Black Box Thinking

by Matthew Syed

Little Peter died at the hands of his mother, Tracey, her boyfriend, Steven Barker, and Barker’s brother, Jason Owen. He had suffered terrible abuse and neglect over the course of his short lifetime. Fifteen months after the tragedy the three perpetrators were found guilty of “causing or allowing the death of a child.” They were sentenced to prison.

  But the very next day the media focused its outrage on a very different group of people. The Sun newspaper ran a front-page headline with the words: “Blood on Their Hands.” Other media outlets vented similar outrage. Was their anger directed at accessories to the murder who had not yet been prosecuted? Were there other shadowy figures in the background who had been involved in Peter’s tragic death?

  In fact, the outrage was aimed at those who had been responsible for protecting Peter: mainly his social worker, Maria Ward, and Sharon Shoesmith, director of children’s services for the area. The Sun created a petition calling for their firing and ran photos of them asking “Do you know them?” with a number to call.2 The petition was signed by 1.6 million people.3

  The local council offices were almost immediately surrounded by a crowd holding signs. Shoesmith received death threats. Ward had to leave her home out of fear for her life. Shoesmith’s daughter was threatened with murder, and had to go into hiding.4

  To those at the receiving end the experience felt like something close to the Salem witch trials. Something terrible had happened. The instinct was to insure that something equally terrible happened to someone else. It was the blame game at its most vivid and destructive.

  Many were convinced that the social work profession would improve its performance in the aftermath of the furor. This is what people think accountability looks like: a muscular response to failure. The idea is that even if the punishment is over the top in the specific instance, it will force people to sit up and take responsibility. As one pundit put it: “It will focus minds.”

  But what really happened? Did social workers become “more accountable”? Were children better protected?

  In fact, social workers started leaving the profession en masse. The numbers entering the profession also plummeted. In one area the council had to spend £1.5 million on agency social work teams because it didn’t have enough permanent staff to handle a jump in child protection referrals.5 By 2011 there were 1,350 reported vacancies in child protection work.6

  Those who stayed in the profession found themselves with bigger caseloads. This meant they had less time to look after the interests of each child. They also started to intervene more aggressively, terrified of the consequences if a child under their supervision was harmed. The number of children removed from their families soared. The cost of missing a signal was just too high. The court system sagged under the weight of new cases and an estimated £100 million was needed to cope with the increase in child protection orders.

  There were nonfinancial consequences too. The children taken from their homes were placed into care and with foster families. This meant that the state had to accept a lower quality of foster families to meet demand. Children are often damaged by leaving their own families. Soon, the media had moved into reverse, running stories about the horrors of loving parents having their kids forcibly removed. One headline was: “In Hiding, the Mother Accused of Abuse for Cuddling Her Child.”7

  In Haringey, North London, the situation was even worse. The number of health visitors almost halved. The workload for those who stayed in the profession, already high, escalated. The number of care applications increased by an astonishing 211 percent between 2008 and 2009.8 The British Association for Adoption and Fostering warned that the continuing increase in care applications by England’s local authorities following the Baby P case “could cause a catastrophe in children’s services.”9

  Crucially, defensiveness started to infiltrate every aspect of social work. Social workers became cautious about what they documented, in case it came back to destroy them. The bureaucratic paper trails got longer, but the words were no longer about conveying information, they were about back-covering. Precious information was concealed out of sheer terror of the consequences. The amount of activity devoted to protecting themselves from future bloodletting undermined attention to the actual task of social work.*

  Almost every respected commentator and academic estimates that the harm done to children following the media-driven attempt to “increase accountability” was high indeed.10 Forward-looking accountability collapsed. The number of children killed at the hands of their parents increased by more than 25 percent in the year following the outcry and remained higher for every one of the next three years.11

  When a public inquiry finally reported on the death of Baby Peter, there were allegations that its findings were prejudged and subject to political manipulation. Even the authors of the report seemed to feel that they could not stand in the way of public anger. They worried what might happen to them if they didn’t appease the appetite for a scapegoat. This is what happens in a blame culture.12

  None of this is to assert that blame was not justified in the Baby P case. Like many public institutions in the UK, the social work system would benefit from a vast cultural change directed at it becoming a truly adaptive organization with forward-looking accountability. This book has looked at what such a system looks like, and how it can be achieved. Once a high-performance culture is in place, increasing discipline and accountability is both positive and, indeed, warmly welcomed by most professionals.

  But trying to increase discipline and accountability in the absence of a just culture has precisely the opposite effect. It destroys morale, increases defensiveness, and drives vital information deep underground. It is like trying to revive a stricken patient by hitting him over the head with a hammer.

  Blame has other, more personal consequences, too, particularly in safety-critical industries. Professionals involved in tragedies, such as clinicians or social workers, frequently suffer from post-traumatic stress disorder, even when they are not to blame. They are emotionally scarred by their involvement in a tragedy. This is a very human response and one that needs sensitive handling.

  But when feelings of guilt are compounded by unjustified accusations of criminality, individuals can be pushed over the edge. This phenomenon is now so prevalent that it has led to the coining of a new term: the “second victim.” Studies show that professionals suffer feelings of distress, agony, anguish, fear, guilt, and depression.13 Other studies reveal the prevalence of suicidal thoughts.14

  Sharon Shoesmith was so terrified by the effect of the Baby P affair on her daughters that she contemplated taking not just her own life but those of her entire family. This was a woman described as strong and resolute before she was engulfed in the blame game. “For a moment you can understand how people wipe out their whole family,” she said. “Your pain is their pain and their pain is your pain. And you just want to get rid of the pain for everybody.”15

  In his seminal book Just Culture Sidney Dekker writes: “The question is whether we want to fool ourselves that we can meaningfully wring such accountability out of practitioners by blaming them, suing them, or putting them on trial. No single piece of evidence so far seems to demonstrate we can.”16

  It is time to stop fooling ourselves.

  II

  To conclude our study of blame let us take one final incident, perhaps the most notorious aviation near-miss of the twentieth century. Aviation doesn’t normally penalize mistakes, as we noted in Part 1. The industry has created a culture where errors are not stigmatized, but viewed as learning opportunities. Indeed, aviation is often held up as an industry leader in terms of its culture.

  But on this occasion the industry turned on the professionals. The so-called November Oscar incident was the first time in history that a British pilot was put on trial for doing what he believed to be his duty in high-pressure circumstances.

  What makes the case so fascinating is that it hig
hlights the temptation of the blame game, even in an industry that understands its dangers. And it reveals, once again, how a simple incident can look very different when you look beyond the superficial explanations.

  William Glen Stewart, who had first flown a Tiger Moth as a nineteen-year-old at the RAF base at Leuchars on the east coast of Scotland,17 was one of the most experienced pilots in the British Airways fleet. On November 21, 1989, he was in command on a routine flight from Bahrain to London Heathrow. Also in the cabin were Brian Leversha, the flight engineer, and Timothy Luffingham, the twenty-nine-year-old first officer.

  The short version of the case against Stewart was simple. Flight B747-136 G-AWNO (code name November Oscar) had taken off from Bahrain and, as the flight reached European airspace, the crew had been informed that the weather at Heathrow was dire. Thick fog had reduced external visibility to a just a few feet.

  Stewart would have to make what is called an “instrument landing.” This is where the lack of visibility obliges the crew to rely on various gauges inside the aircraft to bring the plane safely onto the runway. The procedure, which requires the use of autopilot and other internal systems, is far from easy, although not beyond the competence of Stewart.

  Because of the difficulty of the procedure, however, there are a number of safety protocols that have to be followed on approach, rules and regulations that insure that the captain does not take undue risks under the pressure of a tricky landing. The allegation was simple: Stewart had willfully ignored these rules.

  As they came in to land, the aircraft’s autopilot wasn’t picking up the two radio signals being beamed from the end of the Heathrow runway. These are crucial to a successful instrument landing. The beacons guide the plane on to the correct lateral and vertical course. Without them you could be coming in off-kilter. You could be too high, too low, or too far to the left or right.

  If the plane has not captured these beams, the approach must be abandoned no later than 1,000 feet above the ground. A “go-around” must be initiated, which involves discarding the landing and going back into a holding pattern so that the problem can be fixed or an alternative destination with less severe weather conditions selected. Stewart, however, continued with the descent below 1,000 feet, dropping lower and lower in defiance of the rules.

  By the time November Oscar, which had 255 passengers on board, had descended to 750 feet, the plane was so far to the right of the runway that it was actually outside the perimeter fence and flying parallel to the A4 Bath Road. The crew couldn’t see this deviation because the fog was so thick. The plane was now on a collision course with the line of hotels that run alongside the A4.

  Only at 125 feet did Stewart finally order the go-around, but he was a fraction slow. Even as he was revving the engines and pitching up the nose, the plane sank another fifty feet. So close did it come to the roof of the Penta Hotel that it set off the fire sprinklers in the corridors, something the press would latch on to in the aftermath. The undercarriage of the plane was visible to bystanders through the fog as it reached its lowest point, before thundering back into the sky.

  Car alarms started to whoop in the hotel parking lot. Guests dozing in the hotel were rudely awakened. People on the streets scattered as the plane, its bottom half peeping through the mist, reached its lowest point. Up in the cockpit, Luffingham glimpsed the runway lights way off to the left through the mist as November Oscar regained altitude. After the go-around, the plane landed safely, to the applause of the passengers in the cabin.

  An investigation was quickly initiated. A jumbo jet had come within touching distance of what would almost certainly have been the most devastating accident in British aviation history. Had the plane dropped another sixty inches it would have connected with the Penta Hotel, and almost certainly destroyed it.

  To many of the public Stewart’s culpability seemed obvious. Although he had ultimately averted a major disaster, he had disobeyed protocol. His hands had been on the controls when it flew under the mandatory minimum.

  With this in mind one can see why it would have been tempting to pin the incident on Stewart. The heat was on British Airways and the Civil Aviation Authority, the regulator. By pinning it on the pilot they may have hoped to escape censure for poor oversight and procedure.

  Eighteen months later, on May 8, 1991, Stewart was convicted at Isleworth Crown Court in southwest London. The jury decided that he had been guilty of breaking regulations and almost bringing destruction on southwest London. An experienced pilot had become a criminal.18

  But what really happened on that flight? Was Stewart culpable? Was he negligent? Or was he merely responding to a chain of unforeseen events that could have led almost anyone toward disaster?

  In investigating the incident in depth, we will draw upon the seminal report by the journalist Stephan Wilkinson19 and unpublished papers from the trial, as well as confidential documents from the British Airways internal investigation and interviews with eyewitnesses.

  For the deeper story, it turns out, doesn’t begin as a Boeing 747 approaches Heathrow, or even the moment it took off from Bahrain. Rather, it starts two days earlier, as the crew enjoyed a Chinese meal during a stopover in Mauritius.

  III

  It had been a long trip. The crew had been involved in a series of flights in the days before landing in Mauritius and decided it might be nice to unwind by sharing dinner. William Stewart sat alongside Tim Luffingham, the first officer. Engineer Brian Leversha and his wife, Carol, who had come on the trip as well, were also there. It was an agreeable evening.

  But by the time the crew arrived in Bahrain for the next leg of the trip, almost everyone had been struck down with gastroenteritis. Carol Leversha had the worst symptoms of all. Brian had called the local British Airways approved doctor while they were still in Mauritius, but he had been unavailable. Instead, the doctor had recommended a colleague who, although not on the BA roster, was about to be added to the approved list. He dispensed painkillers to Carol and suggested that she give them to anyone else who started to feel ill.

  Two days later, the flight from Bahrain to London was scheduled for 00:14. The so-called slip time (the gap between landing on the previous flight and departure for the next) added to the difficulties of the crew. They had arrived in Bahrain late at night and had gone to sleep. But they had had a full day, and would normally be getting ready for bed again. Instead, they were to fly an overnight into Heathrow. They were also suffering the after-effects of gastroenteritis. It was far from ideal.

  But the crew were professional. They were not going to allow a stomach bug or tiredness to ground a flight containing 255 passengers. As Leversha (now seventy-five) told me when I met him at his home in rural Hampshire: “Some of the crew had suffered worse than others, but there was a consensus that we had gotten over the worst effects. We all felt that it would have been unprofessional to force BA to send out a replacement crew, with all the disruption that would have caused. We wanted to get the job done.”

  The flight itself was grueling from the start. Strong headwinds shrank the fuel reserves. Soon after taking off, Luffingham, the co-pilot, started to feel unwell. It seemed that the gastroenteritis had returned. He borrowed some pills from Carol Leversha, who was in the jump seat, and asked for permission to leave the cockpit. Stewart agreed. Luffingham made his way back into the First Class cabin to get some sleep and use the facilities, leaving Stewart to fly the plane with just the engineer.

  Stewart considered bringing the plane onto the ground at this stage. He and Leversha debated landing at Tehran, one of the only viable stopping points, but they were worried about the fraught political situation in the Iranian capital. Flying on seemed like the prudent thing to do. After all, it wasn’t unusual for a pilot to fly unaided by a co-pilot if the latter had been taken ill.

  By the time November Oscar reached the skies above Frankfurt, however, the situation took a severe turn for
the worse. They were informed that the weather conditions at Heathrow were appalling. Low fog had destroyed external visibility. It was close to zero-zero conditions. They would have to land on instruments in what is called Category 3 conditions (the most demanding kind of landing).

  This posed an immediate problem. Stewart was qualified to fly a Category 3 approach, as was Leversha. But Luffingham, relatively new with British Airways, was not. As they flew over Germany, Stewart radioed to the British Airways office in Frankfurt to ask for a dispensation for Luffingham: essentially, a verbal waiver that would allow the aircraft to land at Heathrow. Frankfurt made the call to London to find out.

  Somewhere in southwest England in the early hours the British Airways duty pilot was awakened by phone. He agreed to a verbal dispensation. It was not considered a significant risk to agree to the dispensation, given that Stewart was fully qualified to make a Category 3 landing. Indeed, these waivers were handed out as a matter of routine.

  By the time November Oscar had reached British airspace, Luffingham was back in his seat. The plane was put into a holding pattern over Lambourne, to the northeast of London. Leversha, from his position behind the captain, was a tad uneasy. Stewart had been flying virtually solo in the dark for more than five hours, with only a fifteen-minute rest. The weather conditions were dire. Fuel was low. He wondered if they should reroute to Manchester, where the weather was better. “Come on, Glen,” he said. “Let’s shove off to Manchester.”

  Stewart considered it. He asked for weather conditions in Manchester, as well as at London Gatwick, and the crew discussed the options. Stewart was on the point of rerouting when Heathrow finally cleared November Oscar for its approach.

  But suddenly there was another complication. They had been due to approach Heathrow from the west, flying out past Windsor before turning around, and landing in an easterly direction. They had the loose-leaf file with the charts of the required route ready at hand in the cockpit. But now Air Traffic Control told them that the fog had lifted ever so slightly, the weather conditions had changed, and that they should therefore land in a westerly direction.