Teaching the Use of Cost-Benefit Reasoning in Everyday Life

Our research shows that people can apply the costbenefit rules of microeconomic theory to their everyday decisions. Two populations were examined: (a) people who had previously received extensive formal training in the rules and (b) naive subjects who were randomly assigned to receive brief training in the rules. Training affected reasoning and reported behavior in both populations. The results indicate that extremely general rules govern choices across a wide range of domains and that use of the cost-benefit rules can be improved through training. Do people think about leaving bad movies, taking losses on investments, and demolishing outdated buildings in the same way, relying on the same inferential rules? Are there general rules of choice that guide people's decisions across a broad range of domains? If so, can the nature of these rules be altered by instruction? Economists have long assumed that a set of abstract rules of choice exists, and that people make choices according to those rules across the economic spectrum from consumer choices to decisions about time use (Becker, 1976). On the other hand, many psychologists share Thorndike's (1906) view that people's behavior is governed by domain-dependent rules that do not generalize across situations that are very different from one another. Even among psychologists who believe that people have domain-independent rules, a pervasive opinion originating with Piaget is that such rules can only be learned by selfdiscovery methods and cannot really be taught in a formal way (Brainerd, 1978; Newell, 1980). In this paper, we examine whether people use abstract rules in making choices and whether these principles can be taught so that people will use them across a wide range of situations. We first sketch the cost-benefit model of choice and show that there is ample evidence that people depart from it in their everyday choices. We next present original evidence about differential rule usage, comparing people trained in the use of the normative rules with those who are less trained. Finally, we present evidence about the trainability of people by relatively brief interventions, about the degree of domain specificity of training, and about the durability of training. COST-BENEFIT RULES OF CHOICE The basic assumption underlying the microeconomic model of choice is that people choose actions that maximize their overall welfare. The model assumes people are confronted by a set of actions, each of which is associated with outcomes that will occur with some probability. Each of these outcomes has a value to the person that can be compared with the values of the other outcomes by converting the values to a single scale (for example, dollars). Three principles follow from this model, (a) The greatest net benefit principle, which states that the action with the greatest positive difference between the total benefit and total cost of its outcome should be chosen from the set of possible actions, (b) The sunk cost principle, which states that only future benefits and costs should be considered when making a choice, (c) The opportunity cost principle, which states that a cost of undertaking a given course of action is foregoing the expected net benefits associated with other courses of action. These principles may seem incomplete to readers familiar with standard textbook cost-benefit analysis (Mishan, 1976; Morgan & Duncan, 1982). We eliminated certain principles from the list of inferential rules of choice because they were either derivable from the principles listed above, or because they were not sufficiently general to apply to all choice situations. Also, these cost-benefit rules certainly are not exhaustive of the principles that guide choice. Moral, esthetic, and social principles may compete with the cost-benefit rules, and humanistic values may be included as costs and benefits in the economic calculus. Thus, departures from economic reasoning could be due to people's favoring humanistic values or principles rather than ignorance of the normative rules. We will return to the question of whether noneconomically trained people know cost-benefit reasoning but are more likely to favor humanistic values, or whether they are actually less likely to know and use cost-benefit reasoning. Lay Use of the Cost-Benefit Rules of Choice A fundamental assumption of economics is that people make choices consistent with the cost-benefit rules (Friedman, 1953). There now exists, however, a good deal of research indicating that people violate the maximization rules of microeconomics. An early dissenting view introduced by Simon (1955) proposed that people do not attempt to maximize utility but simply attempt to attain some minimal level of satisfaction. In the past few decades, many empirical violations of expected utility theory have been documented, including violations of the independence axiom (Allais, 1953) and transitivity (Lichtenstein & Slovic, 1971; Tversky, 1969) and various violations of utility maximization (Kahneman & Tversky, 1979). These critiques Correspondence and reprint requests to Richard E. Nisbett, 5261 ISR, The University of Michigan, Ann Arbor, MI 48106. 362 Copyright © 1990 American Psychological Society VOL. 1, NO. 6, NOVEMBER 1990 PSYCHOLOGICAL SCIENCE Richard P. Larrick et al. have buttressed the psychologists' case against the view that a single, coherent set of abstract rules underlies choice behavior and have revealed serious shortcomings in choice behavior that produce nonoptimal outcomes. We shall focus on several economic traps that occur when people fail to use the sunk cost and opportunity cost principles. Violations of the Sunk Cost Principle A particularly clear case of an economic trap occurs when people consider past investments of money, effort, or time in their current decisions. For instance, imagine you have bought $15 tickets for a basketball game weeks in advance, but on the day of the game it is snowing and the star player you wanted to see is sick and will not play. The game is no longer of much interest to you. Do you decide to go to see the game or stay home and forget the money you have spent? (Adapted from Thaler, 1980.) In these circumstances, some people will go to the game for the primary reason that they spent a large amount of money on it. What is wrong with this? Once the money is sunk, going to the game will not enable the person to get his or her "money's worth," but will simply inflict additional costs of a dangerous trip and a boring game. The sunk cost principle prescribes that a decision maker should consider only the benefits and costs that are going to be incurred from the time of the decision forward and ignore the past costs. To make it easier to ignore past costs, the person can do a thought experiment: Would I go to the game if someone called me on the day it was scheduled and offered me a free ticket? If not, then a sunk cost trap should be suspected. The sunk cost effect has been demonstrated in studies by Arkes and Blumer (1985), who proposed that an important cause of the trap is peoples' ignorance of the normative sunk cost rule and their reliance on the generally effective rule of "waste not, want not." The "waste not, want not" rule is normative for economic choices that involve decisions about the future. People would be acting counternormatively if they deliberately chose to expend resources in pursuit of some benefit they intended to waste. But the rule is not normative for economic decisions involving irretrievable past costs. Note that this interpretation implies that people do operate with highly abstract rules of choice, but that these rules may differ from the normative ones. A second type of economic dilemma that occurs when people fail to ignore sunk costs is a situation we call the extra cost trap. In this type of trap, people abandon an enjoyable activity because they are attending to past costs. For instance, imagine you are on your way to see one of your favorite plays for which you have bought a ticket in advance, but when you get to the theater, you discover you have lost the ticket. (Adapted from Tversky & Kahneman, 1981.) Would you decide to buy a second ticket or to skip the play? Some people in this situation might skip the play because they feel that they have already spent $20 on a ticket and the play is not worth $40 (which is the total cost if the person buys a second ticket). They feel that it is a waste to spend twice as much on a play ticket as it is worth. The sunk cost principle, however, prescribes that you ignore the sunk $20 ticket, and decide whether the future benefits of seeing the play exceed the future cost, which is $20, the price of a second ticket. A thought experiment is also helpful here: Would you still go to the play if you had lost $20 in cash on the way to the theater? The answer is probably yes, because the value of the performance has not been affected. Violations of the Opportunity Cost Principle Another type of economic trap occurs when people ignore the opportunities they forego by choosing one course of action rather than another. For instance, imagine that you have paid off the mortgage on your home, and you are deciding whether to keep your house or move into an apartment. Many people compare the out-of-pocket costs of living in a house with those of living in an apartment and conclude that keeping the home is much cheaper than renting. However, by tying up wealth in a house, a person foregoes interest income, which is often more than enough to offset the difference in the out-of-pocket costs between owning and renting. The waste not, want not principle leads people to feel that it is a waste to "throw away" money on rent when they could live in their own home for "free." Research on opportunity costs establishes that people often fail to calculate opportunity costs when they are unstated (Neumann & Friedman, 1980) and tend to underestimate them when they are explicitly stated but small (Hoskin, 1983). Additional evidence that people have poor understanding of opportunity costs is provided by research on buying and selling behavior in experimental markets, which shows that the lowest amount of money subjects are willing to accept for a good is higher than the highest amount they are willing to pay for the same good (Kahneman, Knetsch, & Thaler, in press). An analogous discrepancy occurs in people's use of time. People will often perform a service for themselves (for example, mowing the lawn) even though the amount they could earn performing some other activity would be higher than the amount they would have to pay for someone else to perform the service. An awareness of opportunity costs should lead people to pay more for the services of others and to invest in time-saving devices. Evidence that People use Inferential Rules Taken as a whole, the evidence indicates that the economic model is wrong in assuming that people act as though they use the abstract normative rules. Indeed, the findings are consistent with the view that people do not use abstract rules of any type but rely on highly concrete domain-dependent rules. In a recent series of studies, however, Nisbett and his colleagues have found that people do use abstract, domainindependent inferential rules to reason about certain everyday problems, and that their reasoning can be improved by formal instruction (for a review see Nisbett, Fong, Lehman, & Cheng, 1987). The rules examined include statistical rules (Fong, Krantz, & Nisbett, 1986; Fong & Nisbett, in press; Kunda & Nisbett, 1986; Nisbett, Krantz, Jepson, & Kunda, 1983), as well as causal and contractual rules (Cheng & Holyoak, 1985; Cheng, Holyoak, Nisbett, & Oliver, 1986; Cheng & Nisbett, 1990; Morris & Nisbett, 1990). But these rules represent only a VOL. 1, NO. 6, NOVEMBER 1990 363 PSYCHOLOGICAL SCIENCE Teaching Rules of Choice portion of the range of possible inferential rules that people use for everyday decisions. More importantly, judgmental rules affect behavior only indirectly. In the following work, we wanted to extend the inferential rules findings to the choice domain and examine whether knowledge of inferential rules affects actual