A couple of comments on my recent post brought up the subject of cognition error. This reminded me of a chapter on the subject in my book, The Importance of Being Honest (NYU Press, 2008), which I have pasted below:
HOW LAWYERS (OUGHT TO) THINK
Shortly after 7 o’clock on the morning of April 16, 2007, campus police at Virginia Tech University received word of a shooting in the West Ambler Johnson dormitory. Arriving at the scene, officers found that a resident advisor named Ryan Christopher Clark had been killed, and a freshman named Emily Jane Hilscher had been fatally wounded. The police immediately “locked down” the dormitory, but neither they nor the university administrators took additional steps to secure the campus. Classes were not canceled, no perimeter was set up, and students were not notified of the shooting (or warned to stay home) for almost two hours. Even the lockdown at West Ambler Johnson was apparently lifted within about 30 minutes. Meanwhile, a 23 year old student named Seung-Hui Cho was reloading his weapons and mailing a chilling videotape to NBC News. At about 9:30, Cho entered Norris Hall, a large classroom building, armed with two semi-automatic handguns and 17 magazines of ammunition. He began shooting, murdering 32 people, and wounding another 28, before he committed suicide.
In retrospect, it is easy to see that the campus authorities made an awful mistake. Knowing that a killer was on the loose, they failed to take relatively simple precautions that almost certainly would have limited the carnage. Cho would have found many fewer victims if only Norris Hall hadn’t been so full of students and teachers, yet university administrators allowed classes to convene as though nothing had happened in the Ambler Johnson dorm. Tragically, however, the investigation followed a blind lead, causing the Virginia Tech leadership team to believe that no one else was in immediate danger. In fact, the police initially turned their attention away from campus, tracking down an innocent man, while Cho continued to make his deadly preparations.
Among the first potential witnesses questioned was Hilscher’s roommate, Heather Haugh. She told the police officers that Emily had spent the weekend with her boyfriend Karl, a student at nearby Radford University who lived in off-campus apartment, and that he presumably had brought her back to the dormitory earlier that morning. Although the police knew nothing else about Karl (whose last name I am omitting for obvious reasons), that immediately made him a “person of interest.” Further questioning revealed that Karl was an avid gun owner and had recently taken Emily to a shooting range. That that was enough to turn him into the prime, indeed exclusive, suspect. Despite Haugh’s insistence that Karl was not a violent person, the police went racing off to find him – they later stopped his car and searched his home for a murder weapon – leaving the campus essentially unguarded and the unfortunate students uninformed.
How could experienced detectives make such a fatal misjudgment? Far from being irresponsible or negligent, they were in fact pursuing established investigative procedures. Sadly but not inexplicably, their training and instincts took them in precisely the wrong direction because they succumbed to a series of well-documented “cognition errors.” The relatively new field of cognitive psychology has identified a number of ways in which human beings routinely err in evaluating information, either by relying on “heuristics” (or shorthand approaches to problem solving) or by indulging certain “biases” (meaning innate decision-making preferences, not racial or other prejudice).
In the Virginia Tech investigation, the first pitfall was an “availability heuristic,” or the tendency to believe that a newly observed event falls into a frequently encountered, known category. Whenever a woman is killed, especially a young woman, the first suspect is almost always a boyfriend or husband, and statistics bear out that grim assumption. Thus, police are taught to look for signs of domestic violence whenever there is a female homicide victim. The more domestic violence they’ve seen the more likely they are to look for it in each succeeding case, because the image of a jealous or estranged lover becomes increasingly “available.” That appears to be exactly what happened at Virginia Tech, as the investigating officers focused their attention on Hilshcer’s boyfriend. Even though Karl had no history of violence, and Haugh assured them of his “amazing relationship” with Emily, he nonetheless fell into the category that came most easily to mind.
Then the police learned that Karl owned firearms, leading almost inexorably to “confirmation bias.” Having already concluded that Karl was the possible killer, his avid gun ownership – although not uncommon in rural Virginia – powerfully confirmed their suspicions. There is nothing wrong, of course, about putting two and two together premised on a combination of probabilities and facts. But the deductive leap from gun-owning boyfriend to double murderer turned out to be worse than unfounded. Once they fastened on Karl as the killer, the police evidently ruled out or discounted other, less available possibilities. And knowing that he did not attend Virginia Tech, the investigation disastrously veered away from campus.
The third and most dangerous cognitive error, therefore, was the “search satisfaction” fallacy – the inclination to stop looking for solutions to a problem upon the discovery of a satisfactory, if not ironclad, answer. It was not a mistake for the police to pursue Karl, based on everything they knew at the time. But it was a catastrophe when they assumed that they had identified the killer, and therefore ignored the real gunman who was stalking the campus for additional victims.
None of this is intended to disparage, or even second guess, the Virginia Tech detectives, who are surely heartbroken over the loss of so many lives. The point, rather, is that cognition errors are inherent in human perception, and no one (least of all judges or lawyers) is immune.
HARD WIRING
There are many other common cognition errors. The “causation fallacy,” for example, creates a presumption of a (frequently nonexistent) causal relationship between events, merely because one followed the other, just as “the pattern heuristic” may create imaginary relationships among isolated occurrences. Closely related is the problem of “hindsight bias,” which suggests that an event was foreseeable (and therefore preventable) simply because it ultimately occurred. The “representativeness heuristic” may lead to assessments based on generalities, overlooking atypical or inconsistent factors, while “attribution errors” do just the opposite, focusing on a single (usually negative) characteristic to the exclusion of everything else. The “affective error” reflects a natural tendency to overvalue information consistent with our wishes or needs. “Commission bias” results in a preference for action over inaction, and “momentum bias” can prevent us from reevaluating earlier conclusions or otherwise changing our minds.
In practice, cognition errors are difficult to recognize because they are both subtle and inconsistent, potentially leading in many different directions. We might err either because we rely too much on generalities (domestic violence), or alternatively because we zero in on specifics (the boyfriend owned a gun), or both. The unifying factor for all cognition errors is that they are forms of conclusion-jumping, reaching decisions through often helpful but frequently misleading analytical shortcuts.
The prevalence of cognition errors exposes a profound flaw in the standard conception legal fact finding, in which judges or jurors receive information in a linear fashion – first one discrete fact, then another and another – withholding judgment until the very end of the evidence. As we saw in the Virginia Tech investigation, the reality is far different. It is impossible to withhold judgment because new information is immediately evaluated and interpreted in light of preexisting knowledge and reference points, and the mind is constantly drawing seemingly logical (but perhaps inaccurate) connections among facts. Without this sort of conclusion-jumping we could not make ready sense of the world, but unfortunately it can also lead us badly astray. In either case, it is hardwired in our genetic code.
A thought experiment proves the point. Think back to the origins of humanity, when our earliest hominid ancestors first walked upright on the plains of East Africa. Imagine one such individual – we can call her Lucy – who is out on her daily hunt for roots and berries. She spies a lithe and tawny animal in the distance, with a shaggy mane and a long tale. “Hmm,” she might say to herself. “That looks very much like the creature that ate my cousin last week.” At this point, Lucy’s neuro-pathways might take her in either of two directions. Under the traditional, judicial model, she would withhold judgment and investigate further. “Let’s not be hasty,” she would figure. “I’d better get closer to make sure.”
Even if the beast roared and flashed its fangs, the incrementalist proto-human still would try to accumulate more information before taking precipitous action. That approach is obviously an evolutionary dead-end. Investigating a saber-tooth tiger would lead to a speedy demise, and conclusion-reserving hominids would not live long enough to reproduce. Thus, if there ever was such an inherent disposition, the forces of natural selection would surely have eliminated it en route to the development of Homo sapiens.
In contrast, conclusion-jumping hominids would have headed for the figurative hills at the first sight of anything that resembled a carnivorous quadruped, thereby increasing their chances of survival and procreation. The knack for quick decision-making would have been passed along from generation to generation, resulting in the heuristic phenomenon that is still observable today.
Life was nasty, brutish and short in age of Australopithecus when our cognitive reactions first evolved, so Lucy and her heirs were naturally bred to take few chances, avoid over-thinking, and assume the worst. Those preservationist tendencies clearly account for many of the most frequent heuristics – such as commission bias, affective error, and the causation fallacy – all of which would have helped our distant ancestors stay out of trouble most of the time. An occasional false impulse was a relatively small price to pay in exchange for evading the many ever-present life-endangering threats.
Today’s world, however, is not as physically dangerous as it was in Lucy’s time, nor are our decisions as simple as fight or flight. Thus, certain neuro-heuristics that were once well-adapted to survival have endured in the form cognition errors, especially when contemporary professionals – doctors, detectives, military officers, rescue workers, and of course judges – are called upon to make complex high-stakes decisions in which the most deductively obvious solutions may frequently be wrong.
HOW LAWYERS THINK
There is a substantial literature is some professions – medicine, clinical psychology, and social work, for example – addressing the need to recognize and correct for cognition errors. Interestingly, very little has been written on the subject for lawyers, and even less for judges. It is deeply regrettable that judges are not more attuned to potential cognition errors, although it is understandable given that they are subject to extreme confirmation bias. A judge’s resolution of disputed facts is self-confirming, seldom disturbed even on appeal, thus providing little occasion for reexamination. Taking cognition bias to the extreme, Justice Antonin Scalia once remarked about death penalty appeals that, “I have been on the court for 20 years and I have not seen a case where I thought there was the slightest doubt about the person’s innocence.” Statistically, that would seem to be impossible; heuristically, it makes perfect sense.
Practicing lawyers face a different challenge. They need to be aware of their own possible cognition errors, which can cause problems interviewing clients, negotiating with other counsel, assessing witnesses, or evaluating evidence. Uniquely among professionals, however, trial lawyers must also consider the potential cognition errors of judges and jurors, in order to counteract the most likely mistakes or, to put it politely, to accommodate them.
The subject is touchy because the trial process inescapably relies on perceptions, usually second-hand perceptions, to determine facts. Thus, lawyers are often accused of manipulation or deception – taking advantage of a jury’s biases, prejudices, or naiveté – which of course could include exploiting predictable cognition errors.
Possible instances quickly come to mind, especially in tort litigation. Did a patient take a particular medication and later suffer a heart attack? The causation fallacy will certainly imply a link. Did the same thing happen to many others? The pattern heuristic will tend to tie them together. Can the heart attacks be traced directly to the drug? That would be the attribution error. Should the pharmaceutical manufacturer have anticipated the danger? Yes, certainly, considering hindsight bias.
So what is the truth? Did the hypothetical medication cause heart attacks, or is that just a crafty, error-loaded insinuation? There is no abstract answer, of course, because some real drugs are foreseeably harmful and others are not. That is why we have trials. As to the role of cognition errors, sometimes they can influence a determination and sometimes they can be neutralized or remedied. That is why we have lawyers, judges, rules of evidence, and due process.
For better or worse, it is an advocate’s job to draw plausible connections and a judge or jury’s job to sort them out. Fortunately, the legal system comprises a number of correctives that can reduce, if not completely eliminate, the impact of cognition errors. Most important among these – indeed, the bedrock – is the requirement that all testimony be truthful. Lawyers are not free, as some critics charge, to spin stories out of dross, conjuring facts and inventing documents virtually at will. Yes, there are some cheats and fabricators but, stereotypes notwithstanding, the great majority of lawyers are devoted to ethical practice would never stoop to falsifying evidence.
Even when a lawyer plays fast and loose, with either facts or inferences, opposing counsel is there to point out the missing evidence or logical flaws. For every attorney who tries to take advantage of cognition errors, another is ready to rectify false impressions and caution against mistaken conclusions. Because every argument is subject to rebuttal, the adversary system itself minimizes opportunities to exploit fallacies and heuristics.
Judges, too, play a role, weighing the probative value of evidence against the danger of unfair prejudice and its potential to confuse or mislead. Some rules of evidence specifically address certain cognition errors. For example, the attribution fallacy – that is, the inclination to base broad judgments on a single negative trait – is somewhat neutralized by the rule against offering character evidence to prove that a person acted “in conformity therewith on a particular occasion.” In similar fashion, the rules on expert testimony limit the admission of “junk science” that is based more on inference and coincidences than sound data and well-researched etiology.
The most significant, and underappreciated, institutional shield against cognition error is the much maligned citizen jury. An individual judge, however well educated and thoughtful, considers cases from only one perspective; there is no ready counterweight when he or she makes a hasty or ill-founded conclusion. A jury, on the other hand, always brings multiple perspectives to bear on every decision. Despite all the stories of confused or misguided or “runaway” juries, the very process of deliberation tends to highlight, and therefore neutralize, any single juror’s cognition errors. In a group of six or twelve, it is extremely unlikely that everyone will simultaneously use the same heuristics, such as availability fallacies or confirmation errors, because each person begins from a different set of assumptions, preferences, and experiences. While every juror is susceptible to cognition errors, collectively they will be able to catch, correct, and cancel out many of each others’ mistakes.
Nonetheless, cognition errors are inevitable, widespread, usually undetected, and too often dangerous. They can cause serious harm in almost every field, potentially costing lives in medicine and law enforcement, and threatening freedom and property in court. Cognition errors are most pernicious because they always seem sensible – in fact, logical and compelling – when they are made. For lawyers, there is no universal solution to the problem, apart from vigilance, scrutiny, and in you can get one, a well chosen jury.
Excellent analysis of cognitive biases. I hope you enjoy reading my book.
Posted by: Scott Fruehwald | March 08, 2018 at 03:13 PM
Good post. I am always amazed at the cognitive error which takes place when these events happen with unerring regularity but the reaction is always, "It was so unpredictable." "Who would have thought it could happen here?"
Actually it is a virtual certainty that it will happen again, and again, and again. While the police are out looking for the 'real killer' maybe they should ask why these things don't happen in Japan and Hong Kong -- and then look for the 'real killer' in the social conditions of this country.
Posted by: Litowitz | March 08, 2018 at 11:27 PM
So ironic: a comment on a post about cognitive bias suggests that the US ranks high in the world for murders.
See, UNODC murder rates (per 100,000 inhabitants).
The US is 96th.
Murder rates by firearms?
US is 17th, and some major populations, like China's, were not included.
The US does have an alarming suicide rate: white males being particularly vulnerable. Every ask yourself the reason?
Of course not. Bias doesn't permit truth for the true believers.
Oh, and, if anyone has the slightest interest, check out the countries where the murder/gun violence actually is the worst.
Posted by: anon | March 09, 2018 at 12:31 AM
Americans are 10 times more likely to be killed by guns than people in other developed countries, a new study finds. Compared to 22 other high-income nations, the United States' gun-related murder rate is 25 times higher. And, even though the United States' suicide rate is similar to other countries, the nation's gun-related suicide rate is eight times higher than other high-income countries, researchers said. The study was published online Feb. 1 in The American Journal of Medicine.
"Overall, our results show that the U.S., which has the most firearms per capita in the world, suffers disproportionately from firearms compared with other high-income countries," said study author Erin Grinshteyn, an assistant professor at the School of Community Health Science at the University of Nevada-Reno. "These results are consistent with the hypothesis that our firearms are killing us rather than protecting us," she said in a journal news release.
The review of 2010 World Health Organization data also revealed that despite having a similar rate of nonlethal crimes as those countries, the United States has a much higher rate of deadly violence, mostly due to the higher rate of gun-related murders. The researchers also found that compared to people in the other high-income nations, Americans are seven times more likely to die from violence and six times more likely to be accidentally killed with a gun. "More than two-thirds of the homicides in the U.S. are firearm homicides and studies have suggested that the non-gun homicide rate in the U.S. may be high because the gun homicide rate is high," Grinshteyn said. "For example, offenders take into account the threat posed by their adversaries. Individuals are more likely to have lethal intent if they anticipate that their adversaries will be armed," she explained.
Even though it has half the population of the other 22 nations combined, the United States accounted for 82 percent of all gun deaths. The United States also accounted for 90 percent of all women killed by guns, the study found. Ninety-one percent of children under 14 who died by gun violence were in the United States. And 92 percent of young people between ages 15 and 24 killed by guns were in the United States, the study found.
Murder is the second leading cause of death among Americans aged 15 to 24, the study found. The research also showed that murder was the third leading cause of death among those aged 25-34. Compared to those in the same age groups in other wealthy countries, Americans aged 15-24 are 49 times more likely to be the victim of a gun-related murder. For those aged 25-34, that number is 32 times more likely, the research revealed.
The researchers also point out that numerous studies suggest that reduced access to guns would lower the suicide rate in the United States. "Differences in overall suicide rates across cities, states and regions in the United States are best explained not by differences in mental health, suicide ideation, or even suicide attempts, but by availability of firearms," said study co-author David Hemenway, professor of health policy at Harvard University's School of Public Health, and director of the Harvard Injury Control Research Center and the Harvard Youth Violence Prevention Center, all in Boston. "Many suicides are impulsive, and the urge to die fades away. Firearms are a swift and lethal method of suicide with a high case-fatality rate," he noted in the news release.
Posted by: Patrick S. O'Donnell | March 09, 2018 at 01:58 AM
The above is by By Robert Preid for CBS News. As was noted on an NPR post summarizing the global statistics, "It is a little surprising that a country like ours should have this level of gun violence," says Ali Mokdad, a professor of global health and epidemiology at the IHME. "If you compare us to other well-off countries, we really stand out."
Posted by: Patrick S. O'Donnell | March 09, 2018 at 02:05 AM
The stats I posted above are cited from the UNODC, an agency established to assist the UN "in better addressing a coordinated, comprehensive response to the interrelated issues of illicit trafficking in and abuse of drugs, crime prevention and criminal justice, international terrorism, and political corruption."
The UNODC employs "between 1,500 and 2,000 people worldwide, has its headquarters in Vienna (Austria), with 21 field offices and two liaison offices in Brussels and in New York City. The United Nations Secretary-General appoints the agency's Executive Director."
Now, rather than just cutting and pasting something from a CBS News report (that really cherry picks stats to make a point), tell us the reason the UNODC numbers cited above are flawed or untrue. There is no reason to believe that.
And, as usual, there is no relationship between the rebuttal and the claim putatively refuted. The comment above "maybe they should ask why these things don't happen in Japan and Hong Kong -- and then look for the 'real killer' in the social conditions of this country." was addressed by the fact that "these things" do happen in other countries: and, at a greater rate at that.
Posted by: anon | March 09, 2018 at 03:01 PM
The studies conducted by the relevant institutions as summarized in the report were not invoked so as to "tell us the reason the UNODC numbers cited above are flawed or untrue," and it's rather stunning that one would draw such an inference, as that is not in any way implied or claimed in what was provided. Rather, it demonstrates the appropriate scheme of comparison, that is, with other fairly affluent nation-states more or less similar to ours in relevant respects: legal regime, development, welfare and well-being, etc., etc. And as the U.S. is an outlier with regard to gun ownership and gun violence, it suggests rather starkly that we need not live and die with gun violence, that other countries have well demonstrated how it is possible to avoid this sort of violence that plagues our country (and that does not mean 'throwing out' the Second Amendment, although it wouldn't hurt if its second clause was properly interpreted in light of the first clause, in which case individual gun ownership is not a constitutional right, given that we have all manner of 'well regulated militias' at both the state and federal levels), in other words, there is much we can and should learn from these countries as to how to avoid gun violence.
Posted by: Patrick S. O'Donnell | March 09, 2018 at 03:58 PM
PTSOD
You want to debate the "gun: issue (with all tired tropes).
That wasn't the issue to which I responded.
And, as usual, there is no relationship between the rebuttal and the claim putatively refuted. The comment above "maybe they should ask why these things don't happen in Japan and Hong Kong -- and then look for the 'real killer' in the social conditions of this country." was addressed by the fact that "these things" do happen in other countries: and, at a greater rate at that.
As for your knee-jerk recital of stats on "gun violence" (now drifting into a reinterpretation of the Second Amendment!) suffice it to say, that is a different debate. At least you concede that, if we compare the US only to the safest countries in the world, we can get skewed results about the rate of crime and violence.
ANd, again, I would invite those who think "it doesn't happen in other countries" to leave behind all their biases, and start thinking about the truth (e.g., check out the countries where the murder/gun violence actually is the worst and think about your stance in more comprehensive terms).
Posted by: anon | March 09, 2018 at 05:03 PM
anon,
The exquisite exhibition of coherence and logic exemplified in your reply has cowed me into silence: ‘tis a shame we don’t know the source of such brilliance.
Posted by: Patrick S. O'Donnell | March 09, 2018 at 05:37 PM
Right. But, is there a reading list?
Posted by: anon | March 09, 2018 at 05:40 PM
From a Fourth Amendment analysis and exigent circumstances, the officers acted reasonably at the time. Domestic Violence is probably the biggest issue/crime that police/prosecutors/courts and us criminal defense lawyers deal with on a daily basis.
Posted by: Deep State Special Legal Counsel | March 09, 2018 at 07:07 PM