Paul V. Coates -- Confidential File, May 8, 1959

Confidential File Why Eyewitnesses Aren't Too Reliable Yesterday I detailed three recent local cases of innocent persons being jailed by mistaken eyewitness testimony. I told you that the cases weren't rare ones. Today, I'm going to tell you why. To do it, I have to point the finger at some highly respected groups who don't like people who point. For example, the police. Too often, their power of subtle suggestion is used to influence a witness to make a wrong identification. As an exaggerated example, there's the "Arkansas lineup," which has become kind of a bitter joke among criminologists. It apparently was (and -- from what I read in the papers -- probably still is) common police procedure in the state after which it was named. It works like this: A woman reports that a crime was committed against her by a Negro. She gives police a physical description of the suspect. Police find a man -- any man -- who generally answers the description she gave. Then they stick him -- one Negro -- in a lineup with eight white men, and ask, innocently: "You see anybody there who looks like him, lady?" Any conscientious cop would condemn the "Arkansas lineup." But chances are that he, to a lesser degree, is employing the same tactic. I talked about it at length this week with Marshall Houts, a former judge and FBI man, and one of the nation's best authorities on criminal evidence. "Just  by handing a book full of mugshots to an eyewitness or by letting him view a lineup of suspects, the police are planting the suggestion that the criminal is there," he pointed out. "It's a dangerous procedure but obviously it's necessary," he added. Houts' criticism was of the additional persuasion and influence -- often unintentionally -- used too often by police. For example, showing a witness just one suspect or his picture. Or, as is the procedure in San Francisco (you've probably seen it on the television show "Lineup), having an officer recite the past criminal record of a man in a line up for the witness to hear. "Some victims are extremely susceptible to suggestion," Houts said. In his book, "From Evidence to Proof," Houts cites example after example of persons later found innocent whose convictions resulted from "rigged" or unobjective witness-identification methods. Houts' suggestion: "It should be standard police department practice to tape record lineups, and take pictures of them." Houts also recited to me the dozens of cases where as many as 6, 10, 12 witnesses made identification of a suspect -- and were wrong. "And as a general rule," he said, "the real criminal bears very little physical resemblance to the falsely identified party." The problem today, he told me, is to educate judges, lawyers, police officers and the public that eyewitness identification is the type of evidence most susceptible to error. Some Strong Words Coming Up It's also been my experience," Houts said, "that the more positive an eyewitness is, the less likely it is that he's right. "Today, judges and juries put too much credence in the eyewitness. In most cases, eyewitness identification should be considered little more than an investigative lead for the police. "And if eyewitnesses are used in court, I think the judge should instruct the jury in every instance as to the unreliable nature of the evidence." Those are strong words. But they're coming from a man -- as general counsel of Erle Stanley Gardner's Court of Last Resort -- has seen dozens of the terrible injustices caused by equally strong words of "positive" eyewitnesses who were wrong.