AS IMPEACHMENT LOOMED FOR the president of the United States in 1998, a reporter called me to ask if we could trust what President Clinton did when he was working at his desk in the Oval Office, now that we knew what he did under it. Other reporters were less direct, but they had the same concern. Their questions reflected the common belief that qualities like self-control, conscientiousness, and trustworthiness are broad traits that characterize a person’s behavior not only stably over time but also consistently in many different kinds of situations: it assumes that a person who lies and cheats in one kind of situation is also likely to be dishonest in many other situations, whereas one who is conscientious will be predictably conscientious in diverse contexts. These expectations are violated each time the headlines announce the fall of another famous person in a position of public trust who turns out to have a secret life, exposing a side of his personality that appears to be the opposite of his public self. Predictably, a torrent of speculations follows that always raise the question: “Who is he really?”
President Clinton’s pattern was hardly unique. One of the most stunning examples of such an inconsistency in behavior was the fall of Sol Wachtler from his position as chief judge of the State of New York and the New York Court of Appeals to incarceration as a felon in federal prison. Judge Wachtler was famously revered for advocating laws to make marital rape a punishable crime, and was deeply respected for his landmark decisions on free speech, civil rights, and the right to die. After his mistress left him, however, the judge reportedly spent months harassing her, writing obscene letters, making lewd phone calls, and threatening to kidnap her daughter. How did this model of jurisprudence and moral wisdom turn into a handcuffed prisoner on his way to jail? Judge Wachtler attributed his own behavior to his problems with an uncontrollable romantic obsession. One expert asked about Wachtler suggested that he might have a brain tumor the size of a baseball. He didn’t.
The headlines announce similar stories about celebrities and public figures in the entertainment world, religious institutions and pulpits, business, sports, and academia—no area is exempt. Tiger Woods, the golf star hero, was the personification and ideal of rigorous self-discipline, not just in mastering his physical skills but in his sensational capacity to focus his attention. He was a supposedly happily married man, but he ultimately confessed to a private life with mistresses that violated his well-cultivated public image. The sports idol suffered one of the more memorable instant falls from grace, or at least from public celebration—for a while. In time, his descent was followed by that of the world champion marathon cyclist Lance Armstrong, whose career and extraordinary life were tainted by a doping scandal.
“How do you explain these folks?” reporters invariably ask.
They want a short answer for their deadline. I give them the shortest version: President Clinton had the self-control and delay ability to win a Rhodes scholarship, attain a Yale law degree, and be elected to the U.S. presidency, apparently combined with little desire—perhaps no ability, and certainly no willingness—to exert self-control for particular temptations like junk food and attractive White House interns. Likewise, the judge and the golf star had the self-control skills to excel in the pursuit of their most important career goals, but not in other contexts. To be able to delay gratification and exert self-control is an ability, a set of cognitive skills, that, like any ability, can be used or not used depending primarily on the motivation to use it. Delay ability can help preschoolers resist one marshmallow now to get two later, but they have to want to do that.
Whether or not self-control skills are used depends on a host of considerations, but how we perceive the situation and the probable consequences, our motivation and goals, and the intensity of the temptation, are especially important. This may seem obvious, but I emphasize it here because it is easily misunderstood. Willpower has been mischaracterized as something other than a “skill” because it is not always exercised consistently over time. But like all skills, self-control skill is exercised only when we are motivated to use it. The skill is stable, but if the motivation changes, so does the behavior.
Many celebrities and public figures exposed in the headlines probably did not want to resist their temptations. On the contrary, they often seemed to expend considerable effort seeking and pursuing them. Their optimistic illusions and inflated self-worth, shared with the rest of humanity but perhaps even more grandiose in them, made them feel invulnerable. They did not expect to be caught even if they had been in the past. They also believed that if they were discovered, they could still get away with it—which is not an unreasonable expectation for some, given their past experiences. Their histories of success and power may also encourage them to spin entitlement theories that exempt them from the usual rules and encourage them to do what less powerful people can’t. As Leona Helmsley, New York’s billionaire ex–hotel queen, was alleged to have said before beginning her prison term, “Only the little people pay taxes.” If they don’t make remarks like that, their prospects for redemption usually remain excellent even if they are exposed. Modern-day fallen heroes often rise, phoenix-like, from the newspaper ashes that announced their downfall to host television shows, run news and interview programs, or become well-compensated consultants.
The ability to exert self-control and wait for marshmallows implies neither that it will be exercised in every domain and context nor that it will be used for virtuous goals. People can have excellent self-control skills that they use creatively for good purposes valued by society. They can also use the same skills to create hidden families, offshore bank accounts, and secret lives. They can be responsible, conscientious, and trustworthy in some areas of their lives and not in others. If we look closely at what people really do, not at what they say, across different situations with regard to any dimension of social behavior, it turns out that they are not very consistent.
When we look at the people we know, nothing is more obvious than that they differ greatly in their social behavior and characteristics, on whatever dimension we consider. In general, some are much more conscientious, sociable, friendly, aggressive, quarrelsome, extraverted, or neurotic than others. We make these judgments easily, and we mostly agree not only with one another but also with the self-perceptions of the people we are judging. These widely shared impressions of what we are like are very useful, indeed vital, for navigating the social world, and they allow us to make reasonable predictions about what to expect from other people.
Situations also influence social behavior in an important way, depending on how they are perceived. Regardless of how conscientious a person tends to be or not be, most will be more conscientious about being on time when picking up the kids from preschool than when meeting a friend for coffee, and they will be more sociable and extraverted at large parties than at funerals. That kind of variability is evident.
The conception of human traits, however, makes an additional assumption—namely that an individual will be consistent in the expression of a trait across many different kinds of situations in which the trait is desirable. It is assumed that a highly conscientious person will be more conscientious than a less conscientious person consistently across many different kinds of situations. If Johnny is judged to be much more conscientious than Danny “on the whole,” then he should also be ranked higher than Danny on his completion of homework assignments and on his attendance record, as well as on how conscientiously he keeps his room organized at home and how trustworthy he is when he babysits his younger sister. Is this assumption justified? Does the person high in any important psychological trait generally remain above the person low in that trait across many different kinds of situations?
The assumption that people are broadly consistent in what they do, think, and feel in very different situations is intuitively powerful. It is fed by the hot system, which quickly forms impressions from the smallest slices of behavior and generalizes them to anything that can more or less fit. But does it hold up when we use the benefit of the prefrontal cortex to look closely at what people really do across different situations—whether it is President Clinton, our family and friends, or ourselves?
When I was preparing to teach my first course on assessment as a new lecturer at Harvard, I started asking these questions: Can you use your work partner’s conscientiousness at the office to predict how conscientious she will be at home? Can I predict how my colleague—who is known as “the loose cannon” at department meetings—behaves at home with his children? To my own surprise, study after rigorous study failed to support the core trait assumption: people high in a trait in one kind of situation often were low in that trait in another type of situation. The aggressive child at home may be less aggressive than most when in school; the woman exceptionally hostile when rejected in love may be unusually tolerant about criticism of her work; the patient who sweats with anxiety in the dentist’s office may be cool and courageous when scaling a sheer face of a mountain; and the high-flying business entrepreneur may avidly avoid taking social risks.
In 1968, I undertook a comprehensive review of the correlations that had been obtained in dozens of studies trying to link people’s behavior in one situation (such as conscientiousness about meeting obligations and commitments at the office) to their behavior in another (such as conscientiousness at home). The findings shocked many psychologists. They revealed that although generally the correlations were not zero, they were much lower than had been assumed. The researchers who failed to demonstrate the consistency of behavior across different situations blamed their failure on having used imperfect and insufficiently reliable methods. I began to wonder whether the problem might be that their assumptions about the nature and consistency of human characteristics were wrong.
While the debate continued, it did not change the fact that across-the-board consistency of a person’s behavior is generally too weak to be useful if the goal is to accurately predict from his behavior in one type of situation what he will do in another type of situation. Behavior is context-dependent. Highly developed self-control skills may be exercised in some situations and with some temptations, but not in others, as the stories of fallen public figures regularly remind us.
This raises problems in everyday life, and those problems became vivid for me when I needed to hire someone to take care of my young children for two weeks while I had to be abroad. I considered my neighbors’ babysitter, “Cindy.” She said she got good grades in high school, had worked as a lifeguard the previous summer, and did not smoke. She seemed to be a nice kid, and the neighbors agreed. But I also knew, as I described above, that we usually cannot predict accurately from one situation to other, very different situations. For example, how would Cindy behave at parties with her peers when the alcohol was passed around? We also couldn’t predict, based on how she behaved babysitting one particular evening, how she would behave when babysitting my kids for two weeks straight. Yet that is how automatic impressions are formed. We compress bits of information into a compelling simplification, a stereotype that makes us feel that what is true in one situation is also true across other situations. Even highly confident, well-trained experts who often get it right also often get it wrong—especially when they try to predict specific behavior in different new situations.
I did not hire Cindy—she seemed too young—but I did hire a young couple who appeared mature and responsible. They made a very good impression on me during a long interview and visit to meet my children, who liked them. When I returned from the trip, however, I found the house transformed into a major mess, with ten days of unwashed everything waiting. The kids survived, but they were very unhappy and had developed an intense dislike for the couple—who had developed an equally strong dislike for them. My interest in doing more research on the consistency or inconsistency of behavior, particularly self-control and conscientiousness, intensified.
In time, my research team and I did find consistency, but not where it had been expected. We found it by looking closely at what different individuals did when we unobtrusively observed them, hour after hour, day after day, over the course of half a summer in a residential treatment camp for children. It was a natural laboratory for seeing in fine-grain detail how a person’s behavior is expressed over time and across situations in everyday life, and it yielded some surprises that changed the understanding of personality. That story begins at Wediko.