How Promises Work

Living up to a social contract is inordinately valuable, and there's no pressure to exceed it.


Nicholas Epley recently made news when he paid commuters $5 to talk to a stranger on a Chicago train. The people were happier for having done it. But the point of his experiment was to shatter expectations. Most people presumed they would be happier sitting alone than talking to a stranger; but they were not. They also expected, on average, that fewer than half of their fellow passengers would be willing to talk with them. When people actually tried talking, though, no one was rebuffed.

Epley is a broad-shouldered professor of behavioral science at the University of Chicago Booth School of Business. He studies prosocial behavior in ways fascinating and a little unnervingly manipulative, of necessity. In a recent lecture he recounted that people tend to evaluate one another in two general dimensions: how interpersonally warm we seem to be, and how competent we seem to be. His latest work suggests that the way to deliver on both without going overboard on effort is to make promises.

In the journal Social Psychological and Personality Science, Epley and Ayelet Gneezy at University of California at San Diego perform three experiments that explain why promises can be helpful in business, and in limiting the burden of expectations in a world where enough is never enough and more is more. I spoke with Epley recently by phone about broken promises, and the importance of trust.

"You're late," he told me upon answering, exactly at 10:00 A.M.

That was the time I had on my calendar. I panicked, confused. "I'm so sorry."

Luckily it was just a joke, he explained, based on the nature of our interview. The ice broke all around us, and we sank into the frigid depths of pop psychology.

In Epley's latest social experiment, people were asked to consider a hypothetical scenario in which a friend had promised to give feedback on a paper. Either the friend did just as promised, gave really exceptional comments, or did less than they promised. "People were no more positive when someone did more than they said they would," Epley explained. "Breaking a promise seemed to hurt, but exceeding a promise didn't seem to help."

The same thing was true for remembered promises. The researchers asked people to recall promises kept, broken, and exceeded. People were less happy when someone broke a promise than when they kept it, but, again, exceeding the promise didn't create a big gain. At least, not one that was of the same magnitude as the damage of breaking a promise. "Breaking a promise was bad," Epley essentially reiterated. "Exceeding a promise? Meh. Not that much better than just keeping it."

Then in the pièce de résistance, part three, researchers made real promises to subjects in a lab, then broke those promises and evaluated the subjects' reactions.

"For a scientist, this was the most compelling demonstration, by far," Epley said, a staid appraisal of the thrill of a calculated lie for the greater good. To the researchers' knowledge, everything in prior psychology research had involved imagined or recalled promises. No one had actually created a laboratory experiment that manipulated promise keeping. "These promises may not have the ecological validity of a promise made out in the world, but they have the benefit of really happening, right there."

The experimenters gave subjects a set of 40 puzzles and told the subjects they would be paid on the basis of how many they completed. Then other people promised to help out by doing a certain number of the puzzles, pro bono. Some over and under-delivered on their promises. People were upset when the person didn't make good on their promise, but they were barely happier when the person over-delivered than when they did just what they said they would do.

I was surprised by how unimpressed people seem to be by that kind of unexpected boon. But Epley was surprised that I was surprised. He explained that there are actually a lot of findings of this kind in the psychological literature.

To make the point, Epley told me to think about an oil spill that kills a certain number of birds. I did, reluctantly. "How much should the oil company be fined?" he asked, rhetorically. "Well, if they killed 20,000 birds, they should obviously be fined more than if they killed only 2,000 birds. That's ten times more bird carnage." It's true; I did the math. But, he explained, if you ask people in experiments how much the oil company should be fined, and the people only see one of those numbers, the estimates are about the same.

"People are insensitive to the scope of the bird carnage when they evaluate only one number at a time. When you see that an oil company killed 2,000 birds, you don't think spontaneously, 'Oh, that's only one tenth of 20,000.'" It's just more than zero. "Everything else is just kind of gravy."

Distance operates like that, too. Think of the distance from your home to Saturn, and then think of the distance from your home to the next solar system. As Epley put it, "At some point, things just become 'far'—too big for us too handle." Scope insensitivity is a common phenomenon in psychology, and Epley believes that's what's happening here.

But, I asked, what about an employee who over-performs at a job? Doesn't that score more points with a boss than someone who just punches the clock?

"If I'm your boss and I expect you only to work Friday, and you come in Friday and Saturday, then I would evaluate you more favorably," he conceded. But that called up the critical difference between a promise and an expectation. His results are really about promises as interpersonal contracts. Expectations are in only one person's head, and they follow a relatively linear pattern. Less is worse, more is better. But promises have a peak at the level of fairness that is above and beyond the expectation point. That's called the fairness premium.

"If I think you're going to come in on Saturday, and you do, I'll feel fine about it. It's a little more than what you normally do. But if you, as an employee, promise your boss that you'll come in on Saturday and then you do come in on Saturday, I will actually be happier than if I just expected it. I don't get that fairness premium, that comes with a promise kept—that boost from recognizing you as a reliable and trustworthy person."

In that way, a promise is an action with moral implications. Keeping a promise is prosocial. "I've done what I said I would do," Epley explained. "Breaking a promise is kind of a selfish thing."

Does this pattern show up more generally across prosocial actions? "We're happy when people are kind to others, when they share, when they are good to us, and we're not happy when they are selfish," Epley explained. "Mother Theresa is lauded, Bernie Madoff is scorned." But the spectrum from purely selfish to purely prosocial behavior is huge. He posits this situation:

Imagine you're a kid with a cookie and a friend who has no cookie. What happens if you eat it all? Your friend will be upset. What happens if you give all of it away? Your friend will like you a lot. What if you give away half the cookie? Your friend will be just about as happy with you as if you gave him the whole thing. His satisfaction is a pretty flat line if you give anything more than half of the cookie. People judge actions that are on the selfish side of fairness. Maybe because we denigrate do-gooders, or because we're skeptical of too much selflessness, the research shows that, as Epley put it, "It just doesn't get any better than giving half of the cookie."

In one of his other recent experiments, people were asked to give donations at an orchestra concert. People who gave the "suggested donation" of $10 were judged by observers no more highly than people who gave twice as much. And in a scenario where there was no suggested donation, $5 was significantly more favorable than nothing, but $50 was no better than $5. That one is worth remembering if you're going for that benevolent-on-a-budget look this fall.

Before we parted, Epley said he would send me some of his other research papers. If I didn't hear from him by 3:00 P.M., I should email him. I asked if that was another experiment. He said no.

So, should we all make more promises? That way we can give people that fairness premium, gain trust, and always be sure that expectations of us are met? Knowing exactly what is expected of me, doing exactly that, and being lauded for it seems like some kind of dream.

"Well, promises are a risky strategy," he said, unexpectedly distancing himself. "Make the promises you can keep, but you don't need to exceed them. Just do what you say you'll do."

I got the email well before 3:00.

(Image via 2jenn/Shutterstock.com)