Choosing not to Choose
by Cass R. Sunstein
Defaults matter. One of the most well-worn examples, used on the first day classes in everything from statistics to psychology, is that of organ donation. Over 98% of the populations of Austria, Belgium, France, Hungary, Poland, and Portugal are all organ donors. In Denmark, Germany, and the U.K, the percentage doesn’t top 20. Are the French simply more philanthropic than the British? Are Austrians really just better people than the Germans?
The answer is, of course, probably not. The 99%-countries all have organ donation chosen as the default, requiring citizens to explicitly opt out of the process. The others require their citizens to actually opt in.
Choosing Not to Choose explores the power and peril of defaults from the perspective of well-meaning “choice architects.” Sunstein explores the pros and cons of defaults and active choosing, the cognitive biases involved, and the promise and perils of personalized choice. He starts by arguing that defaults give us a way of “choosing not to choose,” and that this “second order decision” is vitally important. According to Sunstein, several studies indicate that the cognitive costs of choice have the most detrimental effects on the poor and stressed. At the same time, however, cognitive biases such as the endowment effect and inertia cause people to tend to stick with defaults, so it’s vitally important that the defaults don’t have their own detrimental effects.
I thought one of the most interesting parts was the section on active choosing. Active choosing, sometimes referred to as “forced choice,” is where the choice architect sets up a situation that necessitates an explicit decision between alternatives. Sunstein points out that active choices, are, in their way, as paternalistic as straightforward mandates: they withhold a desired outcome until the person capitulates and makes their choice. (I found the term ‘paternalism’ interesting in its own right, given its loaded and unacknowledged gender bias.) Active choosing isn’t always better because it usually requires either direct penalties (such as those imposed in the U.S. for not choosing health insurance) or leveraging (where a desired outcome such as a driver’s license is used to force someone to make an unrelated choice such as organ donation). Sometimes, the cognitive cost of choosing is prohibitive. For example, consider all those sites that force you to choose whether you want spam from them before you can sign up. Sure the choice is better than being defaulted into spam, but I’d much rather they defaulted me out and let me choose to receive their junk mail. Even so, they have their place, particularly when the choice architect is biased or lacking in information and defaults have the potential to cause harm. Sunstein also discusses problematic phenomenons such as “decision fatigue” and “reactance,” where peoples’ annoyance at the forced choice or even the default cause them to do the opposite of what they believe the choice architect desires. (Personally, I’m very prone to reactance.)
The last section is on personalized defaults. In describing personalized defaults, Sunstein says that “this idea may seem far-fetched, the stuff of science fiction,” but in my opinion, that’s only if you’re living in the past. Sunstein does do a good job discussing the dangers of narrowed perspectives and the potential for polarization. At the same time, he brings up an interesting point: in the case of high diversity, personalized defaults are often the best way to go.
The style is casual and conversational, and stays far on the “pop” side of the “pop science” spectrum. If you’ve taken a class in algorithmic game theory, then almost everything Sunstein says will seem rather obvious. (The entire issue can be put in terms of game theory: one must first decide upon a definition of the social welfare function, taking into consideration externalities as well as the way to summarize utility, then quantify the costs and “effort costs” of choices and the utilities of the outcomes. A choice architect is effectively trying to construct a stable coarse-correlated equilibrium.) Although interesting enough, the book isn’t all that information-heavy, and most of the book is merely a recapitulation of what Sunstein states in the first chapter. Sunstein’s argument is valuable, but I would have preferred a little more substance and a little less repetition.
My largest concern with the ideas presented was the implicit assumption that the information provided about the choices would itself be unbiased. Except for a brief section on “altering rules” (how people can change from the default) and “framing rules” (the “frames” used to convince people to opt in or opt out), Sunstein generally assumes that choice architects will be able to provide fair information about the choices. He asserts that defaults should always be “transparent and subject to scrutiny,” but I suspect that is itself close to impossible. One of the most provocative issues Sunstein discusses is the idea of “informed choice,” especially when one starts to consider the cognitive biases involved. As Sunstein points out, when one starts trying to define “informed” as the removal of cognitive biases, it’s far too easy to delude oneself that “informed” is thinking exactly like the choice architect.
Overall, if you’re curious about the issues of defaults and choice, then I think Choosing Not to Choose makes for a pleasant companion read to other pop social science books such as Kahneman’s Thinking, Fast and Slow or Watts’ Everything is Obvious Once You Know The Answer.
~~I received this ebook through Netgalley from the publisher, Oxford University Press, in exchange for my honest review.~~