Predictably Irrational

by Dan Ariely

  We set up our candy booth in the afternoon hours when the stream of students was more or less steady, and from time to time, we switched the conditions by alternating the free and the 1¢ signs (the penny price represented what we called the “monetary condition”). We counted the number of students who stopped by our booth and how many Starbursts they either bought or picked up. We found that during an average hour in the monetary condition, about 58 students stopped by and purchased candy, while in an average hour in the free condition, 207 students stopped by to take candy. Altogether, nearly three times the number of students stopped at the booth when Starbursts were free. Just as the theory of demand predicts, the decrease in price resulted in a greater number of people consuming the product. So far so good for the first law of demand.

  Now, given the second law of demand, you’d assume that once the price drops from 1¢ to zero, each of the students who took candy would take more units. And since the number of students who stopped by was almost three times as large, you might expect that together these two forces of demand will make the total demand in our free condition much larger than the demand in the monetary condition.

  So how many more Starbursts did our students pick up

  when they were free? Trick question: They picked up fewer Starbursts!

  When the Starbursts cost a cent apiece, the average number of candies per customer was 3.5, but when the price went down to zero, the average went down to 1.1 per customer. The students limited themselves to a large degree when the candy was free. In fact, almost all the students applied a very simple social-norm rule in this situation—they politely took one and only one Starburst. This, of course, is the opposite of the second law of demand. And how did these two forces of demand work together? In total, the increase due to the greater number of people that stopped by and the decrease due to the reduction in the number of candies that each person took resulted in students collectively taking fewer Starbursts as the price decreased from 1¢ to free.

  What these results mean is that when price is not a part of the exchange, we become less selfish maximizers and start caring more about the welfare of others. We saw this demonstrated by the fact that when the price decreased to zero, customers restrained themselves and took far fewer units. So while the product (candy, in our case) was more attractive to more people, it also made people think more about others, care about them, and sacrifice their own desires for the benefit of others. As it turns out, we are caring social animals, but when the rules of the game involve money, this tendency is muted.

  THE RESULTS FROM our experiment also help explain one of the great mysteries in life: why, when we are dining out with friends, taking the last olive feels like such a big deal.

  Imagine you go to a friend’s birthday party. The appetizers are luscious: there’s a lovely spread of cheese and fruit; dishes of gherkins, kalamata olives, and tapenades; and lots of tiny little crostini. You walk around the room talking to old friends, and the wine is flowing. At some point, you wander into the kitchen and notice the delicious-looking four-decker Red Velvet cake (your personal favorite). As you chitchat with the other guests, you can’t stop thinking about that mouthwatering cake. All you really want to do is abscond with the entire thing, eat as much as you can in the laundry room without anyone knowing, and blame the dog if anyone asks. But what do you do? You balance your own desire with the desires of your friends, and you end up with only a medium-sized slice.

  Recently, I was in an analogous situation with two of my colleagues and friends, Jiwoong Shin and Nina Mazar. If you’ve ever been to a sushi restaurant with friends, you know that as the California rolls and sashimi pieces on the plate in the middle of the table start to dwindle, the people sitting around it gradually become shyer about popping them into their mouths. At the end of our meal, there was one lonely spicy tuna left, and none of us seemed to be willing to put it out of its misery. When the waitress came to bring us the check and take away the plate with the lone sushi, I asked her how often people leave a single piece at the end of the meal. “Oh,” she said, “I find one extra piece left almost every time. I think it is even more common than people finishing all their sushi.”

  Now I have eaten a lot of sushi in my life, and I can’t remember a time either when I was dining alone or when I got my own personal portion of sushi when I left anything on the plate. Somehow, when it’s just me who’s eating, I always manage to finish it all. But when the sushi is served in a large plate in the middle of the table, it just feels like taking the last one would be, well, a bit déclassé. “I really can’t eat any more,” I might say to my friends. “Go ahead. You take it.”

  What is this sushi magic? Simply put, the communal plate transforms the food into a shared resource, and once something is part of the social good, it leads us into the realm of social norms, and with that the rules for sharing with others.

  BACK TO OUR experiments, the next question we wanted to examine was whether the pattern of demand that we observed in the experiments was really due to the change from some payment to no payment. Or would it also happen when we discount prices of candies to anything above zero? According to the theory of social norms, this odd behavior of demand should manifest itself only when the price drops to zero—because only when price is not a part of an exchange do we start thinking about social consequences of our actions. Uri, Ernan, and I decided to take a closer look at this hypothesis in our next experiment.

  This experiment worked much like the previous one, except we returned to the trusty and super-delicious Lindt chocolate truffles (which were by now an experimental staple for us). We offered the chocolates to passing students at a broader range of prices. We’d already seen what economists call a “backward sloping demand curve” when the price was reduced from 1¢ to free (meaning demand went down instead of up when the price decreased). What would happen to demand when we decreased the price from 10¢ to 5¢? From 5¢ to 1¢? Or, as in the first experiment, when we decreased the price from 1¢ to free?

  When we dropped the chocolate prices from 10¢ to 5¢, the predictions of both laws of demand were verified, and in total we saw an increase in demand of about 240 percent. Similarly, when the price went from 5¢ to 1¢, the predictions of both laws of demand were verified, and in total we saw an increase in demand of about 400 percent. But, as we saw in our first experiment when we reduced the price from 1¢ to free, the first law of demand was verified (more people stopped for chocolate) but the second law of demand was disconfirmed (the people who took chocolate took less not more), and in total, with more people taking less, we saw a decrease in demand of about 50 percent.

  What these results mean is that the theory of demand is a solid one—except when we’re dealing with the price of zero. Whenever the price is not part of the exchange, social norms become entangled. These social norms get people to consider the welfare of others and, therefore, limit consumption to a level that does not place too much of a burden on the available resource. In essence, when prices are zero and social norms are a part of the equation, people look at the world as a communal good. The important lesson from all of this? Not mentioning prices ushers in social norms, and with those social norms, we start caring more about others.

  ANOTHER IMPORTANT LESSON from Chapter 4 was about the ability to obfuscate exchanges in nonfinancial terms and, by doing so, avoid squashing (“crowding out,” in economic terms) the benefits of social norms. (Giving your mother-in-law a gift is a good idea, but paying her for a wonderful Thanksgiving dinner is not recommended, even if both gestures would cost you the same amount of money.)

  If obfuscation can provide an important wedge that keeps us humming along the social norm track, and if gifts are one mechanism for obfuscating, what about other such mechanisms? What about effort? Over the years, I have been a beta tester for many software products, and in retrospect, it is amazing to realize how much time and effort and how many computer crashes I have endured from this activity. Could it be that because these soft
ware companies asked me for my time and effort without offering money, I was eager and happy to help them, even at substantial cost to myself?

  Now, effort is clearly not the same as a gift, but could it be that it is also different from money? Perhaps there’s a range with strict financial exchange norms at one extreme, pure social norms at the other, and effort somewhere along this spectrum? What, we wondered, might happen in a nonmonetary exchange involving only effort? Would effort undermine social norms in the same way that financial transactions do? Does an exchange of effort keep the social norms intact, similar to the effect of not mentioning money at all? Or might it fall somewhere in the middle of this spectrum?

  To explore this notion, we set up a new experiment. This time, we had a research assistant roam around the MIT Media Laboratory with a tray that was always stacked with fifty Lindt truffles. In the “free” condition, she simply asked people, “Would you like some chocolates?” (Note the plural, which indicates that it would be acceptable to take more than one.) In the “monetary” condition, she asked, “Would you like some chocolates? The cost is one cent each.” Finally, in the “effort” condition, she handed people a stack of pages with random arrangements of printed letters. For every pair of s’s they found, she said, they could have a Lindt truffle (but they did not have to take one). They could work as long as they wanted, putting in as much or as little effort as they wished. (There were well over fifty pairs of s’s among all the sheets of letters, and it was very easy to spot them.)

  The apparently voracious people in the monetary condition took an average of 30 truffles each. Those in the free condition took a polite average of 1.5. And how do you think effort-as-payment for truffles went over? This group fell somewhere in the middle, but closer to the free condition: our participants took an average of 8.6 truffles each in exchange for finding pairs of s’s (or an average of about 21 fewer than those in the monetary condition).

  These results suggest that effort falls somewhere in the middle of the range. It does not produce the same level of social self-consciousness that free does, but participants did seem to consider the implications of their actions on others when they decided how many truffles to take. We found that when effort is part of the equation, it manages to keep a large part of the social norms, though by no means all of it. As it turns out, the old maxim “Time is money”—or, in our case, “Effort is money”—is not exactly correct. Perhaps a more accurate reframing of our findings would be that effort is somewhere on the spectrum between market and social norms.

  The main lesson here is that because exchanges involving effort can maintain social norms to a larger degree relative to financial exchanges, we might want to think about how to get people to switch from paying for services to investing more of their own efforts. As we go about our daily lives, we are often asked to invest effort in recycling, spending time on a neighborhood watch, helping in our kids’ schools, volunteering in a soup kitchen, and much more. In each of these cases, one could argue that taking part in these activities makes little economic sense. Why not pay someone to recycle for us, watch our neighborhood, help in our kids’ schools, or hand out food in the soup kitchen? Sure, it might be economically inefficient, but investing effort rather than cash might help keep us in the domain of social norms and consequently take into account the welfare of others.

  FOR THE MOST part in this chapter, I’ve discussed instances where prices change from something to nothing. Of course, some things that are generally free or are considered a common resource can be relocated into the realm of market forces. For instance, carbon emissions trading is an area where we ought to consider the intersection between social and market norms.* “Cap and Trade” is a program of economic incentives designed to encourage industries and companies to pollute less; the less they pollute, the fewer pollution allowances they have to buy. Moreover, if companies don’t use all their allowances, they can profit from their cleanliness by selling their extra allowances to companies that pollute more. It’s virtue that pays!

  However, in light of the experiments described in this chapter (as well as those in Chapter 4), we might want to consider the dark side of putting a price on pollution. If a company can be charged for spewing poisons into the environment, it might well decide, after a cost-benefit analysis, that it can go ahead and pollute a lot more. Once pollution is a market and companies pay for their right to pollute, morality and concern for the environment are nonissues. On the other hand, if pollution is something that cannot be purchased or traded, it would more naturally fall into the domain of social norms.

  To be sure, if we want to place pollution under the control of social norms, we can’t stand back and hope that people will start caring. We need to make pollution into an easily measurable and observable quantity and get people to pay attention to it and understand its importance. We could, for example, publically post the pollutant amounts of different countries, states, and companies together with their environmental impact. We could include this information on companies’ financial statements to their shareholders or maybe force companies to post it on their products, much as we do for nutritional information on packaged food.

  I’m not saying that Cap and Trade is necessarily a bad idea, but I do think that when public policy or environmental issues are at stake, our task is to figure out which of the two—social or market norms—will produce the most desirable outcome. In particular, policy makers should be careful not to add market norms that could undermine the social ones.

  NOW THAT WE’VE learned how social norms get people to care less about their own selfish goals and pay more attention to the welfare of others, you might expect me to propose a brilliant idea for injecting more social norms and civility into the Filene’s Basement “Running of the Brides.” I wish I had a solution for getting these women to behave in a more considerate or at least less violent way. But the haunting memories of watching the live event suggest to me that getting a future bride to concentrate on an abstract idea like “other people” as opposed to the concrete reality of a discounted wedding gown might be nearly impossible. (For weeks afterward, I would look into the faces of my female friends and wonder whether they, too, were capable of trampling each other in an abject act of retail lust.)

  And why, you might ask, am I so easily giving up on this social science challenge? Because I suspect that for social norms to operate, people cannot be at their most emotionally piqued state. When you’re focused, mind and body, on one highly emotional objective—grabbing that wedding dress—it’s hard to factor in others’ well-being. As we will see in the next chapter, when emotions run high, social norms inevitably get trampled like so many Vera Wang veils.

  CHAPTER 6

  The Influence of Arousal

  Why Hot Is Much Hotter Than We Realize

  Ask most twentysomething male college students whether they would ever attempt unprotected sex and they will quickly recite chapter and verse about the risk of dreaded diseases and pregnancy. Ask them in any dispassionate circumstances—while they are doing homework or listening to a lecture—whether they’d enjoy being spanked, or enjoy sex in a threesome with another man, and they’ll wince. No way, they’d tell you. Furthermore, they’d narrow their eyes at you and think, What kind of sicko are you anyhow, asking these questions in the first place?

  In 2001, while I was visiting Berkeley for the year, my friend, academic hero, and longtime collaborator George Loewenstein and I invited a few bright students to help us understand the degree to which rational, intelligent people can predict how their attitudes will change when they are in an impassioned state. In order to make this study realistic, we needed to measure the participants’ responses while they were smack in the midst of such an emotional state. We could have made our participants feel angry or hungry, frustrated or annoyed. But we preferred to have them experience a pleasurable emotion.

  We chose to study decision making under sexual arousal—not because we had kinky predilections ourselves,
but because understanding the impact of arousal on behavior might help society grapple with some of its most difficult problems, such as teen pregnancy and the spread of HIV-AIDS. There are sexual motivations everywhere we look, and yet we understand very little about how these influence our decision making.

  Moreover, since we wanted to understand whether participants would be able to predict how they would behave in a particular emotional state, the emotion needed to be one that was already quite familiar to them. That made our decision easy. If there’s anything predictable and familiar about twentysomething male college students, it’s the regularity with which they experience sexual arousal.

  ROY, AN AFFABLE, studious biology major at Berkeley, is in a sweat—and not over finals. Propped up in the single bed of his darkened dorm room, he’s masturbating rapidly with his right hand. With his left, he’s using a one-handed keyboard to manipulate a Saran-wrapped laptop computer. As he idles through pictures of buxom naked women lolling around in various erotic poses, his heart pounds ever more loudly in his chest.

  As he becomes increasingly excited, Roy adjusts the “arousal meter” on the computer screen upward. As he reaches the bright red “high” zone, a question pops up on the screen:

  Could you enjoy sex with someone you hated?

  Roy moves his left hand to a scale that ranges from “no” to “yes” and taps his answer. The next question appears: “Would you slip a woman a drug to increase the chance that she would have sex with you?”

  Again, Roy selects his answer, and a new question pops up. “Would you always use a condom?”

  BERKELEY ITSELF IS a dichotomous place. It was a site of antiestablishment riots in the 1960s, and people in the Bay Area snarkily refer to the famously left-of-center city as the “People’s Republic of Berkeley.” But the large campus itself draws a surprisingly conformist population of top-level students. In a survey of incoming freshmen in 2004, only 51.2 percent of the respondents thought of themselves as liberal. More than one-third (36 percent) deemed their views middle-of-the-road, and 12 percent claimed to be conservatives. To my surprise, when I arrived at Berkeley, I found that the students were in general not very wild, rebellious, or likely to take risks.