Consider the following. A young child, having received a playhouse as a gift and finding its interior too dark, asks her mother how she makes their house so light. “By flipping a switch,” says the mother. The child finds a spare switch in the basement, hangs it on the playhouse wall, and flips it, but gets no light. How charming is the innocence of young children. And how oblivious adults can be to the background conditions that are necessary to make “flipping the light switch” give us light.
The upshot of part three of this essay, and the argument in Choose Wisely, is that someone who asks how to make a good decision and is told to use rational choice theory (RCT)—to quantify the options and attributes, the probabilities and values, and to calculate—is like the child who is told that light comes from flipping a switch. The switch works only if it is connected to the house wiring system, which in turn is connected to the utility’s wiring system, which in turn is connected to an extremely complicated electricity-generating system, which in turn is energized by some sort of fuel. Similarly, RCT works well only if the decision problem is framed well; if the options and attributes are specified well, formulated in quantifiable terms, and the probabilities and values are quantified well. Typical decision problems do not come in that form. They need to be put in that form by a series of substitutions for the original amorphous form they actually come in. These substitutions replace the original problem, step by step, with a version that RCT can handle. These substitutions crucially involve framing, and the decisions we make, with or without RCT, will only be as rational and good as their framing is.
RCT itself has little to contribute to that process of framing. It can tell us to transform the decision we face into something like a casino gamble, but not whether, in doing so, we have preserved the actual character of the decision. The gamble is a paradigm of RCT, in the sense of being an exemplary case. It is also a paradigm in the sense used by philosopher Thomas Kuhn in his landmark book The Structure of Scientific Revolutions, in which he suggests that scientific paradigms establish both the problems to be solved and the methods of inquiry to be used in formulating a solution to those problems. The gambling paradigm strongly determines a form of process that is largely formal and quantitative. But the framing of decisions cannot be accomplished by quantitative and formal methods. Formal methods do not—and more important, cannot—tell us how to frame a problem well, how to specify options and attributes, how to formulate the options and attributes as measurable, or how to quantify the relevant probabilities and values. Nor can formal methods provide a criterion for when we have framed a problem well. And the decision is no better than the framing of the problem allows. Of course, once all the framing is done, solving the problem requires only mathematical calculation, just as bringing light to the house requires only flipping the switch. But to credit flipping the switch with lighting the house is extremely misleading. Framing a problem requires deliberation, a decidedly nonformal process, just as generating and transmitting electricity has little in common with flipping a switch.
To light a room, the light switch must be connected to the power grid. Most of what Schuldenfrei and I tried to do in the book was to spell out how the power grid of rationality works, and what it requires.
* * *
Good judgment
We propose that an adequate account of rationality must replace RCT-type calculation with judgment—replace counting with thinking. Which are the right colleges or jobs to apply to is not a matter of maximizing something, or a decision that can be made by formula: It is a matter of judging what is a good subset of appropriate schools or jobs given a decision-maker’s purposes, and of judging the quality of those purposes themselves. The same is true of most other significant decisions in life.
Imagine that, having just graduated from college, you are offered six different jobs in your field as a management consultant. The jobs vary in a host of respects: starting salary and benefits, location, size of the firm, opportunities for advancement, attractiveness of colleagues as potential collaborators and friends, and the nature of the work you will be doing. Each of these features of the jobs (and no doubt there are others) can itself be decomposed into sub-features. Take location. What is the cost of living in the area? How close is it to family and friends? What about housing and commuting? Restaurants and nightlife? Which job to take is a complex and consequential decision indeed—one that may cast a long shadow into your future.
RCT offers us a way to make such decisions. You might create a spreadsheet. Across the top are columns for each of the features of prospective jobs that matter to you. Below that are columns for the relevant subfeatures that matter. For each of these many columns, you need to assign three numbers. First, how important is this feature or sub-feature to you, say on a 10-point scale? Second, how good or valuable is each job you’ve been offered on each dimension you care about, again on a 10-point scale? And finally, what is the likelihood that each feature you are evaluating will deliver the goods (or bads) that you are expecting? Every decision is a prediction—not only about what will happen, but also about how what happens will make you feel.
It’s a lot of work, but it’s an important decision. The virtue of using RCT in this way is that it may encourage more careful examination of features of various jobs that are important to you. It may also protect you from allowing preconceptions and biases from putting their fingers on the scale. In any case, if you do your due diligence and fill out this spreadsheet, it becomes a simple matter to calculate which is the best job. Push a key on your computer, let Excel do its calculations, and voilà, you know which job to take.
An adequate account of rationality must replace RCT-type calculation with judgment—replace counting with thinking.
You could do the same sort of analysis to decide which college to attend, which discipline to major in, which career to pursue, whether (and whom) to marry, and whether (and when) to have children. And you could do it for more trivial decisions, like where to go on vacation, what restaurant to eat in, and the like. It is, one might say, a precise and objective way to calculate what is essentially a subjective quantity—how much satisfaction (utility) each option is likely to deliver.
I believe, however, that the precision apparently offered by RCT is an illusion. Virtually every number you enter into the spreadsheet requires a significant amount of judgment. It is, at best, a rough estimate about how each job is likely to unfold for you, how important each feature of the jobs will be to you, and how much, in what ways, you will change as you work at the job and mature as a person. In addition, the job you take will have effects on the lives of people who matter to you. How much, and in what ways, should that enter into your calculations? There may also be moral dimensions to your work in that it will have effects on clients and customers. Will you be contributing to social welfare or impairing it? And how much should that matter? Finally (well, not really finally, since the dimensions of this decision are endless), the job you choose may affect other aspects of your life that you care about. A great job whose demands leak into other important features of your life won’t be such a great job.
And this point illustrates what is perhaps the greatest deficiency in the RCT approach to decisions like this. It claims to substitute calculation for judgment. Remember, for each feature of the jobs you are considering, you have to enter a number that represents how good or valuable that feature is. What, exactly, do “good” and “valuable” mean? Location is not valuable in the same way that salary is. Salary is not valuable in the same way that good colleagues are. Good colleagues are not valuable in the same way that work you care about is. Each of these different dimensions of each job likely provides not just a different amount of value but a different kind of value. If so, how can you sum scores across columns and arrive at a grand total for each job? You can’t. RCT provides an abstract term—utility—to capture value. It thus requires you to translate financial, social, moral, and intellectual values that may be reflected in your spreadsheet into the common currency of utility. Does that make sense? I think not.
Creating an RCT-type spreadsheet has its value. It may force you to think more broadly and carefully about many aspects of a decision than you otherwise would. But that virtue is not quantitative. It exists before you enter a single number estimating value or probability into the spreadsheet. The spreadsheet helps you to avoid overlooking something important. But having done that, it is time to substitute judgment and reflection for calculation and thus avoid the false precision that using a spreadsheet encourages.
* * *
Practical wisdom
Aristotle taught us that many, perhaps most, on-the-ground decisions require judgment—what he called practical wisdom. The particulars of a given situation are crucial: Context always matters. Context influences how we should balance our obligations to family and friends with our own opportunities. It influences how differently we should treat each of our kids, or our students, each of whom needs different things. The answer to questions we ask ourselves about issues like these is, almost always, “It depends.” The right thing to do with one person at one point in time may be a catastrophe with another person at another point in time. In the book Practical Wisdom, written with Kenneth Sharpe, I argued that in almost every part of life we care about—work, education, friendship, parenting, politics—when we face decisions, the right answer is usually, “It depends.” No formula substitutes for judgment. A formula, or a rule, is like a road map with enough resolution to distinguish various cities and towns, but not enough to distinguish streets. Such a map may get us to the right city, but not the right address in that city. Finding the city provides a frame within which locating the address becomes possible.
Why does the importance of good judgment constitute a criticism of RCT? I believe that to exercise good judgment reliably, one must cultivate qualities of mind that are most essential to good judgment and good decisions. These are: understanding, reflectiveness, self-knowledge, and values. When RCT leaves all these attributes of rational thinking out, or simply presupposes them, it discourages the cultivation of exactly the qualities we need most.
The kinds of decision problems people are posed in the laboratory, though they come within frames, come within very limited frames. By adding context to the situations, one changes the frames and thus also the character and complexity of the decisions we face. By keeping background information skeletal, researchers make decision problems seem more similar to one another than they really are, and more simple than they really are. In consequence, aspects of thinking like meaning and understanding sink to the background, seemingly irrelevant to the problem at hand.
Our proposed alternative to RCT does not take the form of a formal procedure or anything approximating or modeled on one. Our alternative is based on the notion that any action we take needs to be understood as parts of whole lives, and that a given decision, if it is an important one, has to be made largely on the basis of how it fits into a whole life. Decisions are not, and should not, be made in isolation. We believe the best sort of life is (among other things, and all other things being equal) a life of narrative unity and purpose—a life with worthy goals that, to the best of our ability, we articulate as we make progress toward them. It is a life that is appropriately unified (not obsessively limited) by those goals or purposes. We can abbreviate this desideratum as calling for a meaningful life.
To exercise good judgment reliably, one must cultivate qualities of mind that are most essential to good judgment and good decisions. . . . RCT discourages the cultivation of exactly the qualities we need most.
We think understanding, reflectiveness, and self-knowledge are essential ingredients in a meaningful life. They help us place perspicuous frames around our experiences, which in turn enables us to assess their current and future significance. They help us appreciate the radical uncertainty of many events in the world, which in turn helps us to maintain a flexible and adaptable stance toward the future. They also help us appreciate the inherent ambiguity of many experiences, opening us up to the interpretations and decisions of others. And they help us articulate the values we want to live by, and then to assess how the decisions we face may impact those values. If we lived in a world in which framing is unneeded or to be avoided, radical uncertainty does not exist, ambiguity can be eliminated, and diverse values can all be reduced to utilities, then understanding, reflectiveness, and self-knowledge may not be needed. But that is not the kind of world we live in—or would want to live in.
And a whole life itself has to be evaluated above and beyond the evaluation of the individual decisions that compose it. Judgment about a whole life is not a yes or no, good or bad matter. But having some ideals in mind can facilitate our assessment of our lives in something like the way that geometry helps us understand the physical world. There are no objects in the world that are perfect geometrical shapes. Nonetheless, models from geometry put us in the right ballpark. It’s a great start, but it must be reconciled with the empirical facts on the ground. Thus, the process of thinking we envision is one that shuttles back and forth between the ideal and the real—between the simplified formalisms of a discipline like geometry and the bumps and ridges of lived reality. RCT is missing this back-and-forth. It impoverishes decisions by analogizing them to gambles and stops there, rather than renormalizing them.
We think a similar point can be made about the narrow framing of decisions that enables RCT to be used to make them. It takes judgment to know when and how to frame the decision context, and when and how to change the frame. Often, deliberation about a choice between two options can and perhaps should lead to the realization of a hitherto neglected third option. Perhaps the two original options, on examination, are both inadequate, a discovery that “forces” us to open things up and consider new alternatives.
We can easily imagine something similar happening when the whole RCT process is completed. Suppose the process yields a decision for an option that, looked at freshly, seems simply unacceptable. Is it irrational to simply say, “No. There must have been something wrong with the process that led up to the calculation”? This is similar to rejecting a hypothesis when it leads to a false prediction. Is that not rational? A conclusion like this has no explicit role in RCT, but it should have a role in rational decision-making. Rejecting such reasoning is a least partly the effect of the (false) notion that the real work in deciding is in the calculation, not the thinking that surrounds the calculation. It is, in effect, an argument that we should be seeking reasonableness, not formal, quantifiable rationality as we make our decisions and live our lives.
* * *
Conclusion
There are limits to how much the arguments in an entire book can be captured by a summary—even an extended one like this. Schuldenfrei and I urge you to look at the whole book if the arguments presented here have managed to pique your curiosity. We don’t expect our book to be the death knell of RCT. RCT certainly has and should have its place. But that place is not every place. We hope that our book will stimulate thoughtful conversations about where RCT belongs, and where it doesn’t. And we hope that when you face a decision about what to do on a beautiful Saturday, or throughout a beautiful but complicated life, you will resist the temptation to resort to oversimplified quantification. Quantification can turn any decision into a “no-brainer.” But making decisions is what brains are for.
Adapted from Choose Wisely By Barry Schwartz and Richard Schuldenfrei. Published by Yale University Press. Copyright © 2025 by Barry Schwartz and Richard Schuldenfrei. All rights reserved.
Disclosure: Barry Schwartz is a member of the Behavioral Scientist advisory board. Advisors do not play a role in the editorial decisions of the magazine.
