Black Box Thinking

by Matthew Syed

  But Tyson is not the first to have created a fictitious memory. In a study in Scotland, members of the public were adamant that they could remember a nurse removing a skin sample from their little finger. But this never happened. A week earlier these volunteers had been asked by researchers to imagine a nurse removing the sample. But somehow, on recollection, it had morphed into a real event. They were four times as likely to recall it as real compared with those who had not been asked to imagine it.7

  In a different study, volunteers were asked to look at films of car bumpers in which no windows or headlights were broken. Later, they were asked how fast the cars were going when they “smashed” into each other. Suddenly they started reporting memories of glass shattering when no glass had smashed at all. They had reengineered the memory to encompass the new information provided by the word “smashed.”8

  Memory, it turns out, is not as reliable as we think. We do not encode high-definition movies of our experiences and then access them at will. Rather, memory is a system dispersed throughout the brain and is subject to all sorts of biases. Memories are suggestible. We often assemble fragments of entirely different experiences and weave them together into what seems like a coherent whole. With each recollection, we engage in editing.*

  By retrieving, editing, and integrating disparate memories, we have imagined an entirely new event. People with amnesia, however, are unable to do this. They struggle to remember the past, but they also cannot imagine the future.

  In short, the very fact that memory is so malleable may lead us astray when it comes to recollection. But it could also play a crucial role in imagining and anticipating future events.

  We try to make the memory fit with what we now know rather than what we once saw. In the case of Jean Charles de Menezes, for example, who was shot by police in an Underground station in the aftermath of the London terrorist atrocities in 2005, eyewitnesses said that he had been wearing a bulky jacket, had run away from police, and had vaulted a ticket barrier.

  But it turned out that all of this was untrue. Menezes, an innocent passenger, was actually “wearing a light denim shirt or jacket, walked through the barriers having picked up a free newspaper, and only ran when he saw his train arriving.”9 The witnesses had transposed what they had seen with what they had read about the event subsequently in the newspapers.

  With this in mind it will not seem surprising that when the Innocence Project started to investigate the signatures of wrongful convictions, they discovered that mistaken eyewitness identification was a contributing factor in an astonishing 75 percent of cases.10 People were testifying in open court that they had seen people at the scene of a crime who in fact were elsewhere at the time.

  These witnesses were not necessarily lying. They were not making it up. But then neither was Neil Tyson when he talked about Bush’s stars speech. When the witnesses said they remembered seeing the suspect at the scene of the crime, they were telling the truth. They did remember seeing him there, but they didn’t actually see him there. These are two quite different things.

  This is not to say that eyewitness testimony is worthless; quite the reverse. In certain circumstances it is invaluable in order to secure convictions. Rather, it is to say that memories should be coaxed out of witnesses with sensitivity to the biases that might otherwise contaminate the evidence. The tragedy is that the techniques used by police, until recently, had little of this sophistication.

  The practice of “drive-bys,” for example, has been used and abused for decades: this is where an eyewitness is taken by police to see a suspect on the street, or at their place of work. Given that the witness knows that the police have suspicions about the person—why else would they be going there?—the technique is dangerously suggestive.

  And one obvious problem is that once a person has viewed the suspect they are liable to transpose his face onto that of the real criminal. Each time they recall the crime scene, they will become more certain that the suspect was really there. A tentative identification is rapidly transformed into cast-iron certainty. As Donald Thomson, a psychologist in Melbourne, put it: “Two months down the track, they go into the witness box and say they are absolutely sure.”

  Lineups—where a suspect and a number of fillers are placed side by side in a room—are more reliable than drive-bys, but these, too, have been open to abuse. Often they are conducted by an officer who already knows the identity of the suspect, opening up the possibility that he might inadvertently influence the selection with verbal and nonverbal cues. In other cases lineups have been conducted where only one person, the suspect, matches the description.*

  And so it goes on. There were so many error traps in the methods used by police that entire book chapters have been written about them. If miscarriages of justice had been investigated, these latent problems would have been discovered, and could have been addressed. Instead, these procedures were used, with only minor variations, for decades.

  This was not just bad for suspects, but also for the police, prosecutors, and the public. After all, mistaken identifications cause police to ignore other leads. This often allows the real criminal to roam the streets, perpetrating more crimes.

  The Innocence Project has campaigned for a number of reforms. It argues that lineups should always be administered by an officer who doesn’t know the identity of the suspect. It also calls for sequential lineups, where suspects and fillers are shown one at a time rather than simultaneously.

  When these procedures have been tested, they have significantly reduced mistaken identifications without compromising accurate identifications. A field study in 2011, for example, found that “double-blind sequential line-ups as administered by police departments across the country resulted in the same number of suspect identifications but fewer known-innocent filler identifications than double blind simultaneous line-ups.”11

  Some have disputed these findings and have proposed more tests. But this, in itself, represents progress. Systems are being trialed. People are using experiments. As of 2014, three states are using double-blind sequential administration, and six others have recommended them. This is what an open loop looks like.

  A second error trap identified by the Innocence Project is false confessions, which contributed to 30 percent of wrongful convictions.12 These are often secured from vulnerable people, who are tricked or intimidated into confessing to crimes they didn’t commit. Juan Rivera, you will remember, was a vulnerable young man with a history of psychological problems who confessed after days of interrogation. Police experts said he had experienced a psychotic episode.

  One reform that could help to eliminate false confessions would be to make the videotaping of interrogations compulsory. This would undermine any incentive to bully or mislead suspects into confessions.

  Some police forces worry that such a change might impede their ability to secure confessions from people who are actually guilty. If true, this would count against reform. But a comprehensive review by the Department of Justice found that police departments that had voluntarily taped interviews had not compromised their capacity to secure genuine confessions. As a district attorney in Minnesota put it: “During the past eight years it has become clear that videotaped interrogations have strengthened the ability of police and prosecutors to secure convictions against the guilty.”13

  Another area requiring major reform is forensic science. Some of these techniques, such as hair microscopy, have limited scientific legitimacy. In one murder case, experts “matched” seventeen hairs found at a crime scene with the hair taken from a suspect. He was subsequently convicted. But later testing using hard DNA evidence demonstrated that all seventeen hairs had been misidentified. A pubic hair matched to a male suspect actually belonged to the female victim.14

  It turns out that hair matching is highly subjective. In 2013 the FBI admitted that in more than two thousand cases between 1985 and 2000, analysts may have exaggerat
ed the significance of hair analysis or reported them inaccurately.15 The National Academy of Science has said that hair matching is “unreliable.”16 It was this error trap that condemned Jimmy Ray Bromgard, mentioned in chapter 4, to fifteen years in prison for a crime he didn’t commit.

  And so it goes on. In case after case the Innocence Project discovered predictable pathways to failure; weaknesses that should have been identified and addressed. Other signatures of wrongful conviction include government misconduct, bad advice by lawyers, the use of prison informants (often offered undisclosed incentives to testify against the suspect) and scientific fraud.

  Barry Scheck has suggested reform in each of these areas. But perhaps the most significant reform he has called for is the establishment of Criminal Justice Reform Commissions. These are independent bodies mandated to investigate wrongful convictions and to recommend reforms, along the lines of air-accident investigation teams. As of publication, only eleven states had such commissions.

  In the UK a Reform Commission of sorts was set up in 1995 following a series of spectacular miscarriages of justice, including the Birmingham Six and the Guildford Four. The Criminal Cases Review Commission, an independent body, has the authority to refer questionable verdicts to the Court of Appeal. Between 1997 and the end of October 2013 the commission referred a total of 538 cases.

  Of these, 70 percent succeeded at appeal.

  • • •

  There is an intriguing coda to the Tyson-Bush episode, as Christopher Chabris and Daniel Simons, two psychologists, point out in an essay for the New York Times.17 For it turns out that George W. Bush was wrong about his memories of 9/11, too.

  The former president has often claimed that he saw the first plane crashing into the north tower before going into a classroom in Florida. But he didn’t. There was no live footage of a plane hitting the tower so he couldn’t have seen it before going into the classroom. As Chabris puts it: “Mr. Bush must have combined information he acquired later with the traces left by his actual experience to produce a new version of events, just as Dr. Tyson did.”

  This faulty recollection from Bush also had another effect. People assumed that if he saw footage of the crash before going into the classroom, he must have known about the attacks in advance. Had he also been involved in planning them? people asked. This is the stuff of a now-familiar conspiracy theory. But, in fact, there was no conspiracy. It is just that presidents misremember as well.

  III

  In our discussion of the criminal justice system, we have largely focused on wrongful convictions. But this shouldn’t obscure equally pressing issues. Methods of detection need to be improved to bring unsolved crimes to trial. There is also vital work that needs to be undertaken to reduce the rate at which guilty people walk free. These are tragedies, too, because victims are denied justice and the deterrent effect of the system is undermined.

  There is also the problem of the large number of trials where innocent defendants are put in the dock. The data suggest that the acquittal rate is high. That is often hailed as evidence that the justice system is rigorously acquitting the innocent, but it could also mean that millions of pounds are being wasted on unnecessary trials, with the real culprit still at large.

  The key issue in all of this, however, is not to allow the perceived trade-offs between these objectives to obscure the deeper fact that progress can be made on each of them at the same time. That was the point about wrongful convictions: reforms wouldn’t blunt the teeth of the justice system; on the contrary, reforms would, in many cases, make them sharper.

  There are also other deep-lying problems, features so integral to the fabric of the system that they tend to go unquestioned. Trial by jury, for example, is often held up as sacrosanct, and it may be the most effective form of deliberation in criminal cases. But shouldn’t it be tested? If juries are coming to the wrong conclusions in predictable ways, doesn’t it make sense that procedures should be reformed so that these latent problems are addressed?

  To see how, consider an experiment not on juries, but on judges. Over a ten-month period, Shai Danziger, a neuroscientist at Tel Aviv University, and colleagues analyzed the parole decisions of eight Israeli judges.18 Every day each judge considered between fourteen and thirty-five real-life cases, spending around six minutes on each decision. The verdicts represented 40 percent of the parole decisions made in Israel over the ten-month period. Each judge had an average of twenty-two years of experience.

  Now, judges are supposed to be rational and deliberative. They are supposed to make decisions on hard evidence. But Danziger found something quite different: if the case was assessed by a judge just after he had eaten breakfast, the prisoner had a 65 percent chance of getting parole. But as time passed through the morning, and the judges got hungry, the chances of parole gradually diminished to zero. Only after the judges had taken a break to eat did the odds shoot back up to 65 percent, only to decrease back to 0 over the course of the afternoon.

  The judges were oblivious to this astonishing bias in their deliberations. Criminologists and social workers were also unaware of it. Why? Because it had never been analyzed. As one of the co-authors of the study put it: “There are no checks about the judges’ decisions because no one has ever documented this tendency before. Needless to say, I would expect there to be something put into place after this.”19

  With regard to juries, things are even worse. It is illegal in the UK to even conduct a study on how juries go about their deliberations. The unstated rationale for this prohibition is that if the public find out how juries operate, they might lose confidence in the system. It is an “ignorance is bliss” approach. But this is as intellectually fraudulent as removing the black box from an airplane to insure that people won’t ever find out about pilot error. The result is inevitable: the same mistakes will be made, over and over.

  None of this is to argue that the jury system should be abolished. Many juries do brilliant work under stressful circumstances. It is merely to highlight the almost total lack of evidence as to whether juries are working effectively compared with possible alternatives.* We cannot sustain this approach indefinitely because miscarriages of justice and other high profile mistakes are corroding trust in the system. Criminal justice, like so many other areas of public life, needs to undergo a high-performance revolution based on something that has historically proved almost impossible: learning from mistakes.

  More than twenty years after Juan Rivera was sentenced to life imprisonment for the murder of eleven-year-old Holly Staker, a DNA test was conducted on a blood-stained piece of timber that had been used in a different murder. A man named Delwin Foxworth, who also lived in Lake County, had been savagely beaten with the two-by-four, doused with gasoline, and set on fire. He later died of his injuries having suffered burns over 80 percent of his body.20

  The murderer was never found, but the DNA test was conclusive. The DNA of the blood found on the two-by-four matched that of the semen found in Holly Staker. Police are now almost certain that the man who got away with the rape and murder of an innocent eleven-year-old back in 1992 went on to commit another murder eight years later. Therefore Foxworth may be yet another victim of the wrongful conviction of Juan Rivera—it allowed the real culprit to get away with it and kill again.

  “When we think about miscarriages of justice, we often focus on the person who has been jailed for a crime he didn’t commit,” Steve Art, a New York lawyer, said.21 “But there are other consequences, too. When you convict the wrong person, the real criminal is left to roam the streets, committing crimes with sometimes devastating effects. It is yet another reason why we need to learn the lessons.”

  As for Rivera, he was finally released on January 6, 2012. “I can’t explain it. It’s life all over again,” he said as he walked free. “I just want to experience life. Watch a football game. Just walk on the sidewalk and know that I’m free.” Somebody in the crowd hand
ed him a slice of pizza, which he carried with some embarrassment to a car that had been arranged by supporters.

  His friends have rallied around, but he will never get back the nineteen years he spent in prison. “I would be lying if I said that I have come to terms with what I went through,” he told me. “Even now, I am uneasy and nervous. I can’t sleep at night. I can’t go into crowded supermarkets. When I am walking down the road, I keep looking around. Nineteen years in prison for a crime you didn’t commit leaves a mark.”

  But what about those who were responsible for sending him to jail? How do they feel about it today? Perhaps it should come as no surprise that even now many remain convinced of Rivera’s guilt. In October 2014, Charles Fagan, an investigator who helped obtain Rivera’s confessions, was asked by the Chicago Tribune if he still believed that Rivera committed the murder. “I think so,” he said.22

  And what of the prosecutors? Even after Rivera was released, some Lake County lawyers wanted to put him back on trial. Only with a further conviction would they be able to say that they had been right all along. Only with a conviction could they quell their dissonance. Rivera walking around free was like an accusation against their competence.

  It was left to the Illinois Appellate Court to take what might otherwise seem to be an astonishing step: it barred Lake County from ever prosecuting Juan Rivera for the murder of Holly Staker again.

  Part III

  CONFRONTING COMPLEXITY

  Chapter 7

  The Nozzle Paradox

  I

  Unilever had a problem. They were manufacturing detergent at their factory near Liverpool, in the northwest of England, in the usual way—indeed, the way detergent is still made today. Boiling hot chemicals are forced through a nozzle at super-high levels of pressure and speed out of the other side; as the pressure drops they disperse into vapor and powder.