4

The Behavioral Economics Toolkit: Policy Levers and Intervention Strategies

The theoretical foundations of behavioral economics discussed in Chapter 3 have been used to design policies aimed at changing behavior. An outstanding question, however, is how these principles are translated into interventions and behavior change strategies, including so-called nudges, in real-world settings.

A nudge is defined in the context of behavioral economics as a low-cost, light-touch change in choice architecture—the structures and contexts within which and through which a choice is presented—that shifts people’s behavior without explicitly regulating it and without imposing significant (financial) rewards or punishments (Thaler & Sunstein, 2009). Multiple elements of choice architecture can be altered: the number of options, the framing of options, the information provided about the options, the placement of options in a physical or digital space, and the way people’s attention is drawn to some options (or features of options) more than others (Münscher, Vetter, & Scheuerle, 2016; Johnson, 2021). Many different aspects of choice architecture have been targeted using nudge strategies: they include changing default options; providing information about the behavior of peers or neighbors; simplifying forms or instructions for decision making; and a variety of ways to increase the salience of the desired option, such as changing the order or convenience of options in a decision menu. This approach to behavioral change is sometimes called “libertarian paternalism,” in that the designer of the nudge has determined that one choice would be better than another—would benefit individuals or society—but is not imposing that choice through regulation or other requirements (see Chapter 2).

Researchers, policy makers, and practitioners in many domains have designed and tested interventions that leverage these principles, efforts that have been documented in decades of research. Interventions are planned and coordinated sets of activities designed to change specific behaviors or behavior patterns (Michie, Van Stralen, & West, 2011). Behaviorally informed interventions are those that have been selected or designed specifically to address a behavioral barrier, biased tendency, or decision error that can be explained by the foundational principles of behavioral economics discussed in Chapter 3.¹

Interventions can be designed and implemented by different actors (corporate, governmental, institutional) in different policy domains (e.g., health, education, finance) and take diverse forms (e.g., the design of a promotional brochure or reminder letter, a policy about enrollment or late fees, or a physical change to a choice environment). Interventions can be targeted at the general population or at specific subgroups of interest, may comprise one-time or repeated exposures, and may range from very obvious to quite invisible.

In this chapter we review a necessarily selected group of these intervention strategies or approaches. We focus on the strategies that are most commonly used and for which there is some evidence of effectiveness. We note here that some of these strategies work by leveraging the common cognitive biases and heuristics that affect people’s decision making, such as present bias, availability heuristic, and loss aversion (see Chapters 2 and 3); others work by countering, mitigating, or decreasing the effects of those same biases and heuristics. This latter type can be referred to as a way of de-biasing individuals or counteracting specific sources of bias. We note that the research on the effectiveness of de-biasing is just beginning, and its findings to date are inconclusive (Borchardt, Kamzabek, & Lovallo, 2022).

In addition to drawing heavily on psychology, many of the intervention strategies discussed below also rely on theories and practices from communication, education, and human-centered design² to most effectively either leverage or counter the impact of behavioral factors, such as reference dependence and present bias, on behavior. The emerging interdisciplinary field of behavioral design brings together the findings and methods from all of these fields to optimize intervention designs as effective solutions

___________________

¹ The committee uses the phrase “behaviorally informed” to distinguish between interventions that are specifically designed to address a behavioral attribute and other, possibly very similar, interventions that were not. For example, incentives are recognized as key factors in economic decisions, but, while in some cases they are straightforward (e.g., offering financial rewards for taking an action), in others they are designed to address behavioral factors such as those discussed in Chapter 3.

² Human-centered design is an approach to problem-solving that focuses on users’ perspectives, emotions, and needs in the design of products and services.

to behavior change challenges (Datta & Mullainathan, 2014; Niedderer, Clune, & Ludden, 2017; Bucher, 2020). Indeed, behavioral design of markets and institutions is also emerging as an important area of research on the optimal design of institutions—such as matching markets for medical interns (Roth & Sotomayor, 1992; Roth, 2015) and school choice mechanisms (Pathak, 2017)—that incorporate behavioral elements (also see McFadden, 2009; Duflo, 2017). Similarly, behavioral public policy and behavioral governance focus attention on the ways in which policy makers themselves may be subject to biases and heuristic thinking in the policy formulation process and on how to productively incorporate behavioral insights into that process (see Chapter 13; Grimmelikhuijsen et al., 2017; Gofen et al., 2021).

As we discuss in Part III, more research is needed on many aspects of these kinds of interventions, including studies of their efficacy and effectiveness; implementation; replication; subgroup analyses and explorations of differential responses; and ways to ensure the best fit between intervention strategy and target, context, and behavior type. For all of the strategies we discuss, the evidence is mixed: it is possible to find studies with null effects, failures to replicate, or nuanced results suggesting subgroup effects and contextually specific results, as well as evidence of positive effects. For example, it is likely that some behaviorally informed interventions are more effective for changing a one-off behavior than for creating and sustaining a habit (Neal et al., 2016; Wood & Neal, 2016; Carden & Wood, 2018; Venema et al., 2020). Similarly, while there are settings and situations in which individual behavior decisions are made in isolation, many other settings are characterized by feedback loops, “gaming,” or other interpersonal and networked behaviors that may influence the effectiveness or effects of an individually targeted intervention. More research and discussion are also needed on the limitations of behaviorally informed interventions (especially for the narrow nudge form) and on how focusing on individual behavior change could preclude or crowd out focus on structural or system change (Loewenstein & Chater, 2017; Chater & Loewenstein, 2022).

Although there is much more to learn, the work to date has produced an impressive toolkit for influencing behavior in a policy context. While different fields use different terms for the contents of a toolkit (tools, strategies, approaches), we refer to them as “intervention strategies.” For simplicity, we have chosen to align the intervention strategies covered in this chapter with the foundational principles discussed in Chapter 3. However, most successful intervention strategies take into account more than one of these principles. Our mapping of strategies to the foundational principles is shown in Table 4-1. The rest of the chapter discusses some of the leading intervention strategies that have been developed applying these principles.

TABLE 4-1 Intervention Strategies Mapped to the Five Foundational Principles

ADDRESSING LIMITED ATTENTION AND COGNITION

To address limited attention and cognition, the behavioral challenge for policy makers is that individuals pay only limited attention to important aspects of policy situations, often have a hard time processing information, and make cognitive errors even in simple situations. To meet this challenge, interventions have used defaults, active choice, salience primes, simplification, choice sets, and creating slack.

Defaults

Setting a default (e.g., a default level of salary deduction for a retirement plan, or a healthy default side dish on a restaurant menu) is among the most powerful strategies for behavioral change. Defaults have been shown in several systematic reviews and meta-analyses to have larger effect sizes than other intervention approaches. Defaults work through multiple mechanisms but are particularly effective in conditions of limited attentional or cognitive bandwidth, when individuals cannot or will not take the time to evaluate options and make a fully informed choice (Johnson & Goldstein, 2003). Examples of effective defaults include opt-out or opt-in 401(k) retirement plan enrollments for employees (Madrian & Shea, 2001; see Chapter 10), defaults for prescribing generic medications (Patel et al., 2016b), and default menu options to encourage healthy food choices (Vecchio & Cavallo, 2019).

Active Choice

When it is not possible (or perceived as not ethical) to implement a true default, an active choice intervention can also focus limited attention. Active choice interventions create a stopping point in a process that requires an individual to make a choice before proceeding. A common one is the requirement to elect or decline travel insurance when purchasing a plane or train ticket. A study of HIV testing decisions found that an active choice intervention resulted in testing rates between those of opt-out (default) and opt-in interventions (Montoy, Dow, & Kaplan, 2016).

Salience Primes

When inattention and cognitive load create barriers to behavior change, it can be an effective strategy to make crucial information or features of a choice set more salient (see Chapter 3). Salience prime interventions have been shown to be effective in many domains, as shown in two examples discussed in Chapter 3: the study showing that grocery store price tags that

made state taxes more salient significantly decreased purchases for taxable products (compared with price tags that did include taxes; Chetty, Looney, & Kroft, 2009), and the study showing the importance of providing clear descriptions of the options included in Medicare Part D plans (Kling et al., 2012).

Simplification

As discussed in Chapter 3, interventions that simplify the presentation of information allow people’s limited attentional resources to be directed at a smaller amount of information. In a field experiment conducted in cooperation with the Internal Revenue Service (IRS), different versions of letters were sent to 35,000 taxpayers who filed returns and were eligible for the Earned Income Tax Credit but had not yet claimed it (Bhargava & Manoli, 2015). Remarkably, just a simplification of the two-page notice sent from the IRS with clearer wording increased the take-up in this group by over 10 percent, suggesting the importance of clear communication about government programs.

One popular form of simplification that also reduces cognitive burden and can nudge individuals toward a focused set of tasks in a set order is a checklist. A systematic review of checklists used to improve health care delivery quality found significant reductions in surgical complications and medication errors associated with checklists (Boyd, Wu, & Stelfox, 2017).

Choice Sets

Limited attention for choices can also be focused through the creation, arrangement, and framing through choice sets (Johnson, 2021). Reducing the number of choices available has been shown to improve the selection process and increase people’s satisfaction with their choices (Gourville & Soman, 2005; Reutskaja et al., 2020, 2022). Studies of health insurance plan choice have highlighted how plan choice can be strongly influenced by how different plan options are described and displayed, often leading to suboptimal choices (Bhargava, Loewenstein, & Sydnor, 2017). Choice sets can be designed to reduce comparison friction—the cognitive burden caused by having to compare choices across multiple attributes without sufficient information—or other psychological frictions that impede accessing information or taking up benefits (Kling et al., 2012; Bhargava & Manoli, 2015).

Creating Slack

A final category of intervention strategies to counteract limited attention and cognition is the creation of slack: that is, providing more space

or flexibility for people’s responses. Creating slack is particularly effective for those experiencing scarcity of any kind (time, material resources, even social interaction), because scarcity can make decisions about immediate tasks more costly and challenging, while also amplifying the negative consequences of poor or short-sighted decisions (Mullainathan & Shafir, 2013). Techniques to create slack include introducing buffers of extra time or extra financial resources and reducing the cost of errors or poor decisions by establishing safeguards and backstops or otherwise “replacing cliffs with slopes” (Daminger et al., 2015, pp. 34–35).

ADDRESSING INACCURATE BELIEFS

To address inaccurate beliefs, the behavioral challenge for policy makers is that people often have inaccurate perceptions of the situation, of incentives, of their own abilities, and about the beliefs of others. Intervention strategies that address inaccurate beliefs or perceptions can replace or update inaccurate beliefs with new beliefs or reduce the effect of the inaccurate beliefs on optimal decision making. Several of the strategies mentioned above are relevant here, including framing, defaults, reference groups, simplification (e.g., of complex incentive schemes), choice sets, peer feedback, and social proof.

In addition to those interventions, two others are important for meeting the challenge of inaccurate beliefs: de-biasing and mental models, which are people’s representations of the world.

De-biasing

The primary type of intervention for correcting beliefs is de-biasing. For example, an overoptimism bias—the tendency for a person to believe they will perform better than the average or that good things will be more likely to come their way and bad things less likely—can be addressed through salient presentation of statistics related to past performance of either an individual or a relevant reference group. In a systematic review of interventions to de-bias health-related decisions, two-thirds of the studies designed to counter overoptimism bias were successful (Ludolph & Schulz, 2018). De-biasing approaches have been categorized into motivational, cognitive, affective, and technological strategies, with emerging research showing nuanced evidence about which approaches are the most successful in different contexts (Larrick, 2004; Larrick & Soll, 2008; Croskerry, Singhal, & Mamede, 2013; Soll, Milkman, & Payne, 2015; Delavande, 2023; Fuster & Zafar, 2023; Haaland, Roth, & Wohlfart, 2023).

Mental Models

A large literature on mental models suggests that people’s mental representations of how the world works can strongly influence decisions and behaviors in diverse domains, including health, climate change, and management (Johnson-Laird, 2004; Pfeffer, 2005; Kealey & Berkman, 2010; Wong-Parodi & Bruine de Bruin, 2017). For example, multiple studies have shown that people have powerful mental models of the epidemiology of COVID-19 and of the effectiveness and acceptability of mitigation strategies, including masking, social distancing, and vaccination (Southwell et al., 2020; Greenhalgh, 2021; Berg et al., 2022; de Ridder et al., 2022). While mental models can be useful decision-making heuristics, they can also induce inertia and are susceptible to bias (Guiette & Vandenbempt, 2013; Hovmand et al., 2021). Biased or inaccurate mental models can be a target for interventions, including de-biasing interventions, although research suggests that replacing or updating mental models can be challenging. Techniques to revise or update mental models include visualizations, experiencing surprise, analogies, simulation, and critical reflection (Bostrom, 2008; Gary & Wood, 2011; Thacker & Sinatra, 2019; Vink et al., 2019).

ADDRESSING PRESENT BIAS

To address present bias, the behavioral challenge for policy makers is that people tend to disproportionately focus on present consumption, payments, and general utility, and pay less attention to future payoffs and consequences. Present bias creates a tendency to procrastinate and to choose immediate rewards over longer-term benefits.

Whether this tendency is perceived as laziness or hyperbolic discounting, there are several evidence-based strategies to counter present bias. Disregard of future benefits is likely related to inattention, as described above: that is, individuals simply are not paying attention to future benefits. For this reason, many of the strategies discussed above as remedies to limited attention and cognition are also applicable here, particularly defaults and simplification. Defaults and simplification make it harder for an individual to procrastinate by reducing options and limiting overwhelming and distracting information. Interestingly, however, even when attention is focused on future benefits (e.g., employer matches for retirement savings), there may be minimal take-up of the future-oriented benefit in the moment without an opportunity to act right away and a deadline (Choi, Laibson, & Madrian, 2011).

In addition to these strategies, five others are particularly appropriate to meet the challenge of present bias: reducing friction and hassle factors,

behaviorally informed financial or nonfinancial incentives, lottery-based incentives, commitment devices, and reminders.

Friction and Hassle Factors

Removing friction reduces people’s tendency to procrastinate by reducing the immediate costs (in the form of time or inconvenience) of a behavior. Researchers across diverse domains have studied the response to strategies or interventions that remove sources of friction (also called “hassle factors”), reduce administrative burdens, or reduce switching costs. Because present bias leads individuals to delay action when they experience friction, hassle factors, or administrative burden, reducing those factors can substantially increase compliance. Examples of such interventions include providing individualized information about a benefit so that participants do not have to seek that information themselves (Kling et al., 2012), transitioning cumbersome paper-based benefits administration to electronic formats (Vasan et al., 2021), and allowing users to complete an action online instead of by mail or in person (Bhanot, 2021).

The highly successful Moving to Opportunity (MTO) program, which provides an opportunity for low-income families to move to better neighborhoods using a voucher, demonstrates the importance of hassle factors. Despite robust evidence on the effectiveness of the program, take-up is low. In a subsample of MTO recipients who were randomized to receive services that reduce the hassle costs of moving, the proportion of movers increased from 15 percent without the services to 53 percent with the services that reduced hassle factors (Bergman, Chan, & Kapor, 2020). In contrast, even sizable increases in the value of the voucher did not have large effects, suggesting that the key factor is the disutility from navigating bureaucracy and other transaction costs.

Behaviorally Informed Financial or Nonfinancial Incentives

Many people who read about behavioral economics in the lay press assume the term refers to financial incentives or paying individuals to change their behavior. It is certainly the case that many well-known behavioral economics studies leverage incentives to counter present bias. Because present bias leads people to focus on immediate costs and benefits of an action at the expense of future costs and benefits, a financial incentive that is delivered in the present can make benefits feel more immediate and can counter immediate costs (time, hassle, etc.). Financial incentives are also well-suited to incorporating other behavioral principles, including loss aversion, regret aversion, and reference dependence, into the design of interventions.

A common query, particularly from economists, is whether financial incentives should be considered as part of the behavioral economics toolkit. After all, what’s “behavioral” about changing the cost of taking some action and expecting a response to that change? For example, a frequently cited study showed that a $750 incentive prompted 15 percent of employees of a large private firm to quit smoking: in comparison, only five percent quit smoking in a control group that received only information about cessation and available cessation programs (Volpp et al., 2009). While this result may seem obvious—of course it makes sense that people who were paid to quit did so more frequently than those who were not—it is important to note that for most habitual smokers, the financial savings that accrue from quitting exceed $750 very soon after quitting. It violates assumptions of traditional economics that an offer of $750 would be much more effective at prompting quitting (than no financial incentive). Similarly, participants offered a $10 incentive to take up 401(k) retirement savings enrollment increased their savings more than participants who received information about the future benefits of retirement savings that were worth far more (Bhargava & Conell-Price, 2022).

The distinction that behavioral economists point to is between incentives that are specifically designed to capitalize on knowledge of behavioral characteristics and those that simply increase the benefits or change the price of a particular option. An illustration of this distinction is a study that showed that consumers react more strongly to a tax if it is made more salient to them: in this case, if the consumer saw a price tag on an alcoholic beverage that specified the tax rather than having the opportunity to notice a tax of the same amount being applied at the cash register (Chetty, Looney, & Kroft, 2009). The accumulated evidence suggests that behaviorally informed microincentives are a promising way to motivate actions in settings in which the future benefits of an action (or the current costs of inaction) may otherwise be ignored.

Lottery-Based Incentives

Many variations of financial incentives have also been tested, with features that further leverage behavioral research findings to boost effectiveness. Lottery-based incentive programs leverage the fact that participants may be more motivated by the small probability of winning a large prize than by a guaranteed incentive of a much smaller prize. This phenomenon results from base rate neglect (see below), a manifestation of reference dependence. A regret lottery further leverages regret aversion, the tendency to strongly dislike the possibility of regret, which motivates behaviors that avoid future regret. In regret lotteries, participants are eligible to “win” the lottery (i.e., have their winning number drawn) but are only eligible for

the prize if they have already taken some specified action by the time their lucky number is drawn (Zeelenberg & Pieters, 2004; Haff et al., 2015; Patel et al., 2016a). Deposit contracts leverage the notion of loss aversion or the endowment effect (see section on reference dependence, below; see also Chapter 3) by asking participants to put up their own money in a bet against a future behavioral outcome. If the behavior is achieved, the participant keeps the deposit (and possibly earns additional incentives). If the participant fails to achieve the behavioral target, the deposit is forfeited (Volpp et al., 2008; Barankay et al., 2020).

These examples of variations on financial incentives highlight the importance of behavioral design, given the many design elements required in an incentives program, including the size, conditionality, currency, probabilistic or guaranteed nature, and framing of the incentive. While financial incentives have been shown to be effective for many behaviors, particularly one-off behaviors where procrastination is a key barrier, there is less evidence that they can lead to sustained behavior change after the incentive is removed or can boost motivation for behaviors with which a person has considerable experience or strong preferences (Mantzari et al., 2015; Thirumurthy, Asch, & Volpp, 2019; Luong et al., 2021). There is also mixed evidence about the effectiveness of lotteries in the handful of studies that directly compare lottery-based incentives to a fixed or guaranteed payment (Halpern et al., 2011; Meiselman et al., 2022).

Commitment Devices

A different way to counter present bias is to commit a future self to an action that is difficult to evade or delay when the moment comes. This can be accomplished through commitment devices (such as when Odysseus tied himself to the mast) that can be harder or softer depending on the degree of enforceability of the device (Ashraf, Karlan, & Yin, 2006; Bryan, Karlan, & Nelson, 2010; Schwartz et al., 2014). Commitment devices (also called commitment contracts) can take the form of a savings vehicle that restricts when and how you can access your savings or a deposit you put down that is forfeited if you fail to exert self-control over an undesired or unhealthy behavior (e.g., smoking). An important policy question related to the effectiveness of commitment devices is whether they are only useful for “sophisticated” people (those with accurate beliefs about their degree of self-control or present bias) and much less effective for “naïve” people who hold inaccurate beliefs (see Chapter 3). Naïve people may pay for a commitment device that they are unlikely to benefit from (Bai et al., 2021).

Prompting individuals to make a plan with planning prompts or guiding them to state their implementation intentions to complete a behavior has also been shown to boost completion of a task and overcome present

bias (Milkman et al., 2013; Rogers et al., 2015; Yeomans & Reich, 2017; Mazar, Mochon, & Ariely, 2018; Silva et al., 2018; Ahn, Hu, & Vega, 2021; Robitaille, House, & Mazar, 2021). A well-known example of the effect of planning and implementation prompts can be found in the literature on motivating voting behavior (Nickerson & Rogers, 2010). In another domain, a study demonstrated that a simple planning prompt on a postcard significantly increased use of an employee vaccination clinic in comparison with a postcard that only reminded participants of the clinic dates (Milkman et al., 2011). Pairing a more desired and a less desired action together, known as temptation bundling, can also counter present bias that leads to putting off the less desired action (Milkman, Minson, & Volpp, 2014).

Reminders

Finally, simple reminders can also counter the tendency to delay costly behaviors, especially when they mention a future goal or desired state (Karlan et al., 2016). Reminders also cut through inattention, remedy prospective memory failures (i.e., forgetting), and increase salience of the desired behavior. A systematic review of reminders for health behaviors found moderate effectiveness, with the strongest evidence of effectiveness from reminders that are frequent and those that are accompanied by personal contact with a health care provider or counselor (Neff & Fry, 2009). In a powerful example of the effectiveness of reminders, researchers have been able to reduce failures to appear in court following an arrest summons with behaviorally informed reminders about court dates sent in text messages (Fishbane, Ouss, & Shah, 2020).

ADDRESSING REFERENCE DEPENDENCE AND FRAMING

To address reference dependence and framing, the behavioral challenge for policy makers is to recognize that individuals evaluate risk decisions in terms of a specific reference point and may not treat risk according to traditional economic utility theory. Thus, they are sensitive to the framing of decision problems.

Behavioral research has shown that people do not evaluate decisions in a vacuum or using absolute standards. They are highly influenced by reference points, particularly when considering tradeoff decisions (now or later, sell or buy, etc.) or evaluating risk. Reference dependence implies that policy designers can guide choices by creating or highlighting a specific reference group, drawing attention to a temporal reference point, or using framing to prompt a comparison to a reference point. In addition to the strategy of choice sets discussed above, three other strategies are especially relevant to reference dependence: framing, foot in the door, and fresh start effects.

Framing

A common framing strategy for behavioral interventions is a loss frame, which leverages loss aversion and the endowment effect by drawing attention to the negative consequences of inaction (i.e., the bad things that will happen if you do not do something), which may be more motivating than a gain frame (i.e., the good things that will happen if you do something). Loss-frame messages promoting health behaviors are often more effective for some groups than others—for example, for those more at risk of the consequences of a behavior and those in a positive affective state (Keller, Lipkus, & Rimer, 2003; Cho & Boster, 2008). There is also robust evidence that loss-frame messages are more effective for screening or disease detection behaviors, while gain-frame messages are more effective for prevention behaviors (Rothman et al., 2006) or prosocial behaviors (Castelo et al., 2015).

Foot in the Door

Self-perception theory and the literature on people’s preference for consistent actions are the basis for the use of foot-in-the-door techniques, which invite individuals to agree to a smaller request before presenting a larger one, thus resetting the reference point for the larger decision. While much of the foot-in-the-door literature concerns fundraising or civic behavior, foot-in-the-door interventions have also increased engagement with health promotion activities (Dolin & Booth-Butterfield, 1995; Ybarra et al., 2014).

Fresh Start Effects

The fresh start effect is the tendency to be more motivated to set or achieve goals related to key personal or cultural milestones, such as birthdays; holidays; or the start of a new year, month, or week (Dai, Milkman, & Riis, 2014). Simply sending a reminder or inviting goal setting at a fresh start moment can increase engagement. There is some evidence that framing a message around a fresh start—which encourages individuals to reflect on these temporal landmarks—can increase the intention to pursue a goal (Dai, Milkman, & Riis, 2015).

ADDRESSING SOCIAL PREFERENCES AND SOCIAL NORMS

To address social preferences and social norms, the behavioral opportunity for policy makers is that people often care deeply about the well-being of others and how their own behavior and social standing compares with

those of others, and the social environment has a significant influence on people’s decision making. This finding from behavioral research does not align with the most common traditional economic model, which assumes that individuals care most about their own well-being and ignore the actions of others in society when making their decisions (but some traditional models do; see Chapter 3).

Many intervention strategies specifically leverage the importance and salience of social norms (detailed below) and people’s perceptions of their social standing. Four intervention strategies are particularly relevant for taking advantage of social preferences and social norms: social proof, social comparison, social influence, and reciprocity and altruism. As with all behavioral strategies, interventions related to social norms can be nuanced and may backfire; careful attention to this possibility during both the design and evaluation of these interventions is crucial (Bicchieri, 2016; Bicchieri & Dimant, 2022; Constantino et al., 2022).

Social Proof

Social proof interventions provide descriptive data about other people’s choices and behaviors, such as the number of drinks most students consume at a party (Perkins, 2002) or how many other voters plan to vote or have already voted (Gerber & Rogers, 2009). These interventions have been used to reduce risky or harmful behaviors, such as binge drinking among students, and to increase positive behaviors, such as voter engagement. The mechanisms underlying these interventions include motivating conformity and serving as a simplifying decision-making heuristic.

Social Comparison

Social comparison interventions explicitly compare an individual’s performance with a relevant comparison or reference group. The evidence on social comparison is mixed, with several instances of successful interventions, as well null results and even backfire effects (Beshears et al., 2015). A particularly strong response may be triggered by social comparisons with others if the comparison is not anonymous and individuals are publicly compared with others.

A controversial version of a social comparison intervention provides the information for an individual and a comparison between the individual and a relevant peer group, as in a voting intervention that reminded voters about their own voting record and the voting rate of their neighbors (Gerber, Green, & Larimer, 2008). In this study, the social comparison mailer further informed voters that they are likely to receive an updated mailer later with the updated voter turnout record. This mailer led to a

dramatic 8 percentage point increase in voter turnout, probably the largest such finding in the literature. Voters anticipated that their turnout choice would be shared with neighbors and wanted to be sure to vote so they would be seen by neighbors as public good providers. Social comparison can also be implemented as feedback about peer performance, a common strategy for changing clinicians’ behaviors related to adherence to evidence-based guidelines (Hallsworth et al., 2016; Meeker et al., 2016).

Social Influence

Social influence, which is also known as social modeling, involves having celebrities or respected opinion leaders endorse or model a desired behavior, which also leverages the importance of injunctive or prescriptive social norms. This strategy has been studied extensively for recycling behaviors: in that context, social comparison is more effective than information provision alone, but it does not work better than financial incentives (Osbaldiston & Schott, 2012; Varotto & Spagnolli, 2017).

Reciprocity and Altruism

The importance of social relationships can also be invoked by interventions that leverage reciprocity and altruism. Traditional economic models assume self-interest, but studies of human behavior suggest that people feel motivated and rewarded when they reciprocate and act altruistically. While many laboratory experiments have successfully used altruism appeals to increase prosocial behavioral intentions (e.g., to get vaccinated), there is less evidence from real-world or field studies (Hershey et al., 1994; Rieger, 2020; Cucciniello et al., 2022; for evidence related to organ donation, see also Sallis, Harper, & Sanders, 2018; Robitaille et al., 2021).

CONCLUSION

The range of behaviorally based economic interventions is large and expanding. It includes strategies for which there is excellent empirical evidence of effectiveness (defaults, framing) and others for which the evidence is mixed or less robust (incentives, planning prompts). However, the interventions are not well defined: that is, many of the strategies are defined differently by researchers and practitioners in different policy domains and from different academic disciplines and fields.³ The field of behavioral eco-

___________________

³ For a detailed discussion of this problem, see National Academies of Sciences, Engineering, and Medicine (2022).

nomics has also not yet developed or embraced rigorous design methods for characterizing behavioral barriers in specific contexts and then selecting or matching interventions or strategies for addressing those barriers. It is impossible to know now whether this laudable goal can ever be achieved, given the importance of context and the unique characteristics of many behavior change targets. Published studies describing behavioral interventions often fail to describe the source of the behavioral intervention tested or the design process that produced it, if any. True collaboration among behavioral economists, other behavioral scientists, and behavioral designers (i.e., practitioners and researchers focused on developing behavioral solutions to meet specific needs) has been rare.

Conclusion 4-1: Research is needed to advance methods used to characterize barriers to a specific behavior and then to design behaviorally informed interventions that address those barriers. Such research can address the pressing policy demand for guidance about when to use different behavioral strategies and how to match a policy challenge to an intervention strategy and evidence about the optimal level at which to target behavioral interventions: individuals, practitioners and providers, firms and organizations, or government entities.

REFERENCES

This page intentionally left blank.