As you’ve just seen, people are good at language. Even young children understand and produce sentences expertly. Now, however, we’ll explore three types of thinking—reasoning, judgment, and decision making—in which even educated people make systematic mistakes.

Do you read detective stories? If so, you can anticipate the climax. Before dazzled onlookers, the private eye reviews evidence, analyzes its implications logically, and draws conclusions so compelling that the criminal confesses on the spot.

The detective is engaged in reasoning, which is the process of drawing conclusions that are based on facts, beliefs, and experiences (Johnson-Laird, 1990). In reasoning, a person does not merely remember facts, but combines them to reach conclusions that go beyond the information that originally was perceived (Markman & Gentner, 2001).

Detective stories preview research findings in the psychology of reasoning. Notice not only the powers of the detective, but also the reactions of the onlookers: dazzled. In a good detective story, the facts are available to everyone—including the reader—but only the detective figures things out. Others are stumped. Detective stories suggest that most people aren’t very good at reasoning.

Much psychological research suggests the same thing. A distinguished psychologist concluded that research on logical reasoning skills shows that people have only “a modicum of competence” (Johnson-Laird, 1999, p. 109). Why aren’t we better at reasoning?

CONFIRMATION BIAS. Reasoning often is impaired by confirmation bias. Confirmation bias is the tendency to seek out information that is consistent with whatever initial conclusions you have drawn, and to disregard information that might contradict those conclusions.

329

Confirmation bias can impair reasoning because your initial conclusion might be wrong. To find out for sure, you should seek out disconfirming evidence, that is, evidence that could disprove your guess. If you think “the man who committed the midnight murder in the hotel was the butler,” you’d better be sure there’s no evidence that the butler was somewhere other than the hotel at midnight. If you look only for confirming evidence (“the butler had a motive”), you might miss critical disconfirming evidence (“witnesses saw the butler asleep in a movie theatre at midnight”).

You can see confirmation bias in action by trying an arithmetic problem (from Wason, 1960). Your job is to figure out the rule of arithmetic that generates sets of numbers. You first see three numbers that conform to the rule. You then need to suggest additional sets of three numbers that might help you determine what the rule is. After each set of three, you’re told whether your numbers conform to the rule. When you think you know what the rule is, you can take a guess.

Here are the first three numbers:

2 4 6

What do you think the rule might be? And what new set of three numbers would you suggest to find out what it is?

Here’s what a typical participant did. The participant would first think the rule is “a series of even numbers” and then suggest new sets of even numbers to see if they conform to the rule. “How about 8, 10, 12?” “14, 16, 18?” “20, 22, 24?” Each time, the participant learned that, yes, the numbers conform to the rule the experimenter had used. Yet the rule was not “a series of even numbers.” The actual rule was “any three numbers arranged in order of increasing magnitude.”

Participants thus displayed confirmation bias. They suggested numbers consistent with their initial guess, and failed to suggest other possibilities (e.g., 1, 3, 5) that could disconfirm their initial guess.

Confirmation bias is pervasive. Once people develop opinions and beliefs, they commonly seek out information that confirms their thinking (Nickerson, 1998). Social psychologists (see Chapter 12) propose, for example, that confirmation bias explains people’s tendency to believe media reports are biased against their own political attitudes. When people with either pro-Israeli or pro-Arab opinions viewed identical media coverage of a conflict in the Middle East, both groups thought the coverage was biased against their point of view (Vallone, Ross, & Lepper, 1985). How could people perceive opposite biases—anti-Israeli, anti-Arab—in the same media coverage? Confirmation bias explains it. Before the broadcast began, people already expected the media coverage to be unfair. During the coverage, they attended closely to anything that confirmed this expectation. When the broadcast was over, they had better memory for this confirmatory information. Confirmation bias, then, caused everyone to conclude that the media coverage was unfair to their own point of view.

REASONING ABOUT EVOLUTIONARILY RELEVANT PROBLEMS. Do people always find reasoning difficult? The psychologist Leda Cosmides doesn’t think so. She notes that reasoning problems presented in research—identifying rules of arithmetic, evaluating media coverage—were not relevant during the course of human evolution. Contemporary humans might be better at reasoning when contemplating problems that their evolutionary ancestors faced.

One such problem is not getting cheated when exchanging goods (Cosmides, 1989). People have exchanged goods for thousands of years, either by trading or by using currency. In the past, they might have exchanged livestock for grain. Today, you exchange money for groceries. Throughout human history, it has been important to avoid being cheated. Evolution may have equipped people with the ability to reason accurately about problems that involve the possibility of cheating when goods are exchanged.

330

Cosmides tested this by presenting different problems with an identical logical structure—that is, the same relations among pieces of information. The problems, however, differed in substance, that is, in the topic being discussed:

Participants performed poorly on problems that had nothing to do with cheating (the document filing problem). However, they performed excellently when the problem involved potential cheating by the breaking of a social rule (the cassava root problem; Cosmides, 1989). On the latter task, they avoided confirmation bias.

Why are people good at solving cheating-detection problems? Cosmides suggests that evolution has produced an innate brain mechanism dedicated to the solution of problems involving the exchange of goods. Although this claim is controversial, it is consistent with a finding from brain research: Different parts of the brain are active when people reason about social exchange problems than when they reason about logically similar problems that do not involve the exchange of goods (Ermer et al., 2006).

Does Cosmides’s idea sound familiar? It should. Her idea about the origins of a brain mechanism for solving problems is essentially the same as Chomsky’s idea about the origins of a brain mechanism dedicated to processing grammar. Both claim that evolution gave rise to the brain mechanisms and associated mental abilities that people possess today (Buss, 2009).

Life is full of guesses, big and small. “Will I like this new romantic comedy with computer-animated talking animals enough to spend $10 on a movie ticket?” “Will I get as good an education at moderately expensive State U as at super-expensive Private College?” “Is the relationship I’m in ‘just for now,’ or really serious?” “Should I invest money in the stock market, or are stocks likely to lose value in the years ahead?” To answer each of these questions, you need to make a judgment, that is, you need to draw a conclusion based on an evaluation of evidence.

331

For each question, uncertainty exists. You don’t know for sure how things will turn out. You don’t even know the probability that things will turn out one way or the other: Are the chances that you’ll enjoy the movie more like a 50-50 coin flip or a 100-to-1 long shot?

The question for psychology is how people make these guesses. What are the mental processes through which people make judgments under conditions of uncertainty? There are two general types of possibilities.

A heuristic is a rule of thumb, that is, a simple way of accomplishing something that otherwise would be done through a more complex procedure. For example, if you are in a residential area of a big city and wish to go downtown, you could follow step-by-step directions from an Internet mapping service or you could use a heuristic: “Move toward the tallest buildings.” The heuristic will get you downtown.

Judgmental heuristics are simple mental procedures for making judgments under conditions of uncertainty. Extraordinarily insightful research by Amos Tversky and Daniel Kahneman identified three judgmental heuristics that people use to make a wide variety of judgments: the availability, representativeness, and anchoring-and-adjustment heuristics (Tversky & Kahneman, 1974).

332

AVAILABILITY. In the availability heuristic, people base judgments on the ease with which information comes to mind. The rule of thumb—the heuristic—is that when information about a person, place, or thing comes to mind easily, the person, place, or thing is more common. Suppose someone asks, “Do more people have last names starting with S or Z?” You don’t have to count up all the people in the world. You can instead base your judgments on availability. It’s easy to think of people with S last names, but hard to think of many Z names, so the relative ease tells you that S names are more common.

In general, then, the availability heuristic works well. But sometimes it creates errors. This happens when factors other than actual frequency affect availability. Here’s an example.

What do you think is more common: words (of three or more letters) that start with the letter k or that have k as the third letter in the word? (Take a moment to guess.)

You probably thought something like this: “Hmm, first letter k; that’s easy: kite, kitten, kayak, kazoo, keep, keeper, keeping, keepsake, knife, knifing, knit, knitting…. Geez, there’s millions of ’em. K in third position … um … Bake, rake, make … um … ark.” Words starting with k come to mind easily. Most people, relying on the availability heuristic, thus say that words starting with k are more common. But in reality, words with k as the third letter are more common. The fact that we organize our “mental dictionary” according to words’ first letters makes them more available mentally, despite their being less frequent in reality.

REPRESENTATIVENESS. The representativeness heuristic is a psychological process that comes into play when people have to decide whether a person or thing is a member of a given category. When relying on representativeness, people base these judgments on the degree to which the person or object resembles the category. Suppose you meet a shy, middle-aged man wearing eyeglasses and a sweater. Do you think he is a librarian or a construction worker? If your stereotype of a librarian includes “glasses” and “sweater,” you’ll say “librarian.” The man resembles, or is representative of, the category “librarian.”

The representativeness heuristic often produces accurate judgments. However, like availability, it can cause people to make judgmental errors. Consider this example from Tversky and Kahneman (1983):

Linda is 31 years old, single, outspoken, and very bright. She majored in philosophy. As a student, she was deeply concerned with issues of discrimination and social justice and also participated in anti-nuclear demonstrations. Which is more probable?

Did you say “B”? Six out of seven people did when Tversky and Kahneman asked them. But “B” is wrong because everyone who is “a bank teller and active in the feminist movement” also is “a bank teller.” “B” therefore cannot be more probable than “A”; all “B”s are “A”s (all feminist bank tellers are bank tellers) and some “A”s are not “B”s (some bank tellers are not feminists). If the human mind worked like a statistical computer program, everyone would say “A.” However, the mind instead relies on representativeness and, since Linda resembles people’s conception of a feminist, most people say “B.”

In general, the representativeness heuristic can produce errors when it distracts people from the base rates of categories, that is, the overall likelihood that any item would be in the category. Consider our librarian/construction worker example. There are 8 times more construction workers in the United States than there are librarians (www.bls.gov); the base rate of the category “construction worker” is much higher. So the man—even if he is wearing glasses and a sweater—may well be a construction worker.

333

TRY THIS!

These judgment problems should sound familiar to you; they were in this chapter’s Try This! activity. If you haven’t completed that yet, go to www.pmbpsychology.com and do it now.

ANCHORING AND ADJUSTMENT. The third judgmental heuristic, anchoring and adjustment, comes into play when people estimate an amount (e.g., “What GPA am I likely to earn next semester?” “How many people will show up at the party I’m planning?”). In the anchoring-and-adjustment heuristic (or just “anchoring”), people make estimates by formulating an initial guess (the “anchor”) and adjusting it to reach a final judgment.

Anchoring can create judgmental errors. This happens when irrelevant anchor values bias people’s judgments. Tversky and Kaheman asked participants to estimate the percentage of world nations that are African. Before participants made their estimates, the researchers spun a wheel of fortune containing numbers from 0 to 100 and asked participants to indicate whether the correct answer was more or less than the random value generated by the wheel. The anchor values were random—obviously irrelevant to the task. Yet they biased people’s judgment. Participants adjusted their estimates up from low anchors and down from high ones, but the adjustments were insufficient. Those who saw a high random anchor thought that nearly twice as many of the world’s nations were African as did people who first saw a low random number.

Random anchor values can affect important judgments, such as people’s judgments about how much money to spend on a purchase. Just prior to an auction for an item whose value was less than $100, researchers asked bidders to write down the last two digits of their Social Security number. “Huh?”—you should be saying to yourself. “What difference would writing down a Social Security number make to an auction bid?” Due to anchoring effects, it made a big difference. People’s bids were correlated with their Social Security numbers; people with higher Social Security numbers bid more money (Ariely, Loewenstein, & Prelec, 2004). The numbers came to mind, and served as anchors, when they asked themselves, “How much should I bid?”

Table 8.1 summarizes these three heuristics and reasons why reliance on heuristics can sometimes produce judgmental errors.

334

Cherry or blueberry pie? Rent a DVD or go to a movie theatre? Vacation in an exciting metropolis or a relaxing wilderness? Have surgery for your broken bone or put it in a cast and see if it heals by itself? Life presents decisions, large and small.

Decision making is the process of making a choice, that is, selecting among alternatives. Sometimes the benefits of the alternatives cannot be known for sure; you can’t tell how well the broken bone will heal if put in a cast. Other times, the benefits are known with certainty, yet it’s still difficult to decide. You know what cherry pie and blueberry pie taste like, yet it’s hard to choose between them.

How do people make decisions? To understand the psychology of decision making, it’s best first to consider a “standard model” of decision making and then to look at psychological factors that make the standard model inadequate. (The standard model was, for many years, the explanation of decision making in economics.)

The standard model of decision making has two components. One involves subjective value, the degree of personal worth that an individual places on an outcome. In the standard model of decision making, people choose the alternative available to them that has the highest subjective value. Note that subjective value can differ from objective value. Consider two cases: (1) Two people offer you a temporary one-day secretarial job; one will pay $50 and the other $75. (2) Two people offer you a three-month secretarial job; one will pay $6250 and the other $6275. In both cases, the objective difference between the offers is $25. But in #1, the difference feels big, whereas in #2, it feels small. The subjective value of $25 differs in the two cases.

The second component of the standard theory is that, when making decisions, people consider how their choices will affect their net worth (i.e., the overall monetary assets they possess). You might enjoy a city vacation more than a wilderness vacation. But if either vacation would wipe out your personal savings, you’ll probably skip a vacation for this year.

In the standard model, people’s choices are highly logical. The decision maker compares subjective values, evaluates personal worth, and calculates a logical, rational decision. In reality, however, people aren’t as logical and rational as the standard model suggests. Research by Kahneman and Tversky (1979, 1984) undermined the standard model of decision making.

335

FRAMING EFFECTS. A major blow to the standard model was Kahneman and Tversky’s demonstration of framing effects. In a framing effect, decisions are influenced by the way alternative choices are described, or “framed.” According to the standard model, different framings should be inconsequential, because they don’t affect the actual value of alternatives. However, it turns out that framing has big effects. Consider the following two choices (from Kahneman & Tversky, 1984, p. 343):

Choice #1

Imagine that the United States is preparing for the outbreak of an unusual Asian disease, which is expected to kill 600 people. Two alternative programs to combat the disease have been proposed. Assume that the exact scientific estimates of the consequences of the programs are as follows:

Which of the two programs would you favor, A or B?

Choice #2

Imagine the same circumstance as above (an unusual Asian disease is expected to kill 600 people) and consider the following two programs:

If you’re like most people, you chose Programs A and D. A seems better than B; if 200 people can be saved for sure (A), why choose a program whose most likely outcome is that no one will be saved (B)? D seems better than C; if one program leaves 400 people sure to die (C), why not choose an alternative that might prevent the deaths of everyone (D)?

About 3 out of 4 people chose Program A over B, and Program D over C. This seems sensible—except that A is identical to C, B is identical to D, and thus Choice #1 is identical (in terms of costs and benefits of the programs) to Choice #2. This becomes obvious if you look back at the choices. For example, 200 saved (A) is identical to 400 die (C) because there are 600 people.

The standard model predicts that decisions would be the same in Choices #1 and 2, because the numbers of people saved are identical in both cases. But preferences actually reverse: People prefer the sure thing (save 200) in Choice #1, which is framed in terms of lives saved, and the risky option (maybe save 600, but a big chance of saving no one) in Choice #2, where choices are framed in terms of deaths. People think differently about gains and losses; framing reverses the preferences.

MENTAL ACCOUNTING. Kahneman and Tversky identified a second problem with the standard model. Recall an idea from that model: People base decisions on how alternative choices affect their net worth. This idea, too, fails to describe the way people actually make decisions. Consider the following choices (adapted from Kahneman & Tversky, 1984, p. 347):

Choice #1

Imagine that you have decided to see a play and paid the admission price of $20 per ticket. As you enter the theater, you discover that you have lost the ticket. The seat was not marked, and the ticket cannot be recovered. Would you pay $20 for another ticket?

336

Choice #2

Imagine that you have decided to see a play where admission is $20 per ticket. As you enter the theater, you discover that you have lost a $20 bill. Would you pay $20 for a ticket for the play?

In Choice #1, most people say no, they would not buy another ticket. In Choice #2, almost 9 out of 10 people say that, yes, they would buy one. Psychologically, then, the choices differ enormously. But in terms of how they affect net worth, they are identical. In both, if people see the play, they end up with $40 less net worth than when they started; $20 of that worth was lost and $20 was spent on the ticket to get in. Why, then, do people’s choices differ?

As Kahneman and Tversky explain, when making decisions, people usually do not consider their overall net worth—the total value of all the bank accounts, investment accounts, property, and other assets they own. Instead, they engage in mental accounting, a thinking process in which people divide their assets and expenditures into distinct cognitive categories. For example, you might have a category for “amount of money I can spend on entertainment.” If you buy a $20 ticket and lose it, then that $20 comes out of your entertainment mental account, and you’re less likely to buy another ticket. If you lose a $20 bill, that $20 does not come out of your entertainment mental account; your entertainment mental account, then, still has as much money in it as it did before you lost the bill. So you buy a $20 ticket despite losing the $20 bill.

In summary, Tversky and Kahneman revolutionized science’s understanding of judgment and decision making by humanizing it. Rather than performing calculations like a computer, people use simplifying strategies—heuristics, mental accounts—required by human beings with limited short-term memory capacity.