Dear Editor,
The police line-up for a parade is both the critical means of presenting eyewitness identification in court and one of the most dangerous tools of justice. The pernicious combination has resulted in its continued use despite the recognized dangers to innocent suspects in being mistakenly chosen in the line-up. A British commission headed by Lord Devlin in 1976, appointed to investigate two miscarriages of justice recommended that the line-up not be the sole source of evidence against an accused.
A famous case was the trial of John Demjanjuk in Israel. Demjanjuk was accused of being a sadistic guard nicknamed ‘Ivan the Terrible’ in Treblinka extermination camp. The judges were certain that Demnajuk had been a guard, but the only verifiable evidence identifying him as Ivan the Terrible was the eyewitness testimony of several survivors the camp. He was found guilty by the court and sentenced to death, but escaped execution on appeal because there were questions about the validity of his identification cards issued to him by the SS. Somehow despite their knowledge of numerous misidentifications, three of Israel’s most respected judges condemned a defendant.
It is small wonder, then that jurors often err. Somehow, jurors and judges alike give undue weight to line-up identifications, convicting with little supporting evidence, or even in face of conflicting evidence as in the cases reviewed by the Devlin commission and the Demjanjuk case. A less discussed but also serious problem is the failure to identify a guilty suspect from a line-up.
Judges know and jurors can be told, that witnesses can mistakenly choose an innocent suspect in a line-up, yet both will often believe the witness. Brandon and Davies (1973) listed 70 such cases of mistaken identification and pointed out that they were merely the “tip of a much larger iceberg.” After all such cases come to light only by the chance appearance of new evidence. The police do not usually search for evidence on a crime for which a suspect already has been convicted. Experimental evidence confirms a tendency to believe witnesses, regardless of their accuracy. This is because identification requires a comparison between an image of the offender and a suspect who even if he is the offender, cannot look absolutely identical to the person the witness saw. He likely will be dressed differently, have a different expression on his face, have more or less facial hair, and so on. The memory image of the face may itself change, so small wonder that we err at times in identifying even familiar people. Witnesses have to compare an image in their heads to each person in the line-up and decide whether the similarity is great enough for concluding, identity. A number of methods are available to influence the witness against choosing too easily, such as giving unbiased instructions, and presenting the line-up members sequentially.
However, given the uncertain effect of our efforts on any particular witness, we should be wary of pushing the witness to adopt a criterion that is too strict. Such a criterion will cause the witness to fail to choose the offender. We are faced with the problem of balancing the need to protect the innocent suspect with the need to convict the guilty one.
In sum witnesses do not indentify suspects in line-up. They choose the suspect. As Navon (1990) pointed out, “An identification of the suspect merely indicates that the suspect resembles the perpetrator.” Surely the probability that the suspect is the perpetrator is greater if the suspect is chosen than if he is not, but we have little idea how much greater.
Yours faithfully,
Robert Gates
