The Upside of Irrationality: The Unexpected Benefits of Defying Logic at Work and at Home

by Dan Ariely

The fact that only 45 percent of people returned the extra cash even when they were not annoyed is a sad state of affairs, to be sure.

The results were depressing. As we had discovered in our first experiment, people who were annoyed by the phone call were much less likely to return the extra cash than those whose conversations were uninterrupted. More surprisingly, we found that the tendency to seek revenge did not depend on whether Daniel (the agent) or I (the principal) suffered. This reminded us of Tom Farmer and Shane Atchison. In their case too they were annoyed mostly with Mike, the night clerk (an agent), but their PowerPoint presentation was aimed mostly at the Doubletree Club hotel (the principal). It seems that at the moment we feel the desire for revenge, we don’t care whom we punish—we only want to see someone pay, regardless of whether they are the agent or the principal. Given the number of agent-principal dualities in the marketplace and the popularity of outsourcing (which further increases these dualities), we thought this was indeed a worrisome result.

The Power of Apologies

So why did I get so angry? I suspect that if the customer service representative had said, “Sorry, but sometimes cars break,” and had showed me some sympathy, the whole sequence would have played out very differently. Could it be that apologies can improve interactions and soothe the instinct for revenge in business and in personal exchanges?

Again, we sent Daniel to ask coffee shop customers if they would complete our letter-pairing task in exchange for $5. This time, however, we had three conditions. In the control (no-annoyance) condition, Daniel first asked the coffee shop patrons if they would be willing to participate in a five-minute task in return for $5. When they agreed (and almost all agreed), he gave them the same letter sheets and explained the instructions. Five minutes later, he returned to the table, collected the sheets, handed the participants four extra dollars (four $1 bills and one $5 bill), and asked them to fill out a receipt for $5. For those in the annoyance condition, the procedure was basically the same, except that while going over the instructions, Daniel again pretended to take a call.

The third group was basically in the same condition as those in the annoyance group, but we threw in a little twist. This time, as Daniel was handing the participants their payment and asking them to sign the receipt, he added an apology. “I’m sorry,” he said, “I shouldn’t have answered that call.”

But what about the third group? Surprise!—the apology was a perfect remedy.

The amount of extra cash returned in the apology condition was the same as it was when people were not annoyed at all.

Indeed, we found that the word “sorry” completely counteracted the...

when researchers at the Johns Hopkins Bloomberg School of Public Health in Baltimore11 showed participants videos of doctors responding to medical errors, participants rated the doctors who expressed an apology and who took personal responsibility far more favorably. What’s more, another research team from the University of Massachusetts Medical School found that people expressed less interest in suing doctors who had assumed responsibility, apologized, and planned a means of avoiding the error in the future.12

So we decided on a different approach. We took the medical files of our forty participants and showed them to a doctor, two nurses, and a physical therapist at the rehabilitation hospital where Hanan and I had spent so much time. We asked them to split the sample into two groups, the mildly injured and the more grievously injured. Once that was done, Hanan and I had two groups that were relatively similar to each other in many respects (all participants had been in the army, injured, and hospitalized, and were part of the same veterans country club) but differed in the severity of their injuries. By comparing these two groups, we hoped to see if the severity of our participants’ injuries influenced the way they experienced pain many years later. The severely injured group was made up of people like Noam, whose army job had been to disassemble land mines. At some unfortunate point, a land mine had exploded in his hands, piercing his body with numerous shrapnel wounds and costing him a leg and the sight in one eye. In the mildly injured group were men like Yehuda, who had broken his elbow while on duty. He had undergone an operation that had involved restoring the joint by adding a titanium plate, but he was otherwise in good health. The participants who had been mildly injured reported that the hot water became painful (pain threshold) after about 4.5 seconds, while those who had been severely injured started feeling pain after 10 seconds. More interestingly, those in the m...

Just as our eyes adjust to changes in light and environment, we can adapt to changes in expectation and experience. For example, Andrew Clark showed that job satisfaction among British workers was strongly correlated with changes in workers’ pay rather than the level of pay itself. In other words, people generally grow accustomed to their current pay level, however low or high. A raise is great and a pay cut is very upsetting, regardless of the actual amount of the base salary. In one of the earliest studies on hedonic adaptation, Philip Brickman, Dan Coates, and Ronnie Janoff-Bulman compared the overall life happiness among three groups: paraplegics, lottery winners, and normal people who were neither disabled nor particularly lucky. Had the data collection taken place immediately following the event that led to the disability or the day after the lottery win, one would expect the paraplegics to be far more miserable than the normal people and the lottery winners much happier. However, the data were collected a year after the event. It turned out that although there were differences in happiness levels among the groups, they were not as pronounced as you might expect. While the paraplegics were not as satisfied with life as the normal people and the lottery winners were more satisfied, both paraplegics and lottery winners were surprisingly close to

normal levels of life satisfaction. In other words, though a life-altering event such as a bad injury or winning a lottery can have a huge initial impact on happiness, this effect can, to a large degree, wear off over time.

A SUBSTANTIAL AMOUNT of research over the past decade has reinforced the idea that although internal happiness can deviate from its “resting state” in reaction to life events, it usually returns toward its baseline over time. Though we don’t hedonically adapt to every new situation, we do adapt to many of them, and to a large degree—whether we’re getting...

When Romeo suffered over his breakup with his first girlfriend, Rosaline, you would have thought it was the end of the world. He stayed up all night and shut himself in his room. His parents were worried. When his cousin asked how he was faring, Romeo sounded as if he would die of hopeless love for the girl who’d rejected

him. “She hath forsworn to love,” he complained, “and in that vow / Do I live dead that live to tell it now.” That night he met Juliet and forgot all about Rosaline.

a new car feels wonderful, but sadly, the feeling lasts for only a few months. We get used to driving the car, and the buzz wears off. So we look for something else to make us happy: maybe new sunglasses, a computer, or another new car. This cycle, which is what drives us to keep up with the Joneses, is also known as the hedonic treadmill. We look forward to the things that will make us happy, but we don’t realize how short-lived this happiness will be, and when adaptation hits we look for the next new thing.

that a Chicagoan moving to sunny California will see a dramatic

Given that hedonic adaptation is clearly a mixed bag, how, you might wonder, can we use our understanding of it to get more out of life? When adaptation works in our favor (such as when we get used to living with an injury), we should clearly let this process take place. But what about instances when we wish not to adapt? Can we somehow extend the euphoric feeling of a new car, city, relationship, and so on?

One key to changing the adaptation process is to interrupt it.

In a set of experiments, they measured how small interruptions—which they called hedonic disruptions—influence the overall enjoyment and irritation we get from pleasurable and painful experiences. In essence, they wanted to see if taking breaks in the middle of p...

experience would make...

Leif and Tom found that, in general, when asked about their preferences for breaking up experiences, people want to disrupt annoying experiences but prefer to enjoy pleasurable experiences without any breaks. But following the basic principles of adaptation, Leif and Tom suspected that people’s intuitions are completely wrong. People will suffer less when they do not disrupt annoying experiences, and enjoy pleasurable experiences more when they break them up.

After listening to the sound, the participants evaluated their irritation levels during the last five seconds of the experience. Leif and Tom found that the most pampered participants—those who had endured only five seconds of sound—were far more irritated than those who listened to the annoying sound for much longer. As you may have guessed, this result suggests that those who suffered through the vacuum whooom for forty seconds got used to it and found the last five seconds of their experience to be not so bad. But what happened to those who experienced the short break? As it turned out, the interruption made things worse. The adaptation went away, and the annoyance returned.

You may think that taking a break during an irritating or boring experience will be good for you, but a break actually decreases your ability to adapt, making the experience seem worse when you have to return to it. When cleaning your house or doing your taxes, the trick is to stick with it until you are done.

And what about pleasurable experiences?

The first group received an uninterrupted three-minute treatment. The second group received a massage for eighty seconds, followed by a twenty-second break, after which the massage resumed for another eighty seconds—making their massage time two minutes and forty seconds, twenty seconds less than the uninterrupted group. At the end of the massages all the participants were asked to evaluate how much they had enjoyed the entire treatment. As it turned out, those who underwent the shorter massages with the break not only enjoyed their experiences more but they also said they would pay twice as much for the same interrupted massage in the future.

Although most people in Ann’s position would think about how nice it would be to dress up their apartment and so would go on a shopping spree, by now it should be clear that, given the human tendency for adaptation, she would actually be happier with the intermittent scenario. She can get more “happiness buying power”

out of her money if she limits her purchases, takes breaks, and slows down the adaptation process. The lesson here is to slow down pleasure. A new couch may please you for a couple of months, but don’t buy your new television until after the thrill of the couch has worn off. The opposite holds if you are struggling with economic cutbacks. When reducing consumption, you should move to a smaller apartment, give up cable television, and cut back on expensive coffee all at once—sure, the initial pain will be larger, but the total amount of agony over time will be lower.

The benefits of randomness range from the personal to the romantic to our work life. As the economist Tibor Scitovsky argued in The Joyless Economy, we have a tendency to take the safe and predictable path at work, and by extension in our personal life, and do the things that provide steady and reliable progress. But, Scitovsky argues, real progress—as well as real pleasure—comes from taking risks and trying very different things. So the next time you have to make a presentation, work with a team, or pick a project to work on, try doing something new. Your attempt at humor or cross-corporate collaboration may fail, but on balance it might make a positive difference.

On a more practical level, let’s say you want a particular laptop but decide that it’s too expensive. If you settle for a cheaper one, you’ll most likely get used to it over time. That is, unless the person in the cubicle next to you has the laptop that you originally wanted. In the latter case, the daily comparison between your laptop and your neighbor’s will slow down your adaptation and make you less happy. More generally, this principle means that when we consider the process of adaptation, we should think about the various factors in our environment and how they may influence our ability to adapt.

With this data we could, for example, see how a person who is rated as unattractive by users of HOT or NOT (let’s say a 2) rates the hotness of others, compared with someone who is rated as very attractive (let’s say a 9).

For example, if adaptation did not take place, a person who is a 2 and a person who is an 8 would both see 9s as 9s and 4s as 4s. On the other hand, if people who are aesthetically challenged have adapted by changing their perspective about the attractiveness of others, their view of hotness would differ from those of highly attractive people. For example, if adaptation had taken place, a person who is a 2 could see a 9 as a 6 and a 4 as a 7, while a person who is an 8 would see a 9 as a 9 and a 4 as a 4. The best news for us was that we could measure it! In short, by examining how one’s own attractiveness influences the hotness rating that one gives others, we thought we might discover something about the extent of adaptation. Intrigued by our project, James and Jim provided us with the ratings and dating information of 16,550 HOT or NOT members during a ten-day period. All members of the sample were heterosexual, and the majority (75 percent) were male.*

Next, we set about testing the possibility that aesthetically challenged individuals are simply unaware of the limitations placed on us by our lack of beauty (or at least, that this is how we behave online). To do this, we used a second interesting feature of HOT or NOT called “Meet Me.” Assuming you are a man who sees a picture of a woman you’d like to meet, you can click the Meet Me button above the woman’s picture. She will then receive a notification saying that you are interested in meeting, accompanied by a bit of information about you. The key is that when using the Meet Me feature, you would not be reacting to the other person only on the basis of aesthetic judgment; you would also gauge whether

Adaptation and the Art of the Speed Date The data from HOT or NOT eliminated two of our three hypotheses for the process of adaptation to one’s own physical attractiveness. One alternative remained: like my middle-aged friend, people do adapt by putting less emphasis on their partner’s looks and learn to love other attributes.

speed dating is everywhere:

It makes the quest for everlasting love feel like bargain shopping in a Turkish bazaar.

Our version of speed dating was designed to include a few special features. First, before the start of the event, we surveyed

each of the participants. We asked them to rate the importance of different criteria—physical attractiveness, intelligence, sense of humor, kindness, confidence, and extroversion—when considering a potential date.

At the end of each “date,” participants did not move immediately to the next one. Instead, we asked them to pause and record their ratings for the person they’d just met, using the same attributes (physical attractiveness, intelligence, sense of humor, kindness, confidence, and extrovers...

These measures gave us three types of data. The pre-speed-dating survey told us which attributes each person was generally looking for in a romantic partner. From the postdate responses, we discovered how they rated each person they had met on these attributes. We also knew wh...

First, we examined participants’ responses regarding their general preferences—the ones they provided before the event started. In terms of what they were looking for in a romantic partner, those who were more attractive cared more about attractiveness, while the less attractive people cared more about other characteristics (intelligence, sense of humor, and kindness). This finding was our first evidence that aesthetically challenged people reprioritize their requirements in dating. Next, we examined how each speed dater evaluated each of their partners during the event itself and how this evaluation translated to a desire to meet for a real date. Here, too, we saw the same pattern: the aesthetically challenged

people were much more interested in going on another date with those they thought had a sense of humor or some other nonphysical characteristic, while the attractive people were much more likely to want to go on a date with someone they evaluated as good-looking.

If we take the findings from the HOT or NOT, the Meet Me, and the speed-dating experiments, the data suggest that while our own level of attractiveness does not change our aesthetic tastes, it does have a large effect on our priorities. Simply put, less attr...

The results I’ve described so far were combined across males and females, and you probably suspect that men and women differ in their responses to attractiveness. Right?

Right. As it turns out, most of the gender differences in our HOT or NOT study fell into line with common stereotypes about dating and gender. Take, for instance, the commonly held belief that men are less selective in dating than women. It turns out that this is not just a stereotype: men were 240 percent more likely to send Meet Me invitations to potential females than vice versa.

men care more about the hotness of women than the other way around, which also relates to the finding that men are less concerned wi...

Incidentally, the male tendency to ask many women on dates, and to aim higher (which some may see as negative), can euphemistically be called “men’s open-mindedness in dating.”

In centuries past, a yenta, or matchmaker, performed a very important task in traditional society. A man or woman (and their parents) would tell the yenta to “find me a find, catch me a catch,” as the song in Fiddler on the Roof put it. To narrow the playing field for her clients, the yenta made sure she knew everything possible about the young people and their families (which is why the word “yenta” eventually became synonymous with “gossip” or “blabbermouth”). Once she found a few likely fits, she introduced the prospective husbands and wives and their families to each other. The yenta ran an efficient, viable business, and she was paid for her services as a matchmaker (or “market maker” in economics-speak) who brought people together.

Essentially, markets are coordination mechanisms that allow people to save time while achieving their goals.