Choice and choice overload

Freedom and autonomy are essential for well-being: to be happy and healthy, people need to be able to make choices for themselves and to feel that they can direct their own lives (e.g. Seligman, Reference Seligman1975; Deci & Ryan, Reference Deci and Ryan2000, Reference Deci and Ryan2002; Fischer & Boer, Reference Fischer and Boer2011; Bone et al., Reference Bone, Christensen and Williams2014). Thus, to at least some extent, the rapid proliferation of choice in nearly all domains of life (see Schwartz, Reference Schwartz2016) is a great thing, helping people satisfy their individualized preferences and feel a sense of self-determination and freedom. Yet, although there is substantial evidence that choice is good, more choice is not necessarily better, at least under some circumstances. The growing literature on choice overload has revealed that there can be too much choice, potentially leaving people worse off than they were with somewhat less choice (e.g. Iyengar & Lepper, Reference Iyengar and Lepper2000; Botti & Iyengar, Reference Botti and Iyengar2004; Iyengar et al., Reference Iyengar, Jiang, Huberman, Mitchell and Utkus2004; Botti et al., Reference Botti, Orfali and Iyengar2009; Hanoch et al., Reference Hanoch, Rice, Cummings and Wood2009; Haynes, Reference Haynes2009; Carroll et al., Reference Carroll, White and Pahl2011). Rather than thoroughly review the extensive literature on the paradoxical effects of having too much choice, we briefly highlight some key ideas in order to lay the foundation for considering how such research might be considered by policy-makers and researchers.

The research on choice overload continues to develop, and there are certainly important factors that moderate the effects of increased choice (for review and discussion, see Scheibehenne et al., Reference Scheibehenne, Greifeneder and Todd2010; Chernev et al., Reference Chernev, Böckenholt and Goodman2015; Schwartz, Reference Schwartz2016), but there is substantial agreement that large choice sets can have at least three distinct types of negative consequences. First, they can produce paralysis, overwhelming choosers so much that they fail to make any decision (e.g. Iyengar & Lepper, Reference Iyengar and Lepper2000; Iyengar et al., Reference Iyengar, Jiang, Huberman, Mitchell and Utkus2004; Carroll et al., Reference Carroll, White and Pahl2011). In a study of employee participation in corporate retirement plans, for instance, Iyengar et al. (Reference Iyengar, Jiang, Huberman, Mitchell and Utkus2004) found a 2% drop in participation rate with every 10 additional investment options offered by companies. Second, large choice sets can produce subjectively worse outcomes, magnifying regret and dissatisfaction (e.g. Iyengar & Lepper, Reference Iyengar and Lepper2000; Schwartz et al., Reference Schwartz, Ward, Monterosso, Lyubomirsky, White and Lehman2002; Botti & Iyengar, Reference Botti and Iyengar2004; Botti & Hsee, Reference Botti and Hsee2010; Chernev et al., Reference Chernev, Böckenholt and Goodman2015). Third, large choice sets can produce objectively worse outcomes, in part because the cognitive burden created by having to evaluate too many options undermines decision-making (e.g. Hanoch et al., Reference Hanoch, Rice, Cummings and Wood2009; Tanius et al., Reference Tanius, Wood, Hanoch and Rice2009; Botti & Hsee, Reference Botti and Hsee2010). For example, Tanius et al. (Reference Tanius, Wood, Hanoch and Rice2009) found that both younger and older adults chose worse (e.g. more expensive, less convenient) prescription drug plans when choosing from a larger choice set than when choosing from a smaller choice set.

The aims of choice

We think it is useful to consider the effects of increased choice in relation to the aims of choice. In our view, the aims of choice can be separated into three overlapping categories: preference satisfying, utilitarian, and self-expressive (for similar perspectives, see Shavitt, Reference Shavitt1990; Kahn et al., Reference Kahn, Dhar, Wertenbroch, Ratneshwar and Mick2005). These distinctions are not always sharp, but we believe they are useful, especially in connection with policy. Indeed, policy interventions may be more plausible and effective with regard to some aims of choice than others.

The most obvious aim of choice is to satisfy one's preferences and experience the subsequent hedonic benefits of choice. Preferences can be about anything, of course – from what food to order in a restaurant to what career to pursue – but we think that most examples of ‘preference-satisfying’ choices concern hedonic pursuits. The second aim of choice is instrumental or utilitarian, such that a choice is a means to an end (e.g. a student studies organic chemistry with the goal of becoming a doctor) or a trade-off among different alternatives.

Large choice sets seem to impede both hedonic and utilitarian aims of choice, producing both increased regret and dissatisfaction and objectively worse outcomes that less successfully achieve one's goals. Policy, however, may be able to alleviate some of these negative consequences. Although policy is likely to have little to offer with regard to people's specific preferences (de gustibus non est disputandum), policy likely has much to offer with regard to both helping people better satisfy their preferences and, especially, helping people better achieve their utilitarian goals through instrumental choice. For example, an individual may have deep concerns about the future, but go about saving for it in an ineffective way; for this person, policy interventions can likely have significant effects.

In contrast, the roles of choice overload and policy may differ when it comes to the third aim of choice: to express something about the self. People choose so as to tell the world (and themselves) who they are as people and what they value; expressive choices are statements about identity (e.g. see Hollis, Reference Hollis and Harrison1979; Bruner, Reference Bruner1990). Hollis (Reference Hollis and Harrison1979) tells the story of two American political scientists on sabbatical in a remote town in Italy who travel 3 hours to a major city so that they can cast absentee ballots in the American presidential election – one for the Republican candidate and one for the Democrat. They do this not to achieve some utilitarian objective or to exact some pleasure, but to affirm their identity as responsible citizens. We think that policy has little to offer when choice is about identity – an issue that we explore further in the conclusion of this paper – but it is nonetheless important to consider how large choice sets affect the expressive aim of choice.

Expressive choice may be the most important type of choice: ‘what should I buy?’ is of less significance than ‘who should I be?’. And it is in connection with such identity-shaping decisions that the benefits of freedom and autonomy (i.e. choice) loom largest. There is little doubt (Schwartz, Reference Schwartz2000, Reference Schwartz, Arkin, Oleson and Carroll2010, Reference Schwartz2016) that freedom of choice in self-defining domains has expanded along with freedom of choice in the world of consumer goods. Young people – at least middle-class, Western young people – find themselves with relatively unconstrained choices when it comes to where they live, what they study, what kind of work they do, what religion they practice and how they practice it, what kind of intimate relations they will enter into, what kind of family commitments they will make, and even what kind of person they will be. And having made decisions like these, people are also free to change them. No longer are people ‘stuck’ with the identities and life paths that accidents of birth, or the views of others, have imposed on them. Self-invention and reinvention are now real options – occasional paralysis in the cereal aisle of the supermarket seems a small price to pay for this kind of liberation.

There is some reason to believe that even in domains of relatively trivial choices, large choice sets raise the stakes by transforming one's decisions into statements about identity (Schwartz, Reference Schwartz, Arkin, Oleson and Carroll2010). When there are only a handful of different options for buying jeans, what jeans one buys cannot say much about who one is. There simply is not enough variety in jeans to match the variety of human identities. However, when choice sets increase in size, it becomes more plausible that what jeans one wears says something to the world about who one is (see also Kelley, Reference Kelley and Levine1967). At present, data on this possibility are sparse. Cheek (Reference Cheek2014) studied this hypothesis and found promising but inconclusive results. Cheek and Schwartz (unpublished data) actually asked participants about whether choices in various domains said something about the self, and found that in many domains, many people thought that they did, and people who felt that they had more choice in a given domain often felt that their choices said more about their identity. Thus, we are left with the possibility – albeit in need of further investigation – that large choice sets may impede the hedonic and instrumental aims of choice, but may actually enhance the expressive aims of choice. As with the other aims of choice, the value of more choice is likely non-monotonic (Aristotle, Reference Ross1999; Grant & Schwartz, Reference Grant and Schwartz2011): some choice helps people be more free to shape their identity, but more choice may make previously trivial decisions seem important and self-relevant, putting pressure on people to constantly define and redefine who they are through choice. We return to this possibility and its connections to behavioral science and policy in the conclusion to this paper.

Choice, freedom, well-being, and culture

The notion that one aim of choice is to express the self leads us to a literature that suggests that the functions and meanings of choices, particularly with regard to individual self-expression and well-being, are far from universal.Footnote ¹ The syllogism with which this paper began can seem self-evident and uncontroversial. But as the cultural research of Hazel Markus and collaborators has shown in recent years, when one leaves the familiar confines of affluent Western democratic societies, it is neither.

North American and Western European societies are fundamentally individualist in nature, and choice has become a powerful engine and manifestation of this individualism (Bellah et al., Reference Bellah, Madsen, Sullivan, Swidler and Tipton1985; Markus & Kitayama, Reference Markus and Kitayama1991).Footnote ² In core values and beliefs, in legal and political systems, in educational and care-taking practices, and even in interpersonal relationships, Western cultures reveal a particular common-sense understanding of what it means to be a person. People are understood to be independent individuals who are – or at least should try to be – free from the constraints of history, other people, and society. According to this way of thinking, each individual has his or her own private set of preferences, motives, attitudes, and goals that guide and motivate thoughts, feelings, and actions, and a good society gives people the freedom to pursue these individual motives and goals.

Western cultures thus afford a ‘disjoint’ model of agency, according to which good selves are crafted through actions that promote independence from other people and that exert individual influence on the environment (Kitayama & Uchida, Reference Kitayama, Uchida, Sorrentino, Cohen, Olson and Zanna2005; Stephens et al., Reference Stephens, Fryberg and Markus2011). Under a disjoint model of agency, because one's goals and actions are seen as deriving from internal motives, when people make choices, they express their personal goals, values, and preferences (Kim & Markus, Reference Kim and Markus1999; Kim & Sherman, Reference Kim and Sherman2008). As a result, the concepts of choice and freedom are closely linked (e.g. Markus & Schwartz, Reference Markus and Schwartz2010; Stephens et al., Reference Stephens, Fryberg and Markus2011), and choice is essential to exercising independent agency and to achieving maturity, autonomy, and well-being.

The view of the self as independent is so deeply ingrained in the Western way of life that it may be hard for Western researchers and policy-makers to see that there could be other models of how to be a good self. Yet, there are other models of selves and agency (Markus et al., Reference Markus, Mullally, Kitayama, Neisser and Jopling1997; Fiske et al., Reference Fiske, Kitayama, Markus, Nisbett, Gilbert, Fiske and Lindzey1998; Miller, Reference Miller, Murphy-Burman and Berman2003). In many collectivist societies and cultures, for example, people are not viewed as independent entities separate from others; instead, they are viewed as fundamentally interdependent or in relationships with other people. In collectivist cultural contexts, in which people define the self in relation to others, choice does not represent a route to expressing the individual self. Rather, the cultural emphasis on attending to the needs of others and adjusting accordingly shapes a ‘conjoint’ model of agency, according to which good selves are crafted not through individual choice and self-expression, but through promoting interdependence with others (Kitayama & Uchida, Reference Kitayama, Uchida, Sorrentino, Cohen, Olson and Zanna2005; Stephens et al., Reference Stephens, Fryberg and Markus2011). Under a conjoint model of agency, choice is less likely to be directly associated with freedom and well-being, and the lack of individual choice is not as threatening to autonomy or the self (Savani et al., Reference Savani, Markus and Conner2008; Chen et al., Reference Chen, Vansteenkiste, Beyers, Soenens and Van Petegem2013). In fact, people who construe the self more interdependently and understand agency conjointly may actually prefer to accept choices made by other people rather than make their own independent choices (e.g. Stephens et al., Reference Stephens, Fryberg and Markus2011).

Conjoint models of agency and interdependent self-construals are not limited to non-Western cultural contexts. Indeed, recent research has shown that even within the United States, working-class individuals are more likely to construe the self interdependently, following models of agency that share much of the emphasis on adjusting to others rather than asserting independence that characterizes Western cultural contexts (Vandello & Cohen, Reference Vandello and Cohen1999; Stephens et al., Reference Stephens, Hamedani, Markus, Bergsieker and Eloul2009). Notably, although interdependence is common in both non-Western and American working-class cultural contexts, it has different sources. Interdependence in East Asian contexts is part of the dominant discourse and is fostered by mainstream practices and institutions of the larger society, whereas the interdependence in the American working-class context seems to be primarily a function of holding a lower status within the larger social hierarchy (Triandis, Reference Triandis1989; Markus & Kitayama, Reference Markus and Kitayama1991; Lorenzi-Cioldi, Reference Lorenzi-Cioldi, Stroebe and Hewstone1998; Stephens et al., Reference Stephens, Markus and Townsend2007; Stephens et al., Reference Stephens, Hamedani, Markus, Bergsieker and Eloul2009).

The research just cited (for a more detailed review, see Markus & Schwartz, Reference Markus and Schwartz2010) underlines the fact that the connections between choice, well-being, and autonomy are not as clear cut as research conducted with people from Western, middle-class cultural contexts would suggest. Importantly, although choice is an expressive act across cultures, the specific meaning of choice can vary substantially, as can its implications for well-being. In individualist cultural contexts with a disjoint model of agency, a good self is one who acts independently on the environment and asserts individual preferences and goals; thus, extensive individual choice is essential to well-being and autonomy. In collectivist cultural contexts with a conjoint model of agency, however, individual choice is not as fundamental to crafting a good self, and thus may not be as important a path to well-being and autonomy. Accordingly, proliferation of choice may provide positive consequences for the more affluent members of Western societies, but it is unlikely to hold such positive consequences for both working-class Westerners and people from other parts of the world.

In sum, choice, freedom, autonomy, and well-being are not inextricably linked; they need not imply each other or follow logically one from the other. Hence, the opening syllogism may indeed hold, but only in the specific cultural context upon which the majority of research on choice and well-being has focused. As with many other claims made by social science, the syllogism that we take to be self-evident and universal may actually only apply to a small fraction of the world's population (see Henrich et al., Reference Henrich, Heine and Norenzayan2010).

Additional limits on preferences for choice

Cultural contexts and models of agency and the self represent important caveats to the view that increasing the amount of available choice follows people's preferences and increases their well-being, but there are further limits on the extent to which even people who mainly construe the self independently want extensive choice. For instance, older adults want substantially less choice than young adults (Reed et al., Reference Reed, Mikels and Simon2008, Reference Reed, Mikels and Löckenhoff2013), and young adults mainly want choice in domains in which they feel confident that they can make successful choices (Reed et al., Reference Reed, Mikels and Löckenhoff2012; Patall et al., Reference Patall, Sylvester and Han2014). People also report wanting more choice when thinking about decisions in the abstract, but when actually making decisions, they often prefer having fewer options from which to choose, particularly when considering the feasibility of making a choice (Goodman & Malkoc, Reference Goodman and Malkoc2012; Ackerman et al., Reference Ackerman, Gross and Celly2014).

The type and aims of a given choice can also influence how many options people want. Reed et al. (Reference Reed, Mikels and Simon2008) found that people want more choice in mundane, everyday domains (e.g. flavors of jam) than in the more consequential domain of health care decisions (e.g. prescription drug plans). People are also often more likely to want a larger variety of options when making hedonic or self-expressive choices compared to more utilitarian choices (e.g. Ariely & Levav, Reference Ariely and Levav2000; Ratner & Kahn, Reference Ratner and Kahn2002; Kim & Drolet, Reference Kim and Drolet2003; Rozin et al., Reference Rozin, Fischler, Shields and Masson2006). For hedonic and self-expressive choices, the act of making a choice in itself can serve an important function (e.g. Hollis, Reference Hollis and Harrison1979; Choi & Fishbach, Reference Choi and Fishbach2011), but because utilitarian choices are more focused on the outcome rather than the process of choice, people do not feel the need for as much choice. Indeed, people may be happy to have someone else make such decisions, provided the outcome is satisfactory (Botti & McGill, Reference Botti and McGill2011). Furthermore, a recent meta-analysis on choice overload (Chernev et al., Reference Chernev, Böckenholt and Goodman2015) found that people are more likely to find large choice sets unappealing and overwhelming when they have little familiarity with or knowledge about the choice domain, when the choice is relatively complex and difficult, and when they are relatively uncertain about their preferences.

Thus, taken together, previous research suggests that even people for whom choice may be an important path to self-expression and well-being sometimes prefer to choose from relatively small choice sets. This may be particularly true when choices are difficult, complex, instrumental, and utilitarian, and when people feel unconfident about their ability to make a good decision by themselves. Importantly, these descriptions also apply to many of the choices that are the focus of policy-makers’ research and recommendations, such as retirement plans, medical insurance plans, and investment and savings plans. Accordingly, we now turn to some of the implications of the research reviewed above for public policy.

Choice and public policy

As mentioned in the ‘Introduction’ section, our goal is not to make specific policy proposals or to take explicit sides in the debate regarding paternalist – libertarian or otherwise – policies. Rather, our goal is to consider the connections between choice, freedom, and well-being that have been taken for granted in much of behavioral research and public and private policies. In this section, we outline some implications of the research reviewed above that we view as particularly important for public policy decisions, although we do not attempt to make any specific policy recommendations ourselves.

Public policy in the United States and other Western countries often seems to follow from, as phrased by Botti and Iyengar (Reference Botti and Iyengar2006), “the conviction that social welfare can be pursued through the provision and exercise of choice” (p. 24). There is a wide variety of arguments in opposition to and in favor of providing – and perhaps requiring – a substantial amount of choice for individuals, and a thorough discussion of each is outside the scope of the current paper (for a discussion of many of these arguments, see Sunstein, Reference Sunstein2015). Here, we limit ourselves to proposing five questions that policy-makers may want to consider more thoroughly before deciding that constantly providing individuals with more choices and options from which to choose is the best way to improve decision quality, well-being, and freedom. Our main goal is to help explicitly identify potentially problematic assumptions that often implicitly motivate policies and policy recommendations.

The first question is: do people want more choice? At first glance, much of the research on choice seems to suggest that people indeed love choice and almost always prefer more choice over less, even if receiving more choice leads them to worse outcomes (e.g. Botti & Hsee, Reference Botti and Hsee2010). Yet, we have reviewed research highlighting at least two important points: first, the previously assumed widespread appeal of choice may be limited to young, middle- and upper-class Westerners; and second, even within cultural contexts that place high value on individual choice, people do not necessarily always prefer to have more choice. Even American college students – perhaps the most choice-oriented people in the world – may be happier with fewer options when choices are unfamiliar, complicated, and not self-expressive. Choice of medical insurance plan – which people may feel very uncertain about, says relatively little about one's identify, and is extremely complex – is an example of domain in which it may not be wise to take for granted that people actually want more – or perhaps any – choice (e.g. Botti & McGill, Reference Botti and McGill2011; Sunstein, Reference Sunstein2015). Furthermore, because people may report preferences for different amounts of choice depending on the context in which they are asked about their preferences (e.g. they may prefer less choice if asked right before they make a choice versus far in advance of a possible choice; Ackerman et al., Reference Ackerman, Gross and Celly2014), answering this first question may be more complicated than simply asking individuals how much choice they hypothetically desire.

The second question is: will more choice improve the quality of decisions? From a traditional rational choice theory perspective, providing more individual choice will facilitate utility maximization and preference matching, and people will make better choices for themselves than others would make for them. However, although there is much work left to be done, the research on choice overload suggests that, in some cases, people may make worse decisions when they choose from larger choice sets. In studies on Medicare drug plan choice, for instance, older adults often choose objectively worse plans for themselves when making choices from larger choice sets. People also often feel worse about their decisions from large choice sets, even if the choices they make objectively meet their criteria. Moreover, if having too many options creates choice paralysis, people might not make any decision, and in cases like insurance and retirement plans, having a suboptimal plan may be much better than not having one at all.

The third question is: will more choice improve well-being? Well-being is a complex concept, but if people feel less happy and frequently make worse decisions that impair their health and financial status, it seems unlikely that well-being has improved, and research on choice paralysis and choice overload suggests that when choice is overwhelming, people can both make worse decisions for their present and future well-being (e.g. failing to save for retirement) and feel more regret and self-blame about decisions that prove less satisfying (Schwartz, Reference Schwartz2016). Moreover, well-being may not even be closely related to abundant choice for some people, such as those from collectivist cultural contexts. Thus, the path from choice to well-being that seems so clear in the syllogism with which this paper began may often require more careful reflection accompanied by consideration of empirical evidence.

The fourth question is: will more choice impose too much of a burden? When people have to make more choices or choose from among more options, the cognitive cost of decision-making increases. Some people may be able to absorb this increased cost fairly easily, but for others it may outweigh the potential benefits of increased choice. Recent research on poverty, for example, suggests that the material scarcity experienced by poor people imposes a serious cognitive burden (Mani et al., Reference Mani, Mullainathan, Shafir and Zhao2013; Mullainathan & Shafir, Reference Mullainathan and Shafir2013), and as a result, increasing choice may place an even larger burden on people who are in the most vulnerable situations and would suffer most from the consequences of choice paralysis. Increases in choice that seem straightforwardly positive to people in comfortable financial positions may therefore not have such simple consequences for people who are already struggling to make ends meet.

The final and perhaps most complex question is: will more choice make people more free? This question depends on, among other things, one's definition of freedom, and if freedom is defined in such a way as to require constant individual choice, then providing more choice may be seen as increasing freedom, although perhaps to the detriment of decision quality and well-being. Although we will not address the nuanced debate on the meaning of freedom here, we believe this question is not as straightforward as it appears, particularly in situations in which people may not want to have to choose for themselves or to choose from so many options. Indeed, might requiring people to make choices when they do not want to, even in the name of freedom, itself be a form of liberty-threatening paternalism (e.g. see Sunstein, Reference Sunstein2015)? The debate on the meaning of freedom also highlights that not all people and cultures view freedom in the same way, and, as we reviewed earlier, whether or not freedom is even closely connected to choice can depend on an individual's cultural context and model of agency. At the very least, we believe that this question and the other four we have proposed deserve more explicit and careful consideration than they sometimes receive, and rather than taking one or more of the answers for granted, policy-makers and researchers might do well to consider explicitly what we know about the sometimes tenuous links between choice, well-being, and freedom.

Conclusion: the role of public policy and behavioral science

The central goal of this paper has been to examine the growing body of research that questions the assumption that promoting choice and freedom is always the same as promoting well-being and welfare. In our view, carefully re-evaluating the presumed relationships between choice, freedom, autonomy, and well-being has the potential to greatly benefit policy, giving policy, in turn, greater potential to benefit well-being. Yet it may also be worth considering the limits to what policy can accomplish, and how those limits might influence how we think about the role of behavioral science.

We suggested earlier that increased choice may increase the expressive value of even mundane choices, and that increased freedom provides the opportunity – and perhaps the requirement – that people define the self through choice. Schwartz (Reference Schwartz, Arkin, Oleson and Carroll2010) theorized that the increased pressure on individual self-definition (see also Taylor, Reference Taylor1989, Reference Taylor1992a, Reference Taylor1992b) as a result of changing views of and opportunities for choice and freedom may be the most negative consequence of having too much choice, focusing attention on material aspects of identity and distracting from the social relationships and more meaningful pursuits in life that might make people happier and healthier. Policy may be unlikely to offer much help to individuals who are struggling under the psychological weight of too much choice and freedom in many (including private) domains of life without also compromising the liberating, life-enhancing consequences of increased freedom. And if policy is limited, what else can be done? One possibility is to more consciously spread some of the insights from behavioral science to the public, perhaps raising awareness and influencing cultural attitudes about choice, agency, and freedom. We might, for example, encourage people to engage in more second-order decision-making, getting people to think more carefully about when it is worth spending the time and effort to make a careful choice and when it is not.

Might sharing some of the science reviewed in this paper be a useful step toward reducing the cultural pressure placed on choice as a result of increasing individualism and autonomy (Schwartz, Reference Schwartz2000, Reference Schwartz, Arkin, Oleson and Carroll2010)? It might, but doing so may also shift the traditional role of behavioral science from solely studying people and cultures to (purposefully) influencing people and cultures, and that might not be a shift we are comfortable making. For now, we close by raising the possibility that, in a journal dedicated to exploring how behavioral science can help public policy, it might sometimes be worthwhile to explore how behavioral science can help people outside of policy as well.