Reid detractors also say that police often feed evidence to suspects, which accounts for why false confessors sometimes know details about a crime that they wouldn’t otherwise know. In a recently published study, University of Virginia law professor Brandon Garrett found that in 97 percent of the false-confession cases he studied from the DNA era, the wrongly accused suspects were said to have supplied such telling details—facts either picked up elsewhere or provided by police. Interrogators also tend to be overconfident of their abilities to spot guilty suspects. No study so far (aside from Reid’s own research) has shown the police to be any better than average at picking out liars. In fact, they’re sometimes worse. In one 1987 study, police officers watched videotaped statements of witnesses, and their record at identifying deceptive testimony was no better than the average person’s. Overconfidence can blind investigators to evidence suggesting that the suspect is innocent. The pressure to resolve cases quickly and tidily can have a similar effect, especially in high-profile cases. Simply wearing suspects down is another issue: At some point, a given suspect will say anything just to make the immediate discomfort stop. “Why don’t they beat people anymore?” asks Don Thompson. “It’s not because they’re particularly enlightened now. It’s because the psychological coercion is so much more effective.”
Frank Sterling’s confession, Thompson believes, was marked by a number of these problems. After Sterling says he hit Vi Manville, Patrick Crough asks Sterling what he hit her with. Sterling says, “My hand.” A moment later, Crough says, “Frank, as best as you can remember, and I know this is difficult for you, did something happen with that BB gun?” Only after being prompted that way does Sterling say, “Yeah, I started hitting her with it.” Mark Sennett, the polygraph examiner, lied to Sterling about his brother Glen. Crough teased out the motive and alternately pressured and consoled Sterling. Sterling knew about the supposedly telling crime-scene detail of Manville’s purple jacket, but Thompson says many people in Hilton would have seen her on her daily walks in that jacket. Finally, there was Sterling’s state of mind. Having been held alone, without counsel, in a small interrogation room and questioned for twelve hours, he became isolated, exhausted, and vulnerable to manipulation. Over the years, Thompson and Crough had crossed paths in Rochester, running into each other around town or at the supermarket. “I’m never really comfortable when I’m talking to him,” Thompson says. “He’s an accomplished interrogator, which translates to being an accomplished manipulator.”
Shortly after Sterling’s release, I had dinner with Crough in Rochester. Calm and self-assured, he did what he could to sound gracious about Sterling’s ordeal. But he couldn’t help but also be defensive. He insisted he did good work that night in 1991. “His responses kept the interview going,” he told me. “As a homicide detective, you don’t walk out on an interview when the person’s giving you a little something.” Crough pointed out that he was the one who visited Christie in prison—he volunteered to do it and talked him into giving his DNA sample when Christie didn’t have to do that—and it was his book that helped persuade Christie to confess. After a while, though, some contrition bled through. “Like that hasn’t haunted me?” he told me. “I’ve been doing interrogations in major crimes for twenty years. This is the first time I’ve ever had one go bad on me. That’s not a bad statistic, you know.”
The law-enforcement community insists current interrogation techniques are sound. The courts have upheld tactics like deceit and minimization, Reid president Joseph Buckley notes, and without such methods, police would have a far more difficult time eliciting confessions from suspects who are, in fact, guilty. When false confessions do happen, Byington says it’s not the Reid technique that’s to blame but the misapplication of it. The police’s main mistake with Frank Sterling, he says, was starting in on their suspect before they were reasonably sure he was guilty. Then, when Sterling gave Crough and the others questionable information, they blindly barreled ahead. “When they ask Frank what he was wearing, he says he thinks he was wearing a T-shirt and jeans,” Byington says. “Well, if Frank was wearing a T-shirt and jeans, he’d have frozen to death.” In Byington’s opinion, Sterling had essentially been fed information over twelve long hours, then encouraged to spout it back over twenty minutes of video. In a good confession, Byington says, the suspect should do about 80 percent of the talking, narrating their experience for the benefit of the police, not saying yes and no to a series of prompts.