We may see child abuse as worse than abuse of an adult—or the rape of a virgin as worse than the rape of a divorcée—but the greater harm means that harshly punishing the perpetrator isn’t sufficient to restore a “just world.” To do that, we may need to cast the victim as partially responsible.
—
Clearly our labels can change when we are strongly motivated to revise them, as when a person initially categorized as innocent or good suffers a terrible crime. Why, then, do most labels prove so difficult to peel off?
Research suggests that once we have summed someone up, we search for data confirming that identity and disregard or minimize evidence conflicting with it. Of course, it doesn’t feel that way. It feels as though we are just dispassionately sorting through the details. But really our minds are bending the facts, sawing off inconvenient corners, and tossing away contradictory information so that everything can be fit into our ready-made boxes.
Say you learn that a friend’s roommate is a waitress. Without any conscious effort, that label will prompt you to remember things about her appearance, behavior, and lifestyle that are consistent with her occupation—and distinct from what you’ll remember if you’re told she’s a librarian. And if you know she’s having a phone interview in the next room, the words and phrases you’ll pick out through the wall are completely different from the ones you’ll pick out if you think she’s speaking with her lawyer. In an experiment along these lines, those listening to a degraded audio recording of what they believed to be a criminal suspect were significantly more likely to hear the statement “I got scared when I saw what it’d done to him” as “I got scared when I saw what I’d done to him” than when they thought they were listening to a job candidate.
Labels have an amazing way of reinforcing themselves even when that seems impossible. In a great demonstration, two groups of students were asked to draw a young man’s face while looking at his photograph, but one group was told he was black and the other was told that he was white. What is incredible is that the label influenced certain people’s depictions: those who believed that traits like race are immutable and useful in predicting behavior drew portraits that distorted features to align with the racial category they had been given (see this page).
The consequences of this drive for coherence and confirmation can be disastrous. Medical research suggests, for example, that one of the major causes of physician error arises from the fact that doctors come to a conclusion very early in their assessment of a patient and then stick with their judgment even as they encounter contradictory evidence. Unfortunately, the impact on our criminal justice system is no less severe. “Tunnel vision” is an endemic problem in the investigative world, corrupting even seemingly objective judgments like fingerprint analysis.
In fact, DNA testing—the shining jewel of the forensic world—appears to be susceptible to confirmation bias. It’s hard to believe: in television dramas filled with hunches and subjective judgments, the matching of DNA is portrayed as the cut-and-dried part of the case. You’ve got a feeling a suspect is lying to you; you enter his DNA into the FBI’s Combined DNA Index System (CODIS), which has DNA profiles for some twelve million convicts and suspects and about half a million crime scenes; and after a short commercial break, you know for sure.
But the reality is more complicated. In one recent study, researchers gave experts a DNA sample, telling them that it was part of a gang-rape case in which one of the perpetrators had accepted a plea deal and was testifying against the others. They needed to confirm that one of the suspects who denied any involvement had, in fact, taken part in the rape. As expected, the experts found that his DNA matched the sample, corroborating the witness’s account and the prosecution’s case. However, the researchers then took the same biological evidence and gave it to seventeen other DNA analysts without providing any backstory. What happened? Only one of the seventeen experts agreed with the original group that the DNA was a match. Sequences of amino acids do not rearrange themselves to fit the prosecutor’s frame, but that doesn’t mean there is no room for bias. Someone has to interpret the results. And the first set of analysts couldn’t help but be swayed by the information they had already been given.
Confirmation bias likely played a central role in the case of David Rosenbaum. Once David was labeled a drunk, the responders and medical professionals appeared to focus on finding evidence that supported that description. Interestingly, the first people to encounter David, the Pritchetts, did not smell alcohol on David’s person. Why not? Well, perhaps because they weren’t looking for it, since their theory was that the man on the sidewalk had suffered a stroke. For those assuming that this was an ETOH, though, the odor was very noticeable. As one of the firefighters recounted, “I could smell the alcohol reeking from him, like it was coming out of his pores.”
This, in turn, reinforced the notion that the cause of the man’s condition was intoxication, when in fact the smell of alcohol is not very diagnostic at all. According to the U.S. Department of Health and Human Services, well over half of adults in the District of Columbia count themselves as active drinkers. After dinner on a Friday night, there are thousands of people in the city who, like David, have consumed some alcohol in the previous four or five hours. You could sit in the emergency room at Howard during the night shift and find scores of patients—victims of heart attacks, gunshot wounds, allergic reactions, pneumonia—with the faint odor of booze on their breath.
But the problem wasn’t just that the responders engaged in a biased search for clues; it’s that they evaluated the evidence they did uncover in a selective fashion—treating facts that supported their existing understanding of the case as highly persuasive, while dismissing those that did not. For example, one of the firefighters reported that he immediately ruled out the possibility that he was dealing with a diabetic because he looked for a medical identification bracelet, which he often does “when people can’t talk,” and didn’t find one. The lack of a bracelet seemed to reinforce the intoxication explanation, but it doesn’t actually preclude a diabetic reaction at all: the victim might not own a bracelet, might have taken it off, or might not yet have been diagnosed.
Likewise, the Pritchetts and the initial cadre of cops, firemen, and EMTs all noticed that David was wearing an expensive watch and wedding band. According to Commander Robert Contee of the Second Police District, it followed that “there was no reason to think it was a robbery.” But, of course, that’s incorrect. The robbers could have taken his wallet and either been scared off at that point or not noticed his ring and watch.
Indeed, if responders had gone in with the assumption that they were dealing with a crime, they most likely would have uncovered powerful evidence within minutes of arriving, like a witness down the street who’d seen David’s attackers just minutes earlier and noted that they looked suspicious (going so far as to write down the first digits of their license plate number). With a different starting frame, they might have noticed that the back pocket of David’s pants was torn—a potential sign that his wallet had been ripped out during a robbery. As it was, the first person to spot the tear was a nurse at Howard cleaning David up. And she wasn’t in a position to make the connection because she didn’t know the circumstances in which David had been discovered.
A big part of why labels stick to victims is that, in many cases, we aren’t even aware that we are operating with a label or that it might be wrong. We never revisit our initial assumptions to test whether they still hold (if they ever did). Numerous people who saw David accepted the label of drunk without questioning the frame. And they did so in the face of strong counterindications. If the man was a drunk, where was the booze, and how had he gotten so intoxicated without a wallet or any money? How had he come to rest on the sidewalk in a wealthy residential neighborhood when he wasn’t able to lift himself off the ground on his own? There was actually plenty of evidence suggesting that this was a much more serious condition than ETOH: the man’s inability
to speak, his failure to respond to oxygen, his vomiting, his extremely poor motor control, his pinpoint pupils, his elevated pulse, his bleeding head, and his inability to use the right side of his body, among other things. But each one was ignored or minimized because the established narrative didn’t accommodate it.
The failure to appreciate counterevidence was almost certainly exacerbated by the fact that the responders did not make independent assessments of David. If they had each come across him separately and attempted to determine the cause of his condition, they might well have reached different conclusions, and that could have led them to discover the head injury much more quickly. The lack of independent assessment is puzzling because police departments, with all their experience conducting eyewitness interviews, are well aware that when one person, for whatever reason, makes public his observations and conclusions very early on in the process, others feel pressured to conform. This is why when, say, six people witness a bank robbery, detectives usually elicit statements and conduct lineups with each one individually.
In the end, it’s hard not to ask the “what if” questions in David’s case. What if the criminals had used a knife instead of a pipe? What if David had been carrying his ID? What if his wife had gone looking for him and had been there when the ambulance arrived, or met the hospital staff at Howard? What if the Pritchetts had driven David to the hospital and then conveyed their observations directly to the triage nurse?
All of these things might have radically altered the course of events the night David was attacked, but they didn’t happen—and it doesn’t just come down to bad luck. The errors were predictable and preventable. That fact makes David Rosenbaum’s death especially tragic, but it also offers a glimmer of hope—one that David’s family seized upon. They decided to offer the District of Columbia a deal: they would drop their lawsuit against the city if it created a task force to reform the Department of Fire and Emergency Medical Services to help prevent similar tragedies in the future. A number of these reforms, like ensuring that both EMTs and firefighters are fully prepared to carry out medical protocols and that they are held accountable for their performance, can help future responders avoid the critical mistakes that were made in David’s case. But we need to go further to get at the root of the problem: preventing damaging labels from taking hold in the first place.
When we know that extraneous elements can lead to harmful labels, we should consider intervening to counteract them. Controlling the influence of disgust, for example, may be as simple as providing first responders with menthol salve to put under their noses to neutralize the smell of vomit, feces, or a dead body. We also need to take action when we know that victims with particular personal characteristics—the elderly, minorities, homeless people, addicts—tend to be devalued and receive substandard treatment. We could start by introducing more empathy training designed to get police officers, EMTs, and ER nurses to see members of these groups—who are overrepresented in 911 calls—as human beings worthy of compassion and careful treatment.
Perhaps most important, responders should conduct their own blind assessments of victims—whether that involves giving a cognitive evaluation, searching the crime scene, or conducting a forensic test. Data (“His eyes are not dilating properly, and his pulse is abnormal”) should be shared; labels (“He’s a drunk”) should not.
Treating someone based on a label—such as ETOH—should always be a last resort. When we act upon a single label, we almost always lose vital nuance. With the power failing and the floodwaters rising as Hurricane Katrina made landfall, the medical chairman at Memorial Medical Center in New Orleans had to make an evacuation plan. He used a shortcut to help sort the patients: those who possessed do-not-resuscitate orders would be given lowest priority in leaving the hospital. In that panicked moment, the choice seemed justifiable—as he explained later, these folks seemed to have the “least to lose”—but, in retrospect, using that label as a sorting mechanism was a grave mistake. It did not help doctors distinguish between those who were going to die anyway and those who could be saved; nor did it divide those willing to die from those holding hard to life. A DNR simply informs medical professionals not to engage in CPR if the person stops breathing or her heart ceases beating; the order is not meant to affect the delivery of other medical care. Moreover, a person may choose to have a DNR whether or not she is seriously ill. Using the label, however, gave the chairman the illusion of objectivity.
The patients abandoned on the seventh floor were never just one label. They carried an almost infinite array of tags—recovering, terminal, insured, elderly, forty-something, mother, son, fat, handsome, combative, kind. A few of these were written on their charts, or read from their faces, coughs, and trembling hands. But most were invisible.
No one deserves to be summed up in a word. And no one should die because of an acronym.
2
DANGEROUS CONFESSIONS
The Detective
The back door of the apartment had been bashed open with a blue mop. Someone had washed bloody hands in the sink. And if you walked right up to the front staircase and bent down, you could see where those same hands had brushed the wall near the banister—streaks of red fresh against the white.
Dawn Engelbrecht knew that something was wrong as soon as she saw her five-year-old boy walk into the bar where she worked. Holly Staker was supposed to be watching him and his younger sister, but there he was, little Blake, standing in front of her. A neighbor had noticed him playing outside in the dark and had brought him by.
No one picked up when she called the house, but she got through to Holly’s mother, and the two of them met at 442 Hickory Street.
The house is on the north side of Waukegan, Illinois, a classic brick two-flat set close to the neighboring houses. Most people would call it a nice street, though it has known better times. The houses have steps cut across the wide green verge between the street and the sidewalk. But 442 is different from the others: the footpath follows the left side of the house around to the back. You have to jog right if you want the front door.
It was locked, as Dawn had left it. But when she finally turned the key of the second-floor apartment, she found the television on. A single white tennis shoe lay on the floor. A chair in the dining room was turned over.
They called for Holly, but there was no answer. Taylor, the two-year-old, was fine, though—safe and sound, lying on her brother’s bed. Maybe Holly had just gotten bored and left?
It was not until the police arrived that Dawn thought to look behind the bedroom door.
Eleven-year-old Holly was curled in the fetal position, her hands up by her face, her black stretch pants cast aside, the missing white shoe tangled in one of the stirrups. She had been stabbed twenty-seven times and raped.
A year later, twelve jurors fixed their gaze on Juan Rivera, the man accused of the crime. Juan—a petty criminal, barely twenty—sat motionless. They had heard the evidence, the alibis, the reports, listened as each witness and expert gave his or her account. They had seen the pictures of the bedroom and of Holly’s face, and it was time to decide.
Guilty.
And so it went at the second trial as well, ordered by the Appellate Court of Illinois after Juan’s initial conviction was overturned. Different jury, same result.
For a man like Juan, sentenced twice to life in prison without the possibility of parole, to get a third chance before a jury is as rare as a snowflake in June. But his lawyers found a way: DNA evidence.
It had always been there, those twelve years Juan sat in prison. A vaginal swab had been collected at Holly’s autopsy. But after being labeled, it sat in an evidence locker until 2005, when it was finally tested.
The results of the analysis were startling. The semen in the sample belonged to a single man, and that man was not Juan Rivera.
Proof. Vindication. After so long in the dark, it felt as if someone had lit a lantern inside his body. He would walk out of that prison. He would be reborn. T
here was life to live.
It took four more years to reach that third trial, but the moment finally came—the opportunity for science to redeem the failures of the past and save this condemned man. With the lab test offering exoneration and with no physical evidence linking Juan to the crime, the outcome seemed like a foregone conclusion.
But something curious happened once proceedings got under way. The prosecution didn’t back away; it doubled down, advancing two explanations for the autopsy evidence: either the DNA sample had been contaminated or Holly had had sex with someone else before she was raped and killed by Juan.
The problem for the prosecution was that there was nothing to suggest that the swab had been compromised, and the experts were in agreement that the semen had been deposited shortly before the victim was killed. Semen tends to drain into underwear, but no sperm was found on Holly’s clothes. That meant that the prosecution had to convince the jury that an eleven-year-old girl had sex with a mystery man directly before being violently attacked by Juan Rivera, who managed to leave the mystery man’s semen completely intact in the girl’s body and to provide no trace of ever having touched her himself, let alone of having been inside the apartment where she was found. The account seemed implausible, to say the least.
Once more, twelve jurors retired to the deliberation room to ponder the tragedy at 442 Hickory Street. When they returned, they were met with the gaze of Holly’s twin sister, Heather, and Juan’s brother, Miguel. Juan, on the threshold of exoneration, was silent and ready.
Guilty.
It was a verdict that Juan’s lawyers described as “unfathomable,” but it had happened. Again.
The defense had provided many facts to raise a reasonable doubt in the eyes of the jury. It wasn’t just that none of the copious forensic evidence at the crime scene—blood, fingerprints, hair, and semen—matched Juan. There were also phone records and data from an electronic leg monitor Juan was wearing (a condition of bail for an earlier charge of stealing a car stereo) corroborating his parents’ claim that he had been home the night of the murder, talking to his mother in Puerto Rico. Yet, despite Juan’s alibi and the lack of either eyewitness testimony or physical proof, the prosecution got the conviction it sought because of evidence so damning that it made everything else in the trial seem superfluous: a signed three-page confession.
Unfair Page 4