Noise: A Flaw in Human Judgment

by Daniel Kahneman

People have a remarkable intuitive ability to match intensities across unrelated dimensions, by mapping one intensity scale onto another. You can match the intensity of your affection for different singers to the height of buildings in your city. (If you think that Bob Dylan is especially great, for example, you might match your level of enthusiasm for him to the tallest building in your city.) You could match the current level of political discord in your country to a summer temperature in a city you know well. (If there is remarkable political harmony, you might match it to a breezy seventy-degree summer day in New York.) And if you were asked to express your appreciation of a restaurant by comparing it to the length of a novel instead of the usual 1-to-5-star rating scale, this request would strike you as quite bizarre but not at all infeasible. (Your favorite restaurant might be like War and Peace.) In each case, it is—oddly—quite clear what you mean.

In ordinary conversation, the range of values for a scale is a function of the context. The

The main implication of this view of confidence is that subjective confidence in one’s judgment by no means guarantees accuracy. Moreover, the suppression of alternative interpretations—a well-documented process in perception—could induce what we have called the illusion of agreement (see chapter 2). If people cannot imagine possible alternatives to their conclusions, they will naturally assume that other observers must reach the same conclusion, too. Of course, few of us have the good fortune of being highly confident about all our judgments, and all of us have had the experience of uncertainty, perhaps as recently as your reading about Julie 2.0. We are not all highly confident all the time, but most of the time we are more confident than we should be.

When the triggers of pattern noise are rooted in our personal experiences and values, we can expect the pattern to be stable, a reflection of our uniqueness.

Today, the dominant model of personality, the Big Five model, combines traits into five groupings (extraversion, agreeableness, conscientiousness, openness to experience, neuroticism), with each of the Big Five covering a range of distinguishable traits. A personality trait is understood as a predictor of actual behaviors.

Common sense suggests that while behavior may be driven by personality, it is also strongly affected by situations.

Speaking of Pattern Noise “You seem confident in your conclusion, but this is not an easy problem: there are cues pointing in different directions. Have you overlooked alternative interpretations of the evidence?” “You and I have interviewed the same candidate, and usually we are equally demanding interviewers. Yet we have completely different judgments. Where does this pattern noise come from?” “The uniqueness of people’s personalities is what makes them capable of innovation and creativity, and simply interesting and exciting to be around. When it comes to judgment, however, that uniqueness is not an asset.”

CHAPTER 17 The Sources of Noise

When we began our research, we were focusing on the relative weights of bias and noise in total error. We soon concluded that noise is often a larger component of error than bias is, and certainly well worth exploring in more detail.

The evidence gradually led us to realize that the noisy judgments that different people make are largely determined by something that is neither a general bias of the individual nor transient and random: the persistent personal reactions of particular individuals to a multitude of features, which determine their reactions to specific cases. We eventually concluded that our default assumption about the transient nature of pattern noise should be abandoned.

Though we want to be careful not to overgeneralize from what remains a limited selection of examples, the studies we have assembled, taken together, suggest that stable pattern noise is actually more significant than the other components of system noise. Because we rarely have a full picture of the components of error in the same study, it requires some triangulation to formulate this tentative conclusion. In short, here is what we know—and what we don’t.

In principle at least, level noise—or simple, across-the-board differences between judges—should be a relatively easy problem to measure and address. If there are abnormally “tough” graders, “cautious” child custody officers, or “risk-averse” loan officers, the organizations that employ them could aim to equalize the average level of their judgments. Universities, for instance, address this problem when they require professors to abide by a predetermined distribution of grades within each class. Unfortunately, as we now realize, focusing on level noise misses a large part of what individual differences are about.

Noise is mostly a product not of level differences but of interactions: how different judges deal with particular defendants, how different teachers deal with particular students, how different social workers deal with particular families, how different leaders deal with particular visions of the future. Noise is mostly a by-product of our uniqueness, of our “judgment personality.” Reducing level noise is still a worthwhile objective, but attaining only this objective would leave most of the problem of system noise without a solution.

our ordinary way of making sense of the world around us makes it all but impossible to recognize the role of noise.

right. Like any other unsurprising story, a success story explains itself once the outcome is known.

Noise is inherently statistical: it becomes visible only when we think statistically about an ensemble of similar judgments.

Speaking of the Sources of Noise “We easily see differences in the average level of judgments, but how large is the pattern noise we do not see?” “You say this judgment was caused by biases, but would you say the same thing if the outcome had been different? And can you tell if there was noise?” “We are rightly

PART V Improving Judgments

Three things matter. Judgments are both less noisy and less biased when those who make them are well trained, are more intelligent, and have the right cognitive style. In other words: good judgments depend on what you know, how well you think, and how you think. Good judges tend to be experienced and smart, but they also tend to be actively open-minded and willing to learn from new information.

judgments. Yet some professionals in these domains come to be called experts. The confidence we have in these experts’ judgment is entirely based on the respect they enjoy from their peers. We call them respect-experts. The term respect-expert is not meant to be disrespectful. The fact that some experts are not subject to an evaluation of the accuracy of their judgments is not a criticism; it is a fact of life in many domains. Many professors, scholars, and management consultants are respect-experts. Their credibility depends on the respect of their students, peers, or clients.

Part of the answer is the existence of shared norms, or professional doctrine. Experts often obtain professional qualifications from professional communities and receive training and supervision in their organizations. Doctors who complete their residency and young lawyers who learn from a senior partner do not just learn the technical tools of their trade; they are trained to use certain methods and follow certain norms. Shared norms give professionals a sense of which inputs should be taken into account and how to make and justify their final judgments.

Respect-experts excel at constructing coherent stories. Their experience enables them to recognize patterns, to reason by analogy with previous cases, and to form and confirm hypotheses quickly. They easily fit the facts they see into a coherent story that inspires confidence.

Many debates and misunderstandings arise in discussions of measures of intelligence or of general mental ability (GMA, the term now used in preference to intelligence quotient, or IQ).

The message that emerges from this mass of research is unambiguous. As one review put it, “GMA predicts both occupational level attained and performance within one’s chosen occupation and does so better than any other ability, trait, or disposition and better than job experience.”

Psychologists and neuroscientists distinguish between crystallized intelligence, the ability to solve problems by relying on a store of knowledge about the world (including arithmetical operations), and fluid intelligence, the ability to solve novel problems.

Yet for all its crudeness and limitations, GMA, as measured by standardized tests containing questions on verbal, quantitative, and spatial problems, remains by far the best single predictor of important outcomes.

therefore, high mental ability is apparently a necessary condition for gaining access to high-status professions.

However, this measure fails to capture differences in achievement within these groups. Even among the top 1% of people as measured by cognitive ability (evaluated at age thirteen), exceptional outcomes are strongly correlated with GMA. Compared with those who are in the bottom quartile of this top 1%, those who are in the top quartile are two to three times more likely to earn a doctoral-level degree, publish a book, or be granted a patent. In other words, not only does the difference in GMA matter between the 99th percentile and the 80th or 50th, but it still matters—a lot!—between the 99.88th percentile and the 99.13th.

The conclusion is clear. GMA contributes significantly to the quality of performance in occupations that require judgment, even within a pool of high-ability individuals. The notion that there is a threshold beyond which GMA ceases to make a difference is not supported by the evidence. This conclusion in turn strongly suggests that if professional judgments are unverifiable but assumed to reach for an invisible bull’s-eye, then the judgments of high-ability people are more likely to be close. If you must pick people to make judgments, picking those

People of high mental ability are more likely than others to make better judgments and to be true experts, but they are also more likely to impress their peers, earn others’ trust, and become respect-experts in the absence of any reality feedback.

The only measure of cognitive style or personality that they found to predict forecasting performance was another scale, developed by psychology professor Jonathan Baron to measure “actively open-minded thinking.” To be actively open-minded is to actively search for information that contradicts your preexisting hypotheses. Such information includes the dissenting opinions of others and the careful weighing of new evidence against old beliefs. Actively open-minded people agree with statements like this: “Allowing oneself to be convinced by an opposing argument is a sign of good character.” They disagree with the proposition that “changing your mind is a sign of weakness” or that “intuition is the best guide in making decisions.”

review. First, it is wise to recognize the difference between domains in which expertise can be confirmed by comparison with true values (such as weather forecasting) and domains that are the province of respect-experts. A political analyst may sound articulate and convincing, and a chess grandmaster may sound timid and unable to explain the reasoning behind some of his moves. Yet we probably should treat the professional judgment of the former with more skepticism than that of the latter. Second, some judges are going to be better than their equally qualified and experienced peers. If they are better, they are less likely to be biased or noisy. Among many things that explain these differences, intelligence and cognitive style matter. Although no single measure or scale unambiguously predicts judgment quality, you may want to look for the sort of people who actively search for new information that could contradict their prior beliefs, who are methodical in integrating that information into their current perspective, and who are willing, even eager, to change their minds as a result.

Speaking of Better Judges “You are an expert. But are your judgments verifiable, or are you a respect-expert?” “We have to choose between two opinions, and we know nothing about these individuals’ expertise and track record. Let’s follow the advice of the more intelligent one.” “Intelligence is only part of the story, however. How people think is also important. Perhaps we should pick the most thoughtful, open-minded person, rather than the smartest one.”

Ex post, or corrective, debiasing is often carried out intuitively.

This kind of bias correction is sometimes undertaken more systematically. In the United Kingdom, HM Treasury has published The Green Book, a guide on how to evaluate programs and projects. The book urges planners to address optimistic biases by applying percentage adjustments to estimates of the cost and duration of a project. These adjustments should ideally be based on an organization’s historic levels of optimism bias. If no such historical data is available, The Green Book recommends applying generic adjustment percentages for each type of project.

Ex ante or preventive debiasing interventions fall in turn into two broad categories. Some of the most promising are designed to modify the environment in which the judgment or decision takes place.

Other nudges work on different aspects of choice architecture. They might make the right decision the easy decision—for example, by reducing administrative burdens for getting access to care for mental health problems.

A different type of ex ante debiasing involves training decision makers to recognize their biases and to overcome them. Some of these interventions have been called boosting; they aim to improve

Whether they correct biases ex post or prevent their effects through nudging or boosting, most debiasing approaches have one thing in common: they target a specific bias, which they assume is present. This often-reasonable assumption is sometimes wrong.

The upshot is that ex post or ex ante debiasing—which, respectively, correct or prevent specific psychological biases—are useful in some situations. These approaches work where the general direction of error is known and manifests itself as a clear statistical bias. Types of decisions that are expected to be strongly biased are likely to benefit from debiasing interventions. For instance, the planning fallacy is a sufficiently robust finding to warrant debiasing interventions against overconfident planning.

The problem is that in many situations, the likely direction of error is not known in advance. Such situations include all those in which the effect of psychological biases is variable among judges and essentially unpredictable—resulting in system noise. To reduce error under circumstances like these, we need to cast a broader net to try to detect more than one psychological bias at a time.

We suggest undertaking this search for biases neither before nor after the decision is made, but in real time. Of course, people are rarely aware of their own biases when they are being misled by them. This lack of awareness is itself a known bias, the bias blind spot. People often recognize biases more easily in others than they do in themselves. We suggest that observers can be trained to spot, in real time, the diagnostic signs that one or several familiar biases are affecting someone else’s decisions or recommendations.

The case for relying on a checklist is clear: checklists have a long history of improving decisions in high-stakes contexts and are particularly well suited to preventing the repetition of past errors.

Here is an example. In the United States, federal agencies must compile a formal regulatory impact analysis before they issue expensive regulations designed to clean the air or water, reduce deaths in the workplace, increase food safety, respond to public health crises, reduce greenhouse gas emissions, or increase homeland security. A dense, technical document with an unlovely name (OMB Circular A-4) and spanning nearly fifty pages sets out the requirements of the analysis. The requirements are clearly designed to counteract bias. Agencies must explain why the regulation is needed, consider both more and less stringent alternatives, consider both costs and benefits, present the information in an unbiased manner, and discount the future appropriately. But in many agencies, government officials have not complied with the requirements of that dense, technical document. (They might not even have read it.) In response, federal officials produced a simple checklist, consisting of just one and one-half pages, to reduce the risk that agencies will ignore, or fail to attend to, any of the major requirements.

Decision observation with appropriate bias checklists can help limit the effect of biases. Although we have seen some encouraging results in informal, small-scale efforts, we are not aware of any systematic exploration of the effects of this approach or of the pros and cons of the various possible ways to deploy it. We hope to inspire more experimentation, both by practitioners and by researchers, of the practice of real-time debiasing by decision observers.

Noise, on the other hand, is unpredictable error that we cannot easily see or explain. That is why we so often neglect it—even when it causes grave damage. For this reason, strategies for noise reduction are to debiasing what preventive hygiene measures are to medical treatment: the goal is to prevent an unspecified range of potential errors before they occur. We call this approach to noise reduction decision hygiene.

pandemic). Similarly, following the principles of decision hygiene means that you adopt techniques that reduce noise without ever knowing which underlying errors you are helping to avoid.

The analogy with handwashing is intentional. Hygiene measures can be tedious.

When provided with a latent print, examiners routinely follow a process called ACE-V, which stands for analysis, comparison, evaluation, and verification. First, they must analyze the latent print to determine whether it is of sufficient value for comparison. If it is, they compare it to an exemplar print. The comparison leads to an evaluation, which can produce an identification (the prints originated from the same person), an exclusion (the prints do not originate from the same person), or an inconclusive decision. An identification decision triggers the fourth step: verification by another examiner.

Itiel Dror, a cognitive neuroscience researcher at University College London, was the first to set out to study. He conducted what amounts to a series of noise audits in a field that had assumed it did not have a noise problem.