Oliver (2021) on Prospect Theory and Risk Preferences

Adam Oliver, “Reflecting on Reflection: Prospect Theory, Our Behaviors, and Our Environment.” Behavioural Public Policy, 1-11, 2021.

• A full-blown version of prospect theory includes both diminishing sensitivity in both the gains domain and the loss domain, as well as probability weighting, where low probability outcomes tend to be overweighted in valuing prospects and high-probability outcomes tend to be underweighted.

• Diminished sensitivity on its own implies risk averse behavior in the gains domain and risk seeking behavior in the loss domain.

• But, adding probability weighting to diminished sensitivity leads to what is called the "fourfold pattern" of risk preferences, or, the reflection effect. 

• For high probability gains, diminishing sensitivity and underweighting combine to produce risk averse behavior. But for low-probability gains, overweighting tends to more than offset diminishing sensitivity, leading to risk loving behavior (as with the appeal of lotteries). 

• For low probability losses, probability (over)weighting counters diminishing sensitivity, leading to risk averse behavior, while for high probability losses, (under)weighting combines with diminished sensitivity to lead to risk seeking behavior. 

• Oliver examines whether the fourfold pattern of risk preferences is displayed with respect to life expectancy prospects, as well as to monetary prospects. For monetary  prospects, a thirty-question interview protocol is administered to 60 university-affiliated people, with the (incentivized) questions focusing on eliciting certainty equivalents for prospective risky investment decisions.

• "the respondents generally became more averse to risk as probability increased in the domain of gains and as probability declined in the domain of losses, which... is consistent with the predictions of the prospect theory reflection effect [p. 4]."  

• So for monetary decisions, the fourfold pattern holds up pretty well in the interview results, though less well when looking at low probability gains or losses.

• A second set of 60 interviews with different (though still university-affiliated) people is used to look at preferences in the health domain. Now it is certainty equivalents in terms of lifetime duration that are elicited. Again, the results are largely consistent with the fourfold pattern.

• Risk seeking in the case of high probability losses – money or life expectancy – seems to be the most intense of the risk preferences.

• Though the results are consistent with prospect theory, Oliver is skeptical of the notion that prospect theory explains these results. Instead, he offers an evolutionary story (involving abundance v. scarcity) for why the fourfold pattern might emerge even without full-on prospect theory-style preferences.

• That is, evolution might have favored a pattern of behavior, a heuristic, that calls for risk seeking with high-probability losses. For some people in some circumstances, such a "bias" might still be sensible.

• Two other articles that point to useful heuristics that might be interpreted as irrational biases come to mind: Chen and Schonger on ambiguity aversion and the heuristic not to transact with folks who know a lot more than you do about the transaction; and Smitizsky, Liu, and Gneezy on endowment effects and the heuristic that as a buyer you try to understate your willingness-to-pay and as a seller you tend to exaggerate the value of the good to you.

Bastian and Loughnan (2017) on Eating Animals and Morally Troublesome Behavior

Brock Bastian and Steve Loughnan, “Resolving the Meat-Paradox: A Motivational Account of Morally Troublesome Behavior and its Maintenance.” Personality and Social Psychology Review 21(3): 278–299, 2017.

• The “meat paradox”: meat eating is widespread (about 97% of Americans), but so is revulsion at harming animals.

• How can morally troubling behavior be ignored, normalized, and longstanding?

• Cognitive dissonance is a term that refers to the discomfort people feel by holding beliefs that are inconsistent with their actions. Eating meat is inconsistent with an interest in not harming or killing innocent sensitive beings; so, meat eating is a source of cognitive dissonance.

• Cognitive dissonance leads to negative affect, and this condition motivates the search for ways of reducing or eliminating the dissonance.

• “When people take responsibility, acknowledge harm, and accept the identity relevant consequences of their actions, they will experience dissonance and cease their immoral behavior [p. 280].” So, people try not to take responsibility, not acknowledge harm, and not accept the identity implications of meat eating.

• How to reduce dissonance while still eating meat? We can deny that animals suffer; we can do this selectively: we can categorize some animals as “food animals” and act as if that label means that they don’t count or don’t suffer – even as we would never hurt a puppy. The food animal category makes salient the food dimensions (such as taste), and the other dimensions (sentience) become less salient.

• Another approach to fight dissonance is to act as if we are not responsible for our choice to eat meat, that we effectively have no choice, as with the "3 N’s": eating meet is natural, normal, and necessary. People who abstain from meat then appear deviant, their opinions are suspect.

• Further, we can downplay the amount of meat we eat or suggest that we only consume ethical meat. We emphasize that we are relatively responsible!

• Our approaches to dissonance reduction can increase the commitment to meat eating. Indeed, doubling down on meat eating and believing that it is the proper course to take (and that the alternative is weird or unwise) reduces dissonance and negative affect. 

• The ritual and symbolism then, in meat eating (turkey for Thanksgiving, ham for Christmas!), is not surprising. Not eating meat looks almost infeasible and anti-social.

• The habitual, automatic nature of meat eating renders it no longer a decision, reducing dissonance – and this is in part why meat eating becomes habitual.

• Habits can spread throughout a population – seeing someone eating meat suggests that the eater is not experiencing dissonance, that such behavior is consistent with morality.

• The whole culture, then, can shroud the immorality of meat eating (and reduce or eliminate dissonance): look at how nations build identities around morally questionable acts of war. Meat production is hidden away, meat purchasers and eaters don't have to face harms to animals directly.

• But social contagion can work both ways. The positive feedback (sorry) loop could reverse, and meat eating could quickly fall out of favor.

• The theoretical points built up around the discussion of meat eating are then applied by the authors to the treatment of refugees and to prejudice. The authors see their contribution, I believe, mainly in indicating how methods employed to reduce dissonance can strengthen the commitment to the morally-troublesome behavior. 

Exley and Kessler (2021) on Information Avoidance

Christine L. Exley and Judd B. Kessler, “Information Avoidance and Image Concerns.” NBER Working Paper 28376, January 2021 (revised May 2021).

• Contrary to what a standard economic view would suggest, people often avoid receiving information that is pertinent and seemingly free. A plethora of potential reasons can be offered as explanations for information avoidance.

• People want to think well of themselves, so they might like to muddy the waters around selfish acts, by remaining ignorant (thereby possessing plausible deniability) of the actual extent of their selfishness. This article asks how much information avoidance is driven by concerns with maintaining one’s self-image.

• More specifically, people often avoid learning about how their choices affect the outcomes for other people – even if the information would not affect their decision. When I intend to act selfishly, I might not want to know how much others are harmed by my choice. The opportunity to avoid information about harms to others promotes more selfish behavior. 

• The approach here is to look at information avoidance with respect to two decisions that differ only in that one decision has no prospect of involving a selfish motive (as the chooser's own payoff is not implicated in the decision).  

• A person is asked to choose either option A or option B. The "Self/Other condition" is one in which the chooser will receive a higher payoff from A than for B. Their choice between A or B, however, will also affect the payoff of someone else ("Other"), and it is possible that the other person will receive a much higher payoff with B. At the start, the chooser does no know what option is best for the other person, but does know that A is best for the chooser him or her self. The chooser (sometimes) has the option of just making the choice between A or B, or, learning the Other's payoffs associated with A or B. Will the chooser seek to learn this information?

• If Self and Other turn out to both do better with A, that is termed the "aligned" state; if it turns out that Other prefers option B, that is the "unaligned" state, the preferences of the two folks do not match. (So choosing to get the info means that the chooser learns whether the preferences are aligned or unaligned – if they are aligned, then there is no conflict, the selfish choice is also good for the Other.)

• In the Other/Other condition, alternatively, the chooser is again making a choice in the Self/Other circumstances, except what used to be the chooser's payoffs now go to someone else, a third party – the chooser has no skin in the game, no selfish interest, but the interests of the two Others might conflict. Will the chooser elect to learn whether the interest are aligned or unaligned?

• Four studies, overall n > 4,600.

• In Self/Other, if choosers know (no choice) the info and it is unaligned, about 1/3 make the selfish choice. If the state information is not revealed unless requested, 2/3 of subjects in Self/Other choose not to acquire the information. If the choosers elect not to know, more than half choose the selfish option. That is, people who avoid the info behave in a more selfish way, on average, than if their choice fully revealed their selfishness because of their knowledge of unalignment.

• In the Other/Other condition, where there is no selfish option, information avoidance falls – but only by about 20%; that is, more than half the participants still choose to avoid the information. 

• The Other/Other results suggest that most of the information avoidance is not about self-image concerns, because there are no such concerns in the Other/Other condition.

• If you make info provision an active choice (as opposed to opt-in), information avoidance falls by about half: it sort of seems like participants just aren’t paying much attention without the need to make an active choice?

Blanchflower and Bryson (2021) on Covid and Mental Health

David G. Blanchflower and Alex Bryson, “Biden, Covid and Mental Health in America.” NBER Working Paper 29040, July 2021.

• The onset of Covid brought on an economic cataclysm in the US [and elsewhere, of course]. Weekly claims for unemployment assistance in the US prior to mid-March 2020 were on the order of 250,000. One week later, they were 2.9 million, and the subsequent week, 6 million.

• Unemployment in February 2020 was 3.5%, and in April 2020, 14.7% (or five percentage points higher when absent (effectively laid-off) workers are included). By November 2021, the US returned to mid-2020 levels of unemployment claims, and the October 2021 unemployment rate was 4.6%.

• What did the pandemic, and its economic effects, do to mental health in the US? To answer this question, the authors employ Household Pulse Survey Data (an ongoing survey) from April 2020 to June 2021, covering some 2.3 million Americans.

• Happiness in the US had been declining for decades prior to Covid (using the very happy, pretty happy, not too happy scale); extreme distress (and deaths of despair) had been rising for decades, especially for people under 30.  

• Anxiety, depression, and worry all peaked in November 2020 (the presidential election month); but by April 2021, mental health was back at pre-Covid levels. The Covid-onset surge in mental health problems was largest for the young. 

• The mental health impacts of Covid on young people, combined with schooling disruptions and a bad macroeconomy for labor market entrants, could hold negative long-term economic effects for the "young during Covid" cohort.

Wirtz, Tucker, Briggs, and Schoemann (2021) on Happiness and Social Media

Derrick Wirtz, Amanda Tucker, Chloe Briggs, and Alexander M. Schoemann, “How and Why Social Media Affect Subjective Well-Being: Multi Site Use and Social Comparison as Predictors of Change Across Time.” Journal of Happiness Studies 22:1673–1691, 2021.

• "Social" media hold the potential to increase social interaction, which tends to raise subjective well-being. But the environment in which interactions take place on social media is quite different from face-to-face encounters.

• Facebook, Twitter, and Instagram use are tracked over ten days by an experience sampling (five times per day, contacted via e-mail) method; n=77 semi-captive (college-course-taking) subjects, with 6 to 48 responses per subject. 

• The main measures: positive affect, negative affect, and life satisfaction (“I am satisfied with my life,” strongly agree through strongly disagree); loneliness is also assessed.

• More use of social media increases negative affect (that is, there's more bad feelings), but does not influence positive affect. The suggestion, then, is that social media use decreases subjective wellbeing, not by undermining positive feelings, but by increasing negative feelings. Instagram use hints (but not in a statistically significant way) at some increase in positive affect, too, and “overall” affect.

• Social comparison on social media (seeing all your friends' best moments on Facebook) also correlates with higher negative affect, while also reducing positive affect. It might be social comparison, and not social media use per se, that drives diminished subjective well-being. For instance, Facebook use is no longer significant (in altering happiness) when controlling for social comparison.

• Social media use increases loneliness, and loneliness seems to drive some social media use.

• Direct (face-to-face) interactions remain good for affect, even within this sample.

• Life satisfaction (as opposed to affect) is measured at the beginning and the end of the experiment (that is, there is no experience sampling); there are no significant effects of social media use on life satisfaction changes over the two-ish weeks.

• Facebook use (working through social comparisons) seems to lower self esteem; “passive” use of Facebook (scrolling, not posting or messaging) seems particularly problematic.

Semple (2021) on Life‑Evaluation and Public Policy

Noel Semple, “Good Enough for Government Work? Life‑Evaluation and Public Policy.” Journal of Happiness Studies 22: 1119–1140, 2021.

• We collect happiness data via "life-evaluation" measures such as the Cantril ladder. But a leading strain of public policy thinking is "welfare-consequentialist," where individual utility or welfare is at the center of judging the desirability of different social states. 

• Happiness is not the same as welfare. But can we use life-satisfaction data anyway in evaluating welfare?

• The approach examined by Semple – life-evaluationist welfare-consequentialism (LEWC) – argues that the government should seek to maximize aggregate subjective well-being – and life satisfaction is sufficiently measurable to be implementable. In some circumstances. happiness data are informative of the success of public policy (even when there is no consensus on what actually constitutes welfare). That is, maximizing welfare can often be implemented by maximizing happiness, as measured by subjective wellbeing.

• Practicality is an important argument for the LEWC approach: we have broad, long-term data on subjective wellbeing. Further, where happiness is an imperfect proxy for welfare, it still tends to indicate the sorts of things that make people better off. 

• Happiness research has revealed that mental health problems detract markedly from subjective wellbeing; further, people do not adapt over time to some negative mental health conditions. [Look at the advantage of happiness over willingness-to-pay: "The average person who has not experienced clinical depression simply cannot say, in a reliable way, how much they would be willing to pay to avoid it. By contrast, notwithstanding the problems identified above, those who are clinically depressed, and those who are not, are able to give reasonably consistent evaluations of their own lives [p. 1133]."]

• Targeting mental health with policies that make treatment more available also helps to target those who are worse off, those whose subjective wellbeing is at the bottom of the distribution. And better mental health for parents holds positive second-order impacts for the happiness of their children.

• Approaches to human welfare are all over the map, including preference satisfaction and hedonism and...  But LEWC is grounded in data, it looks at the evidence on what produces happiness. We do not need to agree on what constitutes human welfare to implement LEWC. 

• The LEWC approach, then, offers a form of robustness; it can advance welfare without presuming what constitutes welfare.

• Much public policy is too small or abstract for its effects to be measured by happiness; further, people do adapt over time to many changes in their life circumstances. So, it is sensible to sometimes supplement SWB with preference fulfillment in serving as a proxy for welfare. Preferences for longevity can be used, for instance, to overcome some of the "moment in time" issues connected to happiness measures.

Easterlin and O'Connor (2022) on that Paradox

Richard A. Easterlin and Kelsey J. O'Connor, “The Easterlin Paradox.” In: Zimmermann, K.F. (ed.) Handbook of Labor, Human Resources and Population Economics. Springer, Cham. [Working paper version here.]

• From the Abstract: “The Easterlin Paradox states that at a point in time happiness varies directly with income, both among and within nations, but over time the long-term growth rates of happiness and income are not significantly related.”

• That is, the Easterlin Paradox concerns a disconnect between cross-section and time series data linking happiness and (real) income.

• The life satisfaction (happiness) data used in Easterlin and O'Connor come from the World Values Survey (WVS) and the Gallup World Poll (GWP). This analysis takes place at the national level (so income is real income per capita within a country).

• The WVS question: “All things considered, how satisfied are you with your life as a whole these days?” 1 (dissatisfied) to 10 (satisfied); 67 countries, on average over 27 years, for this study, with about 5 observations per country.

• The relevant GWP item: Cantril ladder, 0 to 10, worst possible life (0) to best possible life (10); 123 countries, 12 to 15 years of annual data.

• A major point of emphasis for Easterlin and O'Connor is that the key for testing is to ensure that you have long time-series data, so that you are not just capturing a boom or a bust (during which the usual cross-section positive connection between happiness and income will prevail).

• Easterlin and O’Connor find that in their data, the Paradox essentially holds when the East European transition countries are not included. (The conditions of transition and the time-span of the data from transition countries are such that even more than ten years of data might essentially be capturing a long boom.) 

• With the transition countries included, or if shorter (10 years or less) time spans are employed, happiness and income do show a positive relationship. Even in these circumstances, however, the effect of changes in income on changes in happiness is small.

• The authors argue that an income reference point (typically, social comparison) is what drives the Paradox. As everyone's income rises, your reference point (other folks' income) shifts up, so your higher income does not bring more happiness. In a recession, alternatively, the relevant reference point is your own previous income, so people on average become less happy. (That is, the Paradox might hold even in the short-run during an expansion, but a recession will cause both happiness and income to fall.) 

• China, Japan, India – all have experienced tremendous income boosts without increases in happiness.

• The authors argue that the “threshold” claim (that increased income does increase happiness until you are fairly well-off, at which point more income doesn't lead to more happiness) also is wrong: the Paradox applies to rich and poor individuals, not just the rich.

• Public policies can still raise happiness! Promotion of employment and a social safety net are good for happiness.

Han and Wibral (2020), “Organ Donation and Reciprocity"

Hua-Jing Han and Matthias Wibral, “Organ Donation and Reciprocity.” Journal of Economic Psychology 81, December 2020.

• Do people who display more positive reciprocity donate organs more readily?

• Do people who display more negative reciprocity donate organs less readily?

• Should people who had previously registered as organ donors receive priority should they need an organ? (Such a policy was adopted in Israel, and apparently increased donor registrations.)

• Does a bad experience at the Department of Motor Vehicles (DMV) undermine organ donor registration (which, in the US, often takes place at the DMV)?

• The authors make use of some helpful German data, drawn from a survey that asked questions related to both organ donation and to reciprocity. Three questions are used to measure negative reciprocity (e.g., "If somebody offends me, I will offend him/her back."), and three questions measure positive reciprocity (e.g., "If someone does me a favor, I am prepared to return it").

• Two dependent variables: willingness to donate (46% yes) and possession of an organ donor card (12% yes); N=683.

• Both positive and negative reciprocity affect willingness to donate (in the expected directions); however, only negative reciprocity influences donor card acquisition.

Besley (2021), “Reciprocity and the State”

Timothy Besley, “Reciprocity and the State.” LSE Public Policy Review 2(1): 1-10, 2021. 

• "Reciprocal obligation lies at the heart of state-citizen relations [p. 1]." We can see this as the state tries to grapple with handling COVID.

• How can a society build reciprocity, which is (often) a sort of informal norm or institution?

• Government sovereignty makes it hard for for the state to commit to its future conduct – it could always reverse course, even if the citizenry is relying on earlier promises (such as pension payments). 

• Incidentally, the social contract theorists (Hobbes, Locke, Rousseau) all invoke reciprocity between the governed and the governors.

• Two elements that can build confidence in governments are (1) constraints on executive power and (2) open competition for leaders (elections).

• "Reciprocity kicks in when states deliver collective goods for citizens and, in exchange, when citizens offer their support, whether by paying taxes, volunteering for military duty, or obeying the law. However, their willingness to do so in the long-term is contingent on state behaviour [p. 5]."

• Taxes and social insurance are two areas where we can apply more concretely general considerations about reciprocity.

• The extent to which citizens believe that it is OK to cheat on their taxes varies quite a bit from country to country. People who trust their government, and believe that taxes are used for the social good, have higher tax morale. 

• To bolster reciprocity, we might want to discourage tax avoidance (which is legal) along with tax evasion (which is illegal) – both can undermine tax morale.

• Social insurance, by sharing risks, brings broad public benefits. Universality helps to promote "quasi-voluntary compliance" that more targeted programs might put at risk.

Newall, Walasek, Hassanniakalager, et al. (2023) on Gambling Risk Warnings

Philip W.S. Newall, Lukasz Walasek, Arman Hassanniakalager, et al., “Statistical Risk Warnings in Gambling.” Behavioural Public Policy 7(2), 219-239, 2023 [pdf available here].

• Warning messages in the gambling arena, when they are provided, tend to be generic and do not facilitate comparison across various betting options.

• Theoretical loss = (house edge) times (total amount bet), or, the expected price of a bet or bets; for instance, in European roulette, the house edge is 2.7%, so the theoretical loss of betting $100 is $2.70.

• Theoretical loss is a long-run concept and not salient for bettors, while their recent gambling experience is quite salient. But given that most of the social harm from gambling is tied to bettor losses, theoretical loss is a decent metric for the social risks of alternative wagers. 

• Expected prices (or theoretical loss) of gambling tends to be hidden from gamblers – but it would be possible to provide (sometimes approximate) theoretical loss information, or other related metrics such as house edge. (For games where player behavior affects the odds, the risk could even be personalized based on player characteristics.) Presumably gamblers (all else equal) would be dissuaded from forms of gambling with high theoretical losses. 

• Should gambling warnings also note volatility and the slow convergence to theoretical loss levels?

• Gamblers cannot identify vast differences in expected price on seemingly identical games.

• Some jurisdictions mandate “return to player”-style info; but players often do not understand it. And sometimes the info is hard to find and to see.

• Should the automatic reinvestment of small wins into a player's available gambling account be banned?

• In sports betting, longshots and “exact score” or other prop bets tend to have higher house edges – and these are heavily advertised.

Michaelsen, Johansson, and Hedesström (2021) on Experiencing Default Nudges

Patrik Michaelsen, Lars-Olof Johansson, and Martin Hedesström, “Experiencing Default Nudges: Autonomy, Manipulation, and Choice-Satisfaction as Judged by People Themselves.” Behavioural Public Policy, First View, 1–22, 19 March 2021. 

• This article examines how people feel about default nudges, especially with respect to their perception of their autonomy and their satisfaction with their ultimate choice. "We present three experiments (total N = 2083) where participants are subjected to opt-out default nudges and compared with participants subjected to opt-in or a no-default active-choice format [pages 3 and 4]."

• Study 1: MTurk folks consider whether to have “green” (and slightly more expensive) appliances in their (imaginary) new flats. There are three conditions: green (environmentally friendly) default (opt out); non-green default (opt-in); and active choice.

• The effect of the default is big, with about twice as many green appliances chosen with opt-out relative to opt-in; active choice leads to more green appliances than opt-in, too, but not as much more as opt-out.

• Participant ratings of their autonomy, satisfaction, and perceived threat to their choice freedom do not differ based on the condition. In particular, the opt-out default, which produces greener choices, does not lead to participant dissatisfaction.  

• Study 2 replicates study 1, but with a twist: now, half of the participants ae told about the default or active choice, and how default nudges are expected to influence choices. How does making the nudge transparent affect its influence on choices and satisfaction?

• It turns out that lots of participants don’t really understand the explanation of how the default nudge is supposed to work.

• The large impact on choosing green appliances from the opt-out default is replicated – and nudge transparency does not change that.  

• The participants give high ratings across-the-board for autonomy, but opt-out folks report slightly higher experienced autonomy and satisfaction than do their opt-in brethren.

• Study 3 aims to raise the stakes (slightly). Participants receive a 20 cent bonus, on top of the 50 cents they are paid to take part.  But they are given an opportunity to donate their bonus to a charity. The opt-out condition is where the bonus is defaulted to the charitable contribution. 

• Defaults continue to have consequences: opt-out folks are more likely to donate than are opt-in; active choice is in the middle; the three conditions have no effect on feelings of autonomy.

• With disclosure (and understanding), opt-in and active choice donations go up. (Again, the number of participants who do not understand the disclosure is fairly staggering.)

• Opt-out folks view the set-up as more threatening to freedom of choice, and full transparency raises everyone's view of such a threat – but despite that threat, people aren't really worked up over it. 

Colby, Li, and Chapman (2020) on Avoiding Healthy Defaults

Helen Colby, Meng Li, and Gretchen Chapman, “Dodging Dietary Defaults: Choosing Away from Healthy Nudges.” Organizational Behavior and Human Decision Processes 161: 50-60, 2020.

• Imagine a burger restaurant that makes carrots the default “side” – as opposed to French fries, say. Are people more likely to choose the carrots than if fries were the default? [Yes]. 

• Note that, unlike some environments with default nudges such as organ donations or pension registration, dining is an ongoing choice; so…

• …would people be less likely to return to a restaurant in which carrots are the default side dish?

• Two real-life(ish) and three hypothetical studies are conducted. In Study 1, university mugs are packaged with either M&Ms or a (healthier) fruit-and-nut bar as defaults. Shoppers know that they can easily override the default by asking the cashier for the alternative filling. 

• The defaults turn out to be almost perfectly sticky (unlike those well-designed M&Ms). But, more mugs are sold when M&Ms are the default. People don't override the healthy default, even though they know they could, but the healthy default seems to put them off the mug purchase altogether.

• If the store elected to employ a healthy default to promote customer health, is it doing the right thing? 

• Study 2 is also a real-life one, involving signing up for meal kit deliveries. What (three) meals will serve as the default for the kit? The experiment compares a default kit of three healthy meals (from one company) with an alternative default of three unhealthy meals (from a different company); once again, it is easy for meal kit purchasers to override the defaults.

• Again, the defaults are sticky, and in particular, the healthy default works in terms of nudging the choice of healthy meals. But, when choosing which company to patronize, there is a significant “dodge” effect: people became more likely to choose the company with the unhealthy default meals. So maybe it is not such a good idea for meal kit companies to set up healthy meals as the default?

• Study 3 is an internet-mediated "lab" experiment about which (hypothetical) restaurants people are likely to patronize. A subject is asked to order food from two competing burger restaurants. One burger restaurant uses turkey burgers as the default (this is the relatively healthy version, though not for the turkey), and the other burger place has been hamburgers as the default. (A similar comparison is run between two pizza places, and two sandwich shops). 

• Once again, (hypothetical) patrons tend to stick with the default, whether it is heathy or unhealthy. But they also are asked which of the restaurants they would return to, and in the case of burger joints, for the most part they would return to the unhealthy default version. People stay away from burger joints with a healthy default – turkey instead of beef burger – even though when at the "healthy" place, they generally go with the default. (It turns out that people are not so reluctant to return to pizza or sandwich shops that offer healthy defaults.)

• Study 4 looks at what happens if potential customers see an ad that lets them know in advance that turkey is the default burger meat. People become more enamored of the alternative restaurant that offers a beef default, and also suspect that ordering will be more of a hassle with the healthy default.

• Study 5 looks at healthy versus unhealthy defaults in a familiar setting, a McDonalds. There is a line, and a subject observes that the five people ahead of him or her in the line each stick with the default burger, where it is healthy or unhealthy. Does this social confirmation influence subsequent orders? 

• Once again, people tend to stick with default – but less so when the default is a turkey burger, when a third of folks request an override. Nonetheless, people indicate that they would dodge the healthy default restaurant in the future – but the dodgers were primarily folks who did not bother to override the healthy default!

• So, defaults can be quite sticky. Set up a health default, and you can expect that it will increase the percentage of customers who choose the healthy option. But there is a partial offset: some customers will dodge the restaurant with the healthy default, reducing the effectiveness of the healthy nudge by some 25%.

de Ridder, Feitsma, van den Hoven, et al. (2020) on Simple, Not Easy, Nudges

Denise de Ridder, Joram Feitsma, Mariëtte van den Hoven, et al., “Simple Nudges That Are Not So Easy.” Behavioural Public Policy, 1-19, 2020.

• This article draws upon a Dutch research program (2014-2018) with the acronym WINK: “Welfare Improvement through Nudging Knowledge [p. 3].” WINK "examines nudging from the perspective of three core disciplines: ethics, public administration and psychology [p. 3]."

• The authors posit that there are three common yet false assumptions made concerning nudges: (1) nudges threaten autonomy; (2) nudges are easy to implement; and (3) nudges are simple and effective.

• Autonomy might mean unfettered choice, but it also could refer to agency (capacity to choose) and self-constitution (identity and living the life one desires). Perhaps people who criticize nudges for threatening autonomy (even though by definition nudges do not preclude any choice options) do so because they have adopted a definition of autonomy in which maintenance of freedom of choice does not end the debate.

• Is it enervating to possess too many choices? Should we welcome nudges to reduce the burden of choice?

• The explanation of nudges can alter judgments about the extent to which they threaten autonomy.

• In general, the authors are fairly hostile to claims that nudges threaten autonomy, and tend to believe that nudges are more likely to enhance autonomy, by overcoming common gaps between  intentions and actions.

• The rush to set up nudge units is one piece of evidence that shows that many policymakers view nudges as both effective and cheap to implement. "In reality, however, the process of designing, testing and implementing nudge interventions is far more complicated, which questions the supposed ‘efficiency’ of nudges [p. 8]."

• Persuasion via information provision is hard… so it is not surprising that many policymakers view nudging as a promising, quick-acting alternative (p. 10). But in general, the effects of nudges tend to be fairly modest. This limited effectiveness might be considered an advantage, as it defuses the argument that nudges are overly manipulative. 

• When folks are unsure of their preferences, nudging works pretty well, and promotes autonomy in both the self-constitution sense and in the agency sense.

de Ridder, Kroese, and van Gestel (2022) on Nudgeability

Denise de Ridder, Floor Kroese, and Laurens van Gestel, “Nudgeability: Mapping Conditions of Susceptibility to Nudge Influence.” Perspectives on Psychological Science 17(2): 346–359, 2022.

• What conditions might influence the extent to which people are susceptible to a nudge? The authors look (primarily) at three issues: (1) awareness of the nudge and/or its purpose; (2) pre-existing preferences; and, (3) are the subjects employing System 1 or System 2 thinking?

• In this article, transparency involves making subjects aware of the presence and purpose of a nudge; this is a stricter condition than Thaler and Sunstein’s publicity principle, that asks if nudges, were they publicly known, would garner public support.

• Default nudges tend to maintain their effectiveness under transparent conditions; further, when nudged opaquely, people informed after the fact don’t seem to view those nudges as an affront to their autonomy. (Though when an ex ante inquiry is made, people often indicate that they will feel their autonomy threatened by an opaque nudge; that is, they forecast a higher level of concern than they will actually experience.) 

• Nudges will be “ineffective” at altering choices if the nudge counsels behavior to which they hold a negative view, or if the nudge counsels behavior to which they already are strongly attached! That is, we needn't be too concerned with nudging people into behaviors that their pre-existing preferences do not endorse.

• If people lack clear preferences, a nudge might “impose” preferences – though starting with a request that people discuss their preferences (a plus!) might undermine any perceived manipulation.

• There is not much evidence suggesting that people are nudged more when they are employing System 1 thinking (through ego depletion, say). That is, the effectiveness of a nudge is not tied to the extent that it prompts unthinking responses.

• People with lower cognitive capacity do not appear to be more susceptible to nudges.

• The final sentence of the text succinctly summarizes the article: "Nudges do allow people to act in line with their overall preferences, nudges allow for making autonomous decisions insofar as nudge effects do not depend on being in a System 1 mode of thinking, and making the nudge transparent does not compromise nudge effects [p. 355]."

Krekel et al. (2021) Have a Course that Raises Wellbeing

Christian Krekel, Jan-Emmanuel De Neve, Daisy Fancourt, and Richard Layard, “A Local Community Course that Raises Wellbeing and Pro-Sociality: Evidence from a Randomised Controlled Trial.” Journal of Economic Behavior and Organization 188: 322–336, 2021.

• Many correlates of wellbeing have been identified, but can we teach people, in an economical manner, how to be happier? Adaptation might preclude any sort of long lasting benefit, even if we could figure out ow to teach people to be happier.

• On the other hand, not all domains of wellbeing are subject to rapid adaptation. Social connections and pro-social behaviors are among those elements of wellbeing that are adaptation-resistant.

• Introducing... “Exploring What Matters”: An eight-week course aimed at the adult population in general; the course is coordinated by lay people (great expertise not required) and “manualised," that is, it is easy to become an effective coordinator.  

• Hundreds of Exploring What Matters courses have been conducted (by August 2020), with a total of more than 5000 participants. The courses have been in 25 countries, though most of them have been in the UK. "'Exploring What Matters' is run by Action for Happiness, a registered charity in England, which was launched in 2011. Its patron is the Dalai Lama, who helped to launch the course in London in 2015 [p. 323]."

• The course material includes mindfulness, gratitude, building good habits, social connections…

• Krekel et al. undertake a randomised control trial (RCT) for six Exploring What Matters courses conducted in London, 2016-2017, n=146. The people who were randomly chosen to be controls were not abandoned, but rather, were placed in a later course. The initial data were collected before people knew whether they were in the treatment group or the control group.

• When participants sign up for the free course, they are asked (but not required) to make a donation to cover the course costs of about 90 pounds. The course meets once a week for eight weeks, with each individual session lasting 2 to 2.5 hours.

• "The course aims at building (i) autonomy .. using a weekly mindfulness exercise, gratitude exercise, and personal reflection, supported by... scientific evidence on that week’s theme; (ii) relatedness by facilitating interpersonal connections and social trust...; and (iii) competence by enabling participants to experience for themselves how behavioural changes to daily routines can make differences to their and other people’s wellbeing, using goal-setting and social commitment tools to help translate motivation into action."

• The researchers look at 15 outcome measures (four on subjective wellbeing (SWB), two on mental health, three on pro-sociality, and six biomarkers collected via saliva samples, such as cortisol levels).

• The course brought about a large increase in SWB on average, like a point higher on a ten-point scale. Mental health improved, but not to the same extent that cognitive behavioral therapy brings. Compassion and social trust rose, too. [These findings have been replicated in a broader sample.] There were no significant change in biomarkers.

• These improvements are highly cost-effective, and scalable!

• Since this course evaluation was undertaken, the course has been revamped (now 6 weeks, recommended donation reduced to 60 pounds) and renamed: Happiness Habits.

• Action for Happiness is great, you should check out their website. Here's their monthly action calendar. 

Banerjee and John (2021) on Nudge Plus

Sanchayan Banerjee and Peter John, “Nudge Plus: Incorporating Reflection into Behavioral Public Policy.” Behavioural Public Policy, 1-16, 2021, available at https://doi.org/10.1017/bpp.2021.6.

• Should we embed a prod for people “to reflect” into System 1 nudges? Such a prod is the “plus.” [The authors themselves seem to equate nudges (or perhaps "classic nudges" with System 1 nudges; the "plus" activates a System 2 element in conjunction with a System 1 nudge.]

• Bundling a nudge with a plus makes nudges more salient.

• People need actual reflection, not just “the potential for reflection [p. 2].”

• Thaler and Sunstein consider that a legitimate nudge must be one that would receive wide assent if everyone were informed about its aim and operation. This restriction would rule out manipulating people in ways that harm them. Nonetheless, such a legitimate nudge might, when it is affecting choice, remain hidden, invoking an automatic, System 1 response.

• Nudge plus wants to extend the transparency, so that just about everyone understands what is going on in the choice moment. Actors maintain their autonomy, though without being required to pay extremely close attention or to engage in strenuous cognitive processing. 

• Commitment devices are a nudge plus, because the need to consider whether to sign up for the device spurs the requisite reflection; likewise with cooling-off periods.

• People often choose not to use GPS on familiar routes – they are thinking, they are not being “tricked into using the device heuristically once again [p. 7].” For them, GPS is a nudge plus.

• Adding a regular prompt to set a future fitness goal turns an otherwise ordinary fitness tracker into a nudge plus.

• But… do people really desire this “experiential learning environment [p. 10],” this near necessity to reflect regularly, at least to some extent?

Duckworth and Gross (2020), "Behavior Change"

Angela L. Duckworth and James J. Gross, “Behavior Change.” Organizational Behavior and Human Decision Processes 161: 39-49, 2020.

• Suppression of impulses is a mug’s game. But changing the environment and adopting strategies concerning how we react hold promise.

• The requisite changes might be self-initiated, or initiated by others, such as employers or the government. Notice that self-deployed strategies require that people be sophisticated about their self-control shortcomings.

• Please put aside problems that arise from lack of knowledge or skill. This article is devoted to looking at situations where "we know what we should do and how to do it [p. 40]." 

• How to overcome the fact that  (temporarily) tempting options might get chosen over alternatives that are preferred by our more considered selves? The authors outline the "process model of behavior change" to point to methods where the "larger later" option is not sacrificed to the temporarily appealing "smaller sooner" choice. Here's their Figure 2 (p. 40) that outlines the process model: 

• 

• Habits or self-adopted strategies might allow you to skip the appraisal stage, leading to an automatic response when a familiar situation enters awareness. Giving into temptation, then, is not the done thing, it doesn't present itself as a viable alternative. 

• But sometimes you both want to take the stairs (for exercise) and want to take the elevator (because you are tired). What can be done to encourage stair-taking or other types of physical exercise?

• The process model suggests that you could try to alter the situation, alter where you put your attention, and/or alter your response strategy – or try to inculcate one of those automatic responses that skips the appraisal stage.

• You could begin the day by putting on comfy shoes; you could do things that would make the lazy alternative less available, like not bringing your car keys or bus pass. 

• You could keep track of your steps and adopt a so-many steps per day goal.

• You could reframe the time-consuming walk as “an energy break.”

• You could adopt (and possibly publicize) a rule that you always walk.

• Others can help. For instance, employers can make the stairwells more inviting, or promote the social norm/company culture of  stair use. 

Zlatev and Rogers (2020) on Returnable Reciprocity

Julian J. Zlatev and Todd Rogers, "Returnable Reciprocity: When Optional Gifts Increase Compliance.” Organizational Behavior and Human Decision Processes 161(S): 74-84, November 2020.

• Sometimes a charity, say, will send potential contributors a small gift along with a request for a donation. They presumably are trying to take advantage of a human propensity to reciprocate gifts, which might in part derive from an aversion of feeling indebted. 

• "Returnable reciprocity" is when the gift that is first sent is explicitly made returnable. The return of the gift by non-contributors is not required, it is perfectly OK to keep the gift and not contribute. Is it possible that by including the option to return the gift, people are more likely to contribute or to otherwise reciprocate?

• Note that with a return option, one has a new channel to avoid the guilt of not reciprocating the gift, beyond contributing – namely, returning the gift.

• The authors conduct six experiments looking at returnable reciprocity in the context of trying to incentivize people to participate in rather virtuous behavior, behavior they would think it admirable to do – like signing up for their employer's wellness program. 

• The first experiment involves a hypothetical choice, of whether you would sign up for a wellness program, if you were first gifted $5. In one condition, you keep the money and either join or not, but in the second condition, you can keep the money and join or not, or return the money (and presumably not join). The returnable condition yields a higher percentage of folks saying they will join. 

• The second experiment, this one a field experiment,  concerns returning a filled-out survey, and being gifted $5 in advance. Once again, the returnable variant increases compliance (returning the completed survey). "Study 2 found that the addition of a single line mentioning the possibility of returning the gift led to a 26% increase in compliance over a traditional reciprocity request [p. 80]."

• Other experiments find that people in the returnable reciprocity condition anticipate feeling more guilty for a failure to comply, and that when given the choice between a standard reciprocity condition and a reciprocal reciprocity condition, people prefer standard reciprocity. Nudging via reciprocal reciprocity might increase compliance, but it might not be a nudge "for good." 

Garcia-Rada, Mann, Hornuf, et al. (2018) on Adaptive Liars

Ximena Garcia-Rada, Heather Mann, Lars Hornuf, Matthias Sohn, Juan Tafurt, Edwin Severin Iversen, and Dan Ariely, “The Adaptive Liar: An Interactionist Approach of Multiple Dishonesty Domains.” CESifo Working Paper No. 7215, 2018.  

• Is “honesty” a stable trait of individuals that is displayed evenhandedly at work, in relationships, and in other life domains?

• The authors posit that honesty is domain-specific, that a person who tends to behave honestly in a work setting might nevertheless behave dishonestly in relationships. Dishonesty is about bad barrels, not bad apples.

• Previous research has shown that individual characteristics influence honesty, as do seemingly minor environmental factors: better lighting, for instance, might reduce cheating.

• One theory: there is a trade-off between personal gains from dishonesty and the maintenance of a positive self-image. (See also here, and then here.)

• Positive interpersonal feedback could arise from dishonest behavior; for instance, honesty is some settings might conflict with a social norm.

• Garcia-Rada et al. propose (as mentioned above) that life domains affect dishonesty – but the influence of domains is nation-specific.

• A survey was given to ten participant groups in each of five countries: China, Colombia, Germany, Portugal, and the US. The overall sample size for the survey was a bit over 1000. Questions concerned honesty in eight "domains": work, government, business, relationships, friends, religion, strangers, and academics.

• The results are that indeed, honesty varies among countries and domains. "Results revealed an interaction effect between country and domain as well as intraindividual variation by domains [p. 21]."

Steffel, Williams, and Tannenbaum (2019) on Presumed Consent for Organ Donations

Mary Steffel, Elanor F. Williams, and David Tannenbaum, “Does Changing Defaults Save Lives? Effects of Presumed Consent Organ Donation Policies.” Behavioral Science and Policy 5(1): 69–88, 2019 [pdf here].

• The chief source of kidneys for transplants come from deceased (cadaveric) donors – approximately one-quarter of kidney transplants involve a living donor. (A cadaveric donor, of course, is often able to donate two kidneys.) 

• There is a kidney shortage. More than 90,000 Americans are on the waiting list for kidney transplants, and thousands of them die from conditions that a timely transplant would have averted.

• Many people who survive the wait list nevertheless wait for years (often more than five years) for their new kidney. During their wait, they are maintained (to some extent) by dialysis. 

• Most Americans, it seems, favor a presumed consent system – an opt-out system – for cadaveric organ donations. [FWIW, I do not, nor do nudgers Thaler and Sunstein.]

• No US state uses a presumed consent system; rather, opt-in (explicit consent) or active choice mechanisms are employed.

• Presumed consent does, at least, “make it easy” for (what is apparently) the majority of the population that wants to be on the organ donor list. Nonetheless, presumed consent can seem, well, presumptuous.

• With presumed consent, family members (next of kin) of potential donors might be unsure of their deceased relative’s wishes, given that there is likely not to be an explicit indicator of those wishes.

• Donor lists tend to grow (markedly) when presumed consent becomes the rule – although many of the new additions to the donor roster did not get there through an active choice.

• The procurement of organs for transplant, and the number of transplants, also seem to be enhanced (though not so markedly) by a presumed consent system.

• Active choice is where people are encouraged to explicitly say yes or no (or perhaps yes or later) to a query about whether they would like to be an organ donor.

• Mandated choice is where the encouragement to answer is somewhat more forceful.

• An active choice system where the choice easily can be postponed or avoided does not seem to enhance donor lists.

• Surviving family members generally get to make the call on donations, even when the deceased opted into the donor list. It might be sensible to nudge family members, if the goal is more donations. 

Kirgios, Mandel, Park, et al. (2020) on Temptation Bundling

Erika L. Kirgios, Graelin H. Mandel, Yeji Park, et al., “Teaching Temptation Bundling to Boost Exercise: A Field Experiment.” Organizational Behavior and Human Decision Processes 161: 20-35, 2020.

• Temptation bundling came to the world's attention (er, my attention) in a 2014 article by Milkman, Minson, and Volpp. Temptation bundling is when you pair a "should" behavior –exercise, for instance – with a "want" behavior, an indulgence such as listening to a captivating, somewhat trashy novel. The idea is that the pairing makes it more likely that you actually will engage in the "should" activity, and it will be more enjoyable than it would be without the tied "want," while you also will reduce the guilt feelings associated without giving in to an indulgence. Hmm, I might have some (vegan, of course) chocolate while blogging...

• The 2014 article had a pretty small sample size (226) and findings that suggested temptation bundling did provide, in the short-term, a boost to engagement with the "should" behavior.

• The 2020 article (which involves as a co-author Katherine Milkman, part of the original temptation bundling team) works with a commercial fitness company to test temptation bundling on a much larger sample. The experiment sticks with the audio books/exercise pairing, but with about 15 times the previous number of participants.

• A new question that is looked at is whether would-be exercisers actually have to be told about temptation bundling, or can an exercise gym company, by giving people an audiobook in the context of encouraging exercise, induce their customers to temptation bundle on their own? The term of art the authors employ is "information leakage." That is, even without being told or advised to bundle, maybe just being given access to an audio book by your gym induces temptation bundling.

• A chain of gyms introduced a 28-day long StepUp Program for their millions of members, with some rewards for participating. The Program was described as “a habit-building, science-based workout program [p. 23].” StepUp served as the locus for tons of research: "Our study was one of 20 preregistered studies embedded in the StepUp Program... [p. 23]." 

• Within the framework of the StepUp Program, more than 2,300 people participated in a temptation bundling experiment. About half of the participants were informed they had been gifted a free audiobook, and they received a code allowing them to choose and download a book. The other half received the same audiobook gift and were encouraged to use the book in a temptation bundle with exercise; this encouragement continued throughout the month.

• Most people (some 86%) didn’t bother to collect their free audiobook. Now I am sad. In some sense, we are back to a small sample: less than 200 subjects in each of the conditions downloaded their free audiobook. 

• Gym attendance is the main variable of interest. The free audiobook offer, whether paired with temptation bundling encouragement or not, increases workouts by something like 10 to 14%, both during the Program and up to 17 weeks later. (This result comes from a variation of the experiment that includes a third control group of participants who are not offered a free audiobook.)

• The marginal impact of the temptation bundling is pretty small (effectively zero during the 28-day treatment period), perhaps inducing some people to go to the gym (in the long run) who would not otherwise have gone. 

• A follow-up lab experiment establishes that information leakage might be what drives the very similar results for the audiobook treatments, with and without the temptation bundling encouragement. (As the lab experiment was not part of the original pre-registration, presumably its utility was understood only after the  original experiment showed near identical results for the two conditions.)

• In my view, this experiment removes a bit of the luster from temptation bundling. I'm a bit disappointed, really – not in the experiment itself, which is superb, but by the meh, alas, results. 

Ainslie (2021) on Willpower

George Ainslie, “Willpower With and Without Effort.” Behavioral and Brain Sciences 44, e30: 1–16, 2021, pdf here. [This "target" article attracted 26 commentaries, which can be found here. Ainslie's response to the commentaries is here.]

• The basic framework: a Smaller, Sooner (SS) option versus a Larger, Later (LL) option. 

• The SS option is tempting in the moment, even if the LL option is valued more highly at other points of time. 

• Why is SS tempting? Perhaps visceral factors take over, or hyperbolic discounting does its thing. “Preference for the [SS] option is often temporary…[p. 2].” 

• Two different mechanisms, suppression and resolve, are commonly associated with willpower. Habit is a third willpower-adjacent mechanism that can be distinguished from resolve (or from a lack of resolve). 

• Suppression involves avoiding (or blocking or ignoring) the re-evaluation of the two alternatives while waiting for the LL. (Cue the marshmallow test!) Suppression “is necessarily unstable [p. 1],” a “game of keep away [p. 5].” Suppression requires attention, and thus precludes other uses of mental resources that might themselves be rewarding. 

• Resolve refers to sticking with the original valuations by cognitively changing the stakes. “Recursive self-prediction” seems to be a leading approach to resolve, and underlies personal rules. Resolve “is intent that is maintained by an enforcement mechanism [p. 5].” One resolve mechanism is to raise the stakes by viewing giving into temptation not as a one-time deviation, but as the start or continuation of a long series of lapses. (Suppression can be one device for building resolve.) 

• Hyperbolic discounting can be a basis for why people choose the LL alternatives when making category-wide choices, and why they try to bundle individual choices into those larger categories. 

• Scale mismatch makes it hard to overcome current temptations – what would be the harm of eating this tempting dessert? (p.5) The harm lies in undermining the credibility of your weight-control plan. So, there is value in viewing your current choice as a test case. 

• Three can be significant value in (legalistic?) distinctions to permit occasional deviations, so that you can give in to temptation to a degree without destroying the credibility of your long-term plan. (What happens in Vegas...)

• "Resolve becomes effortless to the extent that you are confident of maintaining it [p. 8]."

• Habit is a “routine simplification of action [p. 2]” which is not the same as resolve. With habit, the issue is moved out of the choice domain, behavior becomes arational. Ainslie distinguishes routine habits (like showering in the morning) from good habits ("preserved by resolve" and put at risk by unexcused lapses) and bad habits (repeated impulsive behaviors).