional inquiries. Now in court, they claim to be able to identify the robber by his voice. The question arises as to what weight should be given to their purported identification. The Lindberg kidnapping case: The earliest research on the ability of listeners to identify voices was performed in the wake of the Lindbergh kidnapping case, in which the famous aviator Charles Lindbergh purportedly identified the voice of his son’s kidnapper. Listeners heard a 56 word paragraph read by someone behind a screen. At intervals from one day to five months, they were asked to select the voice they heard from among five voices. Accuracy ranged from 83% at one day to 13% (less than chance) after five months. A major weakness of that prior research is that the listeners were told that the voice heard previously was among the test voices, whereas in real life that fact is usually not known with certainty. Another problem was that sometimes the listeners got clues from each other as to whether or not a given speaker was the one heard previously. More sophisticated research: My colleagues and I have subsequently conducted more sophisticated research on the issue of identification of unfamiliar voices.
Our findings raise issues such as the following:
- How long has it been since the listener heard the voice?
- Can the listener identify particular characteristics of the voice, and are these accurate?
- Is the voice of the suspect representative of a particular category of voices?
- How certain is the listener of the identification?
- Was the listener under stress when she originally heard the voice?
- How familiar is the listener with the dialect spoken by the suspect?
- Was the purported identification made in the context of a valid voice lineup?