The Paradox of Choice Read online

  Here’s another, quite remarkable example of the peak-end rule in operation. Men undergoing diagnostic colonoscopy exams were asked to report how they felt moment by moment while having the exam, and how they felt when it was over. Most people find these exams, in which a tube with a tiny camera on the end is inserted up the rectum and then moved around to allow the inspection of the gastrointestinal system, quite unpleasant—so much so that patients avoid getting regular tests, much to their peril. In the test, one group of patients had a standard colonoscopy. A second group had a standard colonoscopy plus. The “plus” was that after the actual examination was over, the doctor left the instrument in place for a short time. This was still unpleasant, but much less so because the scope wasn’t moving. (Note that both groups of patients were having the colonoscopies for legitimate medical reasons; they were not subjecting themselves to these procedures just for the sake of the experiment.) So the second group experienced the same moment-by-moment discomfort as the first group, with the addition of somewhat lesser discomfort for twenty seconds more. And that is what they reported, moment-by-moment, as they were having the procedure. But a short time after it was over, the second group rated their experience as less unpleasant than did the first. Whereas both groups had the same peak experience, the second group had a milder end experience.

  And it made a difference. It turned out that, over a five-year period after this exam, patients in the second group were more likely to comply with calls for follow-up colonoscopies than patients in the first group. Because they remembered their experiences as less unpleasant, they were less inclined to avoid them in the future.

  In the same way, we evaluate positive experiences on the basis of how good they feel at their best, and how good they feel at the end. Thus, you might, in retrospect, remember a one-week vacation that had some great moments and finished with a bang as more pleasurable than a three-week vacation that also had some great moments, but finished only with a whimper. The two extra weeks of relaxing in the sun or seeing the sights or eating great food make little difference, because they recede from awareness over time.

  So how well do we know what we want? It’s doubtful that we would truly prefer intense pain followed by mild pain over experiencing intense pain alone. It’s unlikely that a great one-week vacation is truly better than a great-single-week-followed-by-a-pretty-good-two-weeks vacation. But that’s what people say they prefer. The discrepancy between logic and memory suggests that we don’t always know what we want.

  Another illustration of our lack of self-knowledge comes from a study in which researchers asked a group of college students to choose a series of snacks. Each week they had a three-hour seminar with one break that allowed participants to stretch their legs, use the bathroom, clear their heads, and have something to eat. When the professor asked the students to pick a snack for each of the next three weeks, the students picked a variety, thinking they’d get tired of the same snack each week. In contrast, another group in the same study got to choose their snack every week, and these students, choosing for one week at a time, tended to choose the same thing each week.

  These two sets of participants were faced with different tasks. The students who were choosing one snack at a time simply had to ask themselves what they felt like eating at the moment. Those who were choosing for three weeks had to predict what they would feel like eating two or three weeks from the moment of choice. And they got the prediction wrong, no doubt thinking that their low enthusiasm for pretzels after having just eaten a bag was how they would feel about pretzels a week later.

  People who do their grocery shopping once a week succumb to the same erroneous prediction. Instead of buying several packages of their favorite X or Y, they buy a variety of Xs and Ys, failing to predict accurately that when the time comes to eat X or Y, they would almost certainly prefer their favorite. In a laboratory simulation of this grocery shopping situation, participants were given eight categories of basic foods and asked to imagine doing their shopping for the day and buying one item in each category. Having done this, they were asked to imagine doing it again, the next day, and so on, for several days. In contrast, another group of people were asked to imagine going shopping to buy three days’ worth of food, and thus selecting three things in each category. People in this latter group made more varied selections within each category than people in the former group, predicting, inaccurately, that they would want something different on day two from what they had eaten on day one.

  So it seems that neither our predictions about how we will feel after an experience nor our memories of how we did feel during the experience are very accurate reflections of how we actually do feel while the experience is occurring. And yet it is memories of the past and expectations for the future that govern our choices.

  In a world of expanding, confusing, and conflicting options, we can see that this difficulty in targeting our goals accurately—step one on the path to a wise decision—sets us up for disappointment with the choices we actually make.

  Gathering Information

  HOWEVER WELL OR POORLY WE DETERMINE OUR GOALS BEFORE making a decision, having set them, we then go through the task of gathering information to evaluate the options. To do this, we review our past experience as well as the experience and expertise of others. We talk to friends. We read consumer, investment, or lifestyle magazines. We get recommendations from salespeople. And increasingly, we use the Internet. But more than anything else, we get information from advertising. The average American sees three thousand ads a day. As advertising professor James Twitchell puts it, “Ads are what we know about the world around us.”

  So we don’t have to do our choosing alone and unaided. Once we figure out what we want, we can use various resources to help evaluate the options. But we need to know that the information is reliable, and we need to have enough time to get through all the information that’s available. Three thousand ads a day breaks down to about two hundred per waking hour, more than three per waking minute, and that is an overwhelming amount to sift through.

  Quality and Quantity of Information

  TO ACCOMMODATE THE EVER-INCREASING NUMBER OF ADS, YOUR favorite sitcom has about four fewer program minutes than it did a generation ago. On top of that, the advent of cable TV and its many channels has brought with it the “infomercial,” a show that is an ad masquerading as entertainment. Newspapers and magazines contain hundreds of pages of which just a small fraction are devoted to content. Movie producers now “place” branded products in their films for high fees. Increasingly, sports stadiums are named for a sponsoring company, often at a fee of several million dollars a year. Every race car is tattooed with brand names, as are many athletes’ uniforms. Even public television now has ads, disguised as public service announcements, at the start and end of almost every show.

  Unfortunately, providing consumers with useful decision-making information is not the point of all this advertising. The point of advertising is to sell brands. According to James Twitchell, the key insight that has shaped modern advertising came to cigarette manufacturers in the 1930s. In the course of market research, they discovered that smokers who taste-tested various cigarette brands without knowing which was which couldn’t tell them apart. So, if the manufacturer wanted to sell more of his particular brand, he was either going to have to make it distinctive or make consumers think it was distinctive, which was considerably easier. With that was born the practice of selling a product by associating it with a glamorous lifestyle.

  We probably like to think that we’re too smart to be seduced by such “branding,” but we aren’t. If you ask test participants in a study to explain their preferences in music or art, they’ll come up with some account based on the qualities of the pieces themselves. Yet several studies have demonstrated that “familiarity breeds liking.” If you play snippets of music for people or show them slides of paintings and vary the number of times they hear or see the music and the art, on the whole people will rate the
familiar things more positively than the unfamiliar ones. The people doing the ratings don’t know that they like one bit of music more than another because it’s more familiar. Nonetheless, when products are essentially equivalent, people go with what’s familiar, even if it’s only familiar because they know its name from advertising.

  If people want real information, they have to go beyond advertising to disinterested sources such as Consumer Reports. Its publisher, Consumers Union, is an independent, nonprofit organization whose mission is to help consumers. It does not allow any of its reports or ratings to be used in advertising, nor does the magazine contain any commercial advertising. When it was launched about seventy-five years ago, Consumer Reports offered comparisons among things like Grade A milk and Grade B milk. Today it offers comparisons among 220 new car models, 250 breakfast cereals, 400 VCRs, 40 household soaps, 500 health insurance policies, 350 mutual funds, and even 35 showerheads. And this barely scratches the surface. For every type of product that Consumer Reports evaluates, there are many that it passes over. And new models appear with such frequency that the evaluations are at least slightly out of date by the time they are published. The same limitation is true, of course, of other, more specialized guides—travel guides, college guides, and the like.

  The Internet can give us information that is absolutely up-to-the-minute, but as a resource, it is democratic to a fault—everyone with a computer and an Internet hookup can express their opinion, whether they know anything or not. The avalanche of electronic information we now face is such that in order to solve the problem of choosing from among 200 brands of cereal or 5,000 mutual funds, we must first solve the problem of choosing from 10,000 web sites offering to make us informed consumers. If you want to experience this problem for yourself, pick some prescription drug that is now being marketed directly to you, then do a web search to find out what you can about the drug that goes beyond what the ads tell you. I just tried it for Prilosec, one of the largest-selling prescription medications in existence, which is heavily advertised by its manufacturer. I got more than 20,000 hits!

  And there is good evidence that the absence of filters on the Internet can lead people astray. The RAND Corporation recently conducted an assessment of the quality of web sites providing medical information and found that “with rare exceptions, they’re all doing an equally poor job.” Important information was omitted, and sometimes the information presented was misleading or inaccurate. Moreover, surveys indicate that these web sites actually influence the health-related decisions of 70 percent of the people who consult them.

  Evaluating the Information

  EVEN IF WE CAN ACCURATELY DETERMINE WHAT WE WANT AND THEN find good information, in a quantity we can handle, do we really know how to analyze, sift, weigh, and evaluate it to arrive at the right conclusions and make the right choices? Not always. Spear-headed by psychologists Daniel Kahneman and Amos Tversky, researchers have spent the last thirty years studying how people make decisions. Their work documents the variety of rules of thumb we use that often lead us astray as we try to make wise decisions.

  Availability

  IMAGINE THAT YOU’RE IN THE MARKET FOR A NEW CAR AND THAT YOU care about only two things: safety and reliability. You dutifully check out Consumer Reports, which rates Volvo highest for safety and reliability, so you resolve to buy a Volvo. That evening, you’re at a cocktail party and you mention your decision to a friend. “You’re not going to buy a Volvo,” she says. “My friend Jane bought one about six months ago, and she’s had nothing but trouble. First there was an oil leak; then she had trouble starting it; then the tape player started mangling her tapes. She’s had it in the shop maybe five times in the six months she’s owned it.”

  You might feel lucky to have had this conversation before making a terrible mistake, but actually, maybe you’re not so fortunate. Consumer Reports makes its judgments about the reliability of cars by soliciting input from its thousands and thousands of readers. It compiles this input into an estimate of reliability for each make and model of car. So when Consumer Reports says that a car is reliable, it is basing its conclusion on the experience of thousands of people with thousands of cars. This doesn’t mean that every single Volvo driver will have the same story to tell. But on average, the reports of Volvo owners are more positive about reliability than the reports of the owners of other cars. Now along comes this friend to tell you about one particular Volvo owner and one particular Volvo. How much weight should you give this story? Should it undo conclusions based on the thousands of cases assessed by Consumer Reports? Of course not. Logically, it should have almost no influence on your decision.

  Unfortunately, most people give substantial weight to this kind of anecdotal “evidence,” perhaps so much so that it will cancel out the positive recommendation found in Consumer Reports. Most of us give weight to these kinds of stories because they are extremely vivid and based on a personal, detailed, face-to-face account.

  Kahneman and Tversky discovered and reported on people’s tendency to give undue weight to some types of information in contrast to others. They called it the availability heuristic. This needs a little explaining. A heuristic is a rule of thumb, a mental shortcut. The availability heuristic works like this: suppose someone asked you a silly question like “What’s more common in English, words that begin with the letter t or words that have t as the third letter?” How would you answer this question? What you probably would do is try to call to mind words that start with t and words that have t as the third letter. You would then discover that you had a much easier time generating words that start with t. So words starting with t would be more “available” to you than words that have t as the third letter. You would then reason roughly as follows: “In general, the more often we encounter something, the easier it is for us to recall it in the future. Because I had an easier time recalling words that start with t than recalling words with t as the third letter, I must have encountered them more often in the past. So there must be more words in English that start with t than have it as the third letter.” But your conclusion would be wrong.

  The availability heuristic says that we assume that the more available some piece of information is to memory, the more frequently we must have encountered it in the past. This heuristic is partly true. In general, the frequency of experience does affect its availability to memory. But frequency of experience is not the only thing that affects availability to memory. Salience or vividness matters as well. Because starting letters of words are much more salient than third letters, they are much more useful as cues for retrieving words from memory. So it’s the salience of starting letters that makes t-words come easily to mind, while people mistakenly think it’s the frequency of starting letters that makes them come easily to mind. In addition to affecting the ease with which we retrieve information from memory, salience or vividness will influence the weight we give any particular piece of information.

  There are many examples of the availability heuristic in operation. When college students who are deciding what courses to take next semester are presented with summaries of course evaluations from several hundred students that point in one direction, and a videotaped interview with a single student that points in the other direction, they are more influenced by the vivid interview than by the summary judgments of hundreds. Vivid interviews with people have profound effects on judgment even when people are told, in advance of seeing the interviews, that the subjects of the interview are atypical. Thus seeing an interview of an especially vicious (or humane) prison guard or an especially industrious (or slothful) welfare recipient shifts people’s opinions of prison guards or welfare recipients in general. When spouses are asked (separately) a series of questions about what’s good and bad about their marriage, each spouse holds him or herself more responsible than his or her partner, for both the good and the bad. People’s natural egocentrism makes it much easier to bring their own actions to mind than those of their partner. Because our own actions ar
e more available to us from memory, we assume they are more frequent.

  Now consider the availability heuristic in the context of advertising, whose main objective is to make products appear salient and vivid. Does a particular carmaker give safety a high priority in the manufacture of its cars? When you see film footage of a crash test in which a $50,000 car is driven into a wall, it’s hard to believe the car company doesn’t care about safety, no matter what the crash-test statistics say.

  How we assess risk offers another example of how our judgments can be distorted by availability. In one study, researchers asked respondents to estimate the number of deaths per year that occur as a result of various diseases, car accidents, natural disasters, electrocutions, and homicides—forty different types of misfortune in all. The researchers then compared people’s answers to actual death rates, with striking results. Respondents judged accidents of all types to cause as many deaths as diseases of all types, when in fact disease causes sixteen times more deaths than accidents. Death by homicide was thought to be as frequent as death from stroke, when in fact eleven times more people die of strokes than from homicides. In general, dramatic, vivid causes of death (accident, homicide, tornado, flood, fire) were overestimated, whereas more mundane causes of death (diabetes, asthma, stroke, tuberculosis) were underestimated.

  Where did these estimates come from? The authors of the study looked at two newspapers, published on opposite sides of the U.S., and they counted the number of stories involving various causes of death. What they found was that the frequency of newspaper coverage and the respondents’ estimates of the frequency of death were almost perfectly correlated. People mistook the pervasiveness of newspaper stories about homicides, accidents, or fires—vivid, salient, and easily available to memory—as a sign of the frequency of the events these stories profiled. This distortion causes us to miscalculate dramatically the various risks we face in life, and thus contributes to some very bad choices.