The Science of Self-Control

March 7th, 2012 by David Berreby

A new book parses the implications of willpower.

Brought to you by Liberty Mutual's
The Responsibility Project

A marshmallow. A four-year-old. A dilemma.

In the late 1960s, in a famous experiment, the Stanford psychologist Walter Mischel left children individually alone in a room with a marshmallow on a table and the following instructions: they could have two marshmallows if they waited for the researcher to come back, but only the one if they ate the marshmallow right away. Some kids gave in and grabbed the treat; others, though, managed to hang on for a full 15 minutes to win the bigger reward.

As the years passed, Mischel followed up with those children, and he found that those who had withstood temptation led remarkably different lives from the others. "The children who had managed to hold out the entire 15 minutes went on to score 210 points higher on the SAT,"
Roy F. Baumeister and John Tierney recount in Willpower: Rediscovering the Greatest Human Strength (Penguin, 304 pages, $27.95). They also earned higher salaries, were more popular with peers and teachers, had a lower body-mass index, and were less likely to abuse drugs.

Many years ago, our Victorian forebears talked incessantly of duty, self-sacrifice, and "character building." The message of this lively and eccentric book is that they were right. Describing scores of experiments, leavened with tales of superhuman self-control and discipline (the magician and endurance artist David Blaine, for instance, recalls how he got through hours encased in ice and other stunts), Baumeister and Tierney argue that self-control confers immense, lifelong benefits. The best and latest research, they argue, supports "a simple rule: The best way to reduce stress in your life is to stop screwing up." And the way to stop screwing up is to get ahold of ourselves: Focus on what's important, keep a lid on our emotions, check impulses and stick to the task at hand, so we can meet deadlines, keep promises and, in general, be as responsible as we hope to be. Self-control is good for us, they write. Society will be better off, and each of us will be happier if there is "a lasting revival of this virtue."

Were the successful kids in Mischel's experiment just lucky to have stronger wills? Not at all, write the authors. Willpower, that ability to force yourself to do (or not do) things, can be cultivated and trained, like a muscle. College students in one experiment, asked to avoid slouching and watch their posture for two weeks, scored better on a measure Baumeister often uses to gauge willpower: How long they can squeeze a handgrip before giving up. "By overriding their habit of slouching," Baumeister and Tierney write, "the students strengthened their willpower and did better at tasks that had nothing to do with posture." It is, they say, empirical evidence for the old idea of "character building."

But "character" isn't synonymous with willpower, the authors note. That's because even the most powerful will – like that of Blaine, who went without food for 44 days – is limited. Baumeister believes that willpower is like a muscle in this sense as well: It can be fatigued through use. The children in Mischel's experiment who stared hard at the marshmallow all failed to hold out. It was the ones who distracted themselves – for instance, by turning away from the tempting sight – that succeeded.

Baumeister and Tierney describe willpower as a limited stock that can be used up; they mean the model quite literally. When you engage in an act of self-control, they write, a brain region most active during that task, the anterior cingulate cortex, works harder, uses more glucose, and eventually slows down. In one experiment, subjects were asked to do impossible puzzles as a measure of their perseverance; those who first drank sweetened lemonade lasted longer than those who’d been given a sugarless drink. "No glucose, no willpower," the authors write. "The pattern showed up time and again as researchers tested more people in more situations."
This willpower-as-fuel model has some odd implications. It suggests that diabetics, whose blood-sugar levels rise and fall outside the normal range, will have more self-control issues than other people. (And in fact they do, the authors write.) It also suggests that the simple fact of having eaten can affect how conscientious you are. One study of Israeli parole decisions found that prisoners who appeared right before the judges' midmorning snack had only about a 15 percent chance of gaining parole; those who came up after the judge’s food break had a 65 percent chance.

Given that willpower needs fuel and can be tired out by use, it can't serve as the bedrock of self-control. “Willpower is humans’ greatest strength, but the best strategy is not to rely on it in all situations,” the authors write. “Save it for emergencies." In fact, they add, people with good self-control are not those who constantly use their strong wills to face down temptation. Rather, they're people who have set up their lives so that their precious willpower is expended less often. In several studies cited by Baumeister and Tierney, "People with strong self-control spent less time resisting desires than other people did."

So the more self-management you can off-load onto habitual, nearly thought-free activities, the more willpower you conserve for what's really important. For this reason, Baumeister and Tierney recommend keeping a tidy desk. Their habit-building theory informs the practical tips they offer for changing bad habits, which stress clear goals (so you can define success and failure) and self-monitoring (and, even better, self-monitoring that involves other people, so you can't cheat the system).

This counsel, of course, is familiar from countless other advice books. The Willpower's claim to originality rests instead on its grand unified theory of self-control – and there it proves less convincing than it is fun to read. Though each piece of the model is presented with plenty of supporting evidence and authorial brio, the parts seem to pull against one another, leaving an overall idea that's so loose it could justify most any statement. Suppose, for instance, that I want to stop typing right now and eat an apple. Is this a good move, because the jolt of glucose will refuel my willpower to finish this review? Or a bad move, because I’m giving in to an impulse? If I force myself to stop using certain robust Anglo-Saxon words, am I training my willpower or depleting it?

The trouble here is not that the book raises these questions. It's that it seems to justify both yes and no answers to them. It's as if, in their eagerness to make the ideas cohere, Baumeister and Tierney haven't let themselves be rigorous about their consistency. Self-control is a knotty empirical and philosophical subject – who is controlling whom, after all? And the lab experiments that psychologists devise to plumb real-life self-discipline (like making people work on impossible puzzles, or measuring how long they can keep their hands in ice water) may not capture the real experience. Often what these techniques actually measure is the subjects’ obedience to the rather strange requests of psychologists, which may be a different creature than self-discipline or responsibility.

In a set of famous experiments in the 1960s, the Yale psychologist Stanley Milgram convinced his volunteers to give torturous electric shocks to a stranger, simply because the experimenter gently told them to do as instructed. Those shocks weren't real, but their lesson lingers: Suppressing one's feelings, quelling one's doubts, getting on with the task and doing what you signed up to do are not always the most responsible way to act.

David Berreby blogs writes the Mind Matters blog at and has written about science for The New Yorker, The New York Times Magazine, and many other publications. He is the author of Us and Them: The Science of Identity, published by the University of Chicago Press.