The Game of Life: Discussing Determinism in Behavioral Science

In the debate between free-will and determinism, no answer seems to be satisfactory. If we have free-will, then the world seems chaotic, unpredictable, and dangerous. But on the other hand, if our actions are pre-determined, and everything occurs in accordance to strict logics and causalities, then our future may already be decided for us. The kind of anxiety that follows this realisation, is summed up humorously in the following limerick:

There once was a man who said, “Damn,
It grieves me to think that I am,
predestined to move,
in a circumscribed groove,
and am in fact not a bus, but a tram”.

In the case of determinism, all our “brooding and agonizing” over what is the right thing to do may seem bitterly pointless, because it is in fact our propensities to certain stressors that decide how we will act (Pinker, 2003). As such, behavioral science may hammer home this sense of existential dread, due to its commitment to uncovering the causal patterns between human behavior and its surrounding stimuli. Whilst the discipline may not always adhere to such hard-deterministic logic, it is important to pre-empt how those unacquainted with the field may respond to its casual assumptions.  

You need not oppose or deconstruct the idea of behavioral determinism in order to make people engage more positively with the prospect that their behavior can be predictable. Instead, I suggest that people may benefit from the heuristic findings of an area of recreational Mathematics, known as John Conway’s Game of Life. Its take on chaos, order, and predictability undermines the strict distinction between free-will and determinism, by painting a fascinating picture of how one’s future can be incalculable, even whilst abiding by simple deterministic laws. Conway’s simulation can prove to be a productive PR tactic for those looking to alleviate existentialism and bring behavioral science to a wider audience.

Am I a Bus, or a Tram?

Behavioral science, put simply, is the study of human behavior. The discipline scrutinises human movement scientifically, as a means of approximating the stimuli that determine patterns of individual and group activity (Banerjee, 1995). A core tenant of its scientific rigor therefore is its focus on causality, in that different behaviors are responses to an identifiable stimulus, and occur in a systematic fashion that lends itself to predictive analysis (Simkins, 1969).

This belief that human behavior can be explained in terms of cause and effect — like the atomic space in which it occurs — can be traced as far back as Ancient Greece, with philosophers such as Democritus and Plato (Osler, 2003). Many disciplines and paradigms have similarly emerged with the understanding that human behavior, with all its baffling complexities, can be explained in terms of deterministic associations and laws, the knowledge of which can improve public policy, personnel wellbeing, and corporate efficiency.

Determinism in behavioral psychology, for example, can take the form of environmental determinism, where the source influencing someone’s behavior is external to them — like when Bandura (1961), suggested that violent parents produce violent children (McLeod, 2013). Additionally, deterministic factors can be internal to the person, resulting from unconscious desires and motivations in the subconscious mind (as argued by Freud), or genetic and biological predispositions, such as specific genes that lead to high IQ levels (Chorney et al., 1998, cited in McLeod, 2013), or different personality types that foster specific behaviors (Alarcón, Foulks and Vakkur, 1998).

In addition, the notion of situational determinism in economics claims that the behavior of actors is influenced by the “logic of the situation” (Gustafsson, Knudsen & Mäki, 2003: p17), or that different internal preferences at different times lead actors to pursue a set course of action (Mäki, 2003). By observing these patterns, researchers can predict when people are most likely to behave in certain ways, and what changes in an environment can promote the actions most desired by interested groups.

Overall, whilst these disciplines are in no way uniform in their approach or unanimous in their findings, the general idea that stands out to people who may not know much about behavioral science is that our conscious actions result from environmental and internal stimuli — and are, as such, open to prediction. But what does this mean? Are these causalities set to play out regardless of our desire to act freely? Are our futures already planned? Behavioral science may not always be concerned with these questions, but they undoubtedly play on the minds of those not fully acquainted with the field.

Few would fret over the causal inferences of the natural sciences, such as the Laws of Thermodynamics, genetic predispositions to medical disorders, or the prediction of tomorrow’s weather. However, through implying that behavior is the determined outcome of a multitude of internal and external factors, audiences may interpret that they are not in control of their own decision making. Referring to the anonymous limerick, we may prefer the idea that we are free-willing autonomous agents with the ability to act as we wish, with the freedom that a bus (technically) has over its own movement, and may be hostile to the idea that we are instead destined to act in set ways because of our personal environments and internal cognition, like the constrained movement of a railed tram.

To address this dread, I invite readers to, in the same way I was as an undergraduate, become acquainted with “The Game of Life” — a mathematical automation game introduced by Cambridge mathematician John Conway.

The Game of Life: The Predetermined Chaos of Moving Shapes on a Grid

The game of life consists of a two-dimensional rectangular grid of usually white cells, or pixels, each with the ability to turn on (turn black) or off (stay white). Picture it like a computer screen.

Fig 1.

Off pixels will turn on if they are bordered by exactly three live pixels. Pixels will remain in their starting state if they have exactly two live neighbours, and live pixels will turn off if they either have fewer than two live neighbours and more than four in total (for more see Conover, 2009). The exact mathematical principles of the Game of Life are quite complicated, and may seem irrelevant to behavioral science. However, it is not the maths at play that is important, but the inferences we can gain about the relationship between pre-set laws and the kinds of patterns than can be observed from their enactment.

There is nothing else dictating the activity of the pixels on this grid. No pre-set pattern or graphics program is put in place. Only a small shape or a few pixels are turned on in a small location on the grid before the simulation is run. What follows is sensational. Whilst only adhering to a couple of very basic deterministic laws, the grid lights up in what looks like early 8-bit computer graphics, displaying detailed and explosive patterns that make you speculate that the complexity of the graphical action must all be pre-planned. But, as previously stated, none of the patterns are anticipated or deliberate. All that was pre-set were a few very basic principles determining under what conditions a pixel is to turn on or off. The collection of shapes, formation of entities, and the movement of what looks like the arrangement of cells in the body or avatars on a computer game all have their routes in very simple precepts.

What Conway’s Game of Life shows is that, whilst something can be purely deterministic, in that it can only act in a certain prescribed way, its behavior can still be unpredictable and unimaginably complex. The deterministic nature of the simulation is coupled with a huge sensitivity to a group of pixel’s initial condition, and the slightest difference in the starting state (where the pixels are turned on prior to the simulation, and what shape they form), radically effects subsequent states and patterns once the simulation is run (Ibid). A square shaped collection of live pixels, will move in an immeasurably different manner to a structure of a different shape in a different position on the grid. The kinds of patterns that are shown have a seemingly infinite number of possibilities in their movement, and how they interact with other strands of patterns.

Applying the Game of Life to Behavioral Science

Now, when applying this to behavioral science, this is not to say that human behavior is reducible to a handful of consistent laws and directives, like the pixels in The Game of Life. As stated earlier, Conway’s simulation provides an accessible heuristic tool to demonstrate that causality is more complicated than people may think. Though, in The Game of Life, we can say with certainty that any square on the grid will turn on under conditions A, stay as they were in condition B, and turn off in condition C, we cannot then predict with similar certainty what shapes and patterns we can expect to see from a specific shape occupying a certain starting point.

Similarly, just because behavioral science teaches us that we can expect individuals to behave in a certain way given a certain environment, it is not then to say that we can anticipate the extent of the ensuing behaviors and how they will change and adapt when met with others adhering to a similar deterministic logic.

Applying the Game of Life Beyond Behavioral Science

For example, from approaching the criminal justice system with research about how changing criminal law does not deter individual criminal conduct (Robinson, 2004), or proposing to a business that a customer’s religious commitment determines the importance they place on personnel friendliness/assistance when evaluating their services (McDaniel and Burnett, 1990), the audience may infer that these associations are certain, whilst the sample demographic may resent being portrayed as so predictable. Behavioral outcomes that end up deviating from expectations may fuel resentment toward the usefulness of the science, and individuals may express discontent as being treated as mere formulaic axioms.

Using the heuristics of simulations like The Game of Life, can convey that attributing behaviors of interest to fundamental deterministic laws does not mean that we can fully account for or predict how those laws will play out in a wider context. Essentially, The Game of Life can help to teach the importance of vigilance and open mindedness about the predictability of asserted causalities. This way, audiences adapting to new insights from behavioral science will not be so surprised or feel “cheated” if the outcomes are different to what may have been approximated.


The AI Governance Challenge

This is the essential take-away from Conway’s Game of Life for behavioral science audiences. Just because human behavior may have a quite simple deterministic origin, it is not to say that resulting behavior will be uniform, stable, and always predictable. The future therefore, is still largely unknown. We may provide hypothetical conjectures about the basic laws of human interaction, but there are still many such laws that we do not know. Conversely, we may not know the full extent of outcomes relating to those laws that we do know. This isn’t a critique of behavioral science research itself, as such limitations of the causal and predictive claims of any piece of research are always at the forefront of scientific assessment and enquiry. Instead, this heuristic device applies to the perceptions of the audiences who could benefit from a better understanding of the field.

In the Game of Life, as in life itself, we may formulate certain axioms about how things behave in a given environment, and from these axioms infer certain isolated outcomes (e.g., Condition A will make pixel X turn on). But, on the whole, the product still involves a high degree of context-specific randomness (e.g., from where did the game begin?). The dual challenge facing behavioral science is to formulate those axioms such that they reflect the true sequence of human action in a given environment, but also to acknowledge the limitation of those insights when extrapolating to other contexts. Ultimately, as a tool for formulating policy, behavioral science may be less concerned with whether human beings are more like buses or trams, and more concerned with making sure we get where we have set out to go.

The Role Of Thought Confidence In Persuasion

The aim of persuasion is to bring about behavioral change by altering someone’s attitude or opinion. People tend to have attitudes about everything, from food to religion, fashion to race – but, of course, our attitudes are not always the same. This makes understanding persuasion important as a tool to resolve political conflicts, increase social harmony, and reduce health problems through public health campaigns. Historically, persuasion studies have focused on how much people engage with the message (extent of thinking) and how positively people receive the message (direction of thinking).

A more recent wave of research, however, is taking a step back, and examining what is called meta-cognition.

Meta-cognition is ‘people’s awareness of and thoughts about their own or others’ thoughts or thought processes’ (Petty & Briñol, 2004).

One branch of this research is called thought confidence, otherwise known as the ‘self-validation hypothesis’ (Petty, Briñol & Tormala, 2002) – something that has been shown to affect the degree of attitude change. This emerging concept can therefore further our knowledge of how best to instil long-lasting attitude change.

 The Elaboration Likelihood Model

This research stems from one of the dominant models of persuasion, the Elaboration Likelihood Model (ELM) (Petty & Cacioppo, 1986), which argued that your mind can take one of two routes when encountering a persuasive message. The route you take depends on how motivated and able you are to process the message. If your motivation and ability are low because, for example, you are uninterested or distracted, then you will take the peripheral route. This usually results in only a temporary attitude change because your attitude will be affected by heuristic cues, or mental shortcuts, that are often associated with the source delivering the message.

However, if your motivation and ability are high because, for example, the topic is very relevant to you, then you will take the central route. This means that your attitude will be affected by the content of the message, and typically leads to permanent attitude change. Under these conditions, the source can influence attitude change through various psychological processes, which include self-validation. The confidence we have in our thoughts about a message can be manipulated by three source factors: credibility, similarity, and power.

 Don’t you believe me? – Credibility

Credibility comprises trustworthiness and expertise, and often refers to the believability of the source. Tormala, Briñol & Petty (2006) discovered that a high credibility source increased a message’s perceived validity. They claimed that this made participants feel more confident in their thoughts about the message, which then enhanced its persuasiveness.

But what is more interesting is the finding that high credibility sources were only more persuasive than low credibility sources when they provided a strong argument. As part of the same study, participants read a message promoting a new aspirin product, which contained strong or weak arguments, and wrote their thoughts down about the message. Half of each group were told that the message was taken from a pamphlet produced by a federal agency that conducts research on medical products (high credibility), while the other half were told it was from a class report written by a student (low credibility).  

All participants then had to report the confidence they had in their original thoughts. They found that ‘if a message is weak and produces primarily unfavourable thoughts, high source credibility can actually backfire and be less persuasive than low source credibility’ (Tormala, Briñol & Petty, 2006). It was argued that this was because a high credibility source increased confidence in the negative thoughts generated about the weak argument.

 Great minds think alike – Similarity

Social attractiveness – in particular, similarity – is another integral dimension of persuasion, because it promotes cognitive responses. Unlike credibility, the individual becomes the source of validation, rather than the source of the message itself.

Petty et al. (2002) manipulated the extent to which participants perceived their thoughts as similar to other participants’. Undergraduates from Ohio State University read a strong or weak message in favour of comprehensive exams and then listed their thoughts about the message. Half of the participants in each group were told that their thoughts had been rejected for future reference because they were only 8% similar to other students (relatively low). The other half were told that their thoughts had been accepted for future research because they were 87% similar to other students (relatively high).

All participants were then asked to think back and rate their confidence in their original thoughts. Results revealed that attitude change was greatest for participants who were told that their thoughts were similar to other participants’. They argued that this was because those who knew that others held similar thoughts — whether positive or negative — had more confidence in those thoughts. Participants then relied on them more during attitude formation compared with those who perceived their thoughts as dissimilar to others.

 Who’s in charge here? – Power

The last dimension is power, and, much like similarity, it is about the individual as a source of validation and how their perceived power is manipulated. Power is vital in persuasion because it plays a pivotal role in human relationships, and studies have shown that high power sources are perceived as more persuasive than less powerful sources.

Taking this one step further, Briñol et al. (2007) suggested that ‘power leads to more confidence in whatever actions one is considering’. Participants were required to list thoughts that were either in favour of or against a new vaccination policy. Half of the subjects were asked to recall an incident when they had power over someone else and conversely the other half were asked to recall an incident when someone else had power over them. Finally, they were told to rate their confidence in their initial thoughts. The attitudes of high-powered individuals showed more differences between pro-arguments and counter-arguments than the low-powered individuals because, once again, they felt more confident relying on their recently generated thoughts.

Interestingly, however, when the power manipulation took place before people read the message, high-powered individuals showed fewer attitudinal differences because they were so confident that they did not engage with the message. On the other hand, the low-powered individuals became motivated to focus and displayed greater attitude changes.

 ELM and thought confidence in action

The ELM has provided a more sophisticated understanding of how attitude change works. It has helped develop health promotion campaigns and interventions pertaining to issues such as dental flossing (Updegraff, Sherman, Luyster & Mann, 2007) and eating disorders (Withers, Twigg & Wertheim, 2002). It has also been applied to the domain of social change, for example, with the impact of entertainment-education messages (Slater & Rouner, 2002) and workplace aggression (Douglas et al., 2008) as well as commerce, including the role of personality traits and perceived values in online shopping (Chen & Lee, 2008) and trust in mobile banking (Zhou, 2012).


The AI Governance Challenge

More specifically, Briñol & Petty (2009) claimed that ‘research conducted on self-validation has examined the effect of thought confidence with regard to a variety of attitude objects and measures, increasing the potential applicability of these results in the real world’. Indeed, there is evidence that its applicability is far-reaching. Covarrubias & Fryberg (2015) used the self-validation hypothesis to explain why Native American students’ sense of school belonging increased with the number of role models they identified. Blankeship, Nesbit & Murray (2013) applied the research to understand the relationship between aggression and driving, where those who displayed anger were more confident in their thoughts during a provoking situation. Finally, Santos and Rivera (2015) found that people were more likely to punish an anti-social target when they were exposed to the idea of ethics and legality, because it validated their negative thoughts.

The art of persuasion has been a point of both interest and mystery for centuries due to its powerful ability to change people’s minds. The discovery that thought confidence affects attitude change highlights the complexity of the process because it is no longer simply a question of getting people to think positively about a message, but how best to increase the confidence in their thoughts. Its application to the world of persuasion is still in its infancy, but the self-validation hypothesis undoubtedly sheds light on the mechanics of attitude change and offers a promising solution to developing effective persuasive messages.

The “Mystery” of Intuitive Decision Making

A High Stakes Hunch

I had just got the news: the bank I worked for was going to sell the business I was supporting. This was going to be the seventh transition of my 26-year banking career. Should I try to find a new job in the industry or start my own business? Somehow, I just knew the answer right away: I wanted to start a business that could potentially have a huge impact. It was only later that I could come up with reasons (or rationalizations) for taking such a leap.

Was this intuitive decision rational? When is it best to use your gut feelings to decide? How should you use your intuition to best effect? What is intuition anyway? 18 months later, I can tell you my decision was high-quality because it allowed me to move quickly, without wasting lots of time on unhelpful analysis, so I could get to work on designing the next stage of my career and life.

Intuition Versus Deliberation

Intuition has been described as “unconscious intelligence.”[1] It’s a nonconscious feeling that quickly, automatically, and, without effort, motivates you to act. It doesn’t lend itself to logical argument, reasons, or even language.[2] It’s holistic, concrete, and was formed deep in our evolutionary history.[3][4] It is often about recognizing a pattern from experience and knowing instinctively which rule of thumb to apply.[5] In my case, based on my experience making past career decisions, I used the take-the-best heuristic (rule of thumb) and chose to start a social enterprise because it could have the greatest potential impact compared to my other options.

[Intuition is] the mystery of knowing without knowing

Deliberation is the opposite of intuition. It’s slow, hard work, more or less divorced from emotion, and can be defined by careful analysis and logical arguments. It’s something humans have evolved more recently. In my case, I used analysis only after the fact to confirm (or rationalize) my decision to myself, my family, and friends. Immediate deliberation would have only slowed me down, confused me with irrelevant information, and increased the risk that I’d convince myself to make the wrong decision.[7][8]

When Can You Rely on Your Intuition?

It’s best to use intuition to make decisions when you have the right kind of experience and context for a given decision. Have you had enough practice to learn which decision rules usually work and which don’t? Did you get consistently complete, accurate, and timely feedback, which you were able to appreciate without bias? [9] Over my longish career, I’ve made a lot of career decisions, the consequences of which —positive and negative—were clear, and which I tried to accept non-judgmentally.

Intuition is better than deliberation when the situation is fundamentally uncertain. We’re talking about “Knightian uncertainty” where some of the alternatives, probabilities, and outcomes are unknown, making algorithms and analyses worthless. This describes many, if not most, personal and business decisions.

Consider the investment allocation decision. The future is almost completely unknown and profoundly unknowable. But, we do have a lot of investment options. This decision is a good candidate for the equal-weights (or 1/N) rule: just spread your investment equally among the options. [10] It turns out this strategy performs at least as well, if not better, than more complex analyses in many investment contexts. [11] Finally, use intuition when it’s OK that the answers are approximate, and in domains that are more qualitative and holistic such as ethics, aesthetics, and social decisions. [12]

Using intuition is tricky for analytical people like me. First, you need to be aware of your feelings and the “vibes” of the situation. For a given choice, does it make you feel more comfortable, relaxed, engaged, safe? Or does it rather cause tension, revulsion, malaise, or pressure? [13]

Mindfulness Sharpens Your Intuitive Decision Making

The practice of mindfulness can help you use your intuition effectively. Mindfulness is, “The state of being openly attentive to, and aware of, what is taking place in the present, internally and externally.” [14] It also means that you are observing non-judgmentally and with calm detachment, as if “from a balcony.”[14]

Mindfulness can help reduce the risk of cognitive biases causing mistakes. For example, being accepting of one’s own limitations can reduce the risk of overconfidence. Some research suggests mindfulness can also help offset present bias, stereotyping, and the sunk cost fallacy, while improving learning from experience. [14] Mindfulness can also address behavioral hurdles: For example, my yoga instructor uses her practice to “cool her emotions”, which helps her avoid procrastination and connect with her intuitions to make better decisions.

Give Deliberation a Vote Too

Speed is a key feature of intuition. But this doesn’t mean that deliberation should be ignored. On the contrary, if there’s time and it’s the kind of decision that can be fruitfully analyzed, then deliberation can be a good check on intuition. This is especially the case if you don’t have enough quality experience to rely on. Use deliberation to ensure that you’re being reasonably consistent, you’ve considered the downsides, and controlled for biases — for example, by seeking disconfirming evidence. [15] You can also use it after the fact to rationalize and get buy-in from stakeholders who just aren’t persuaded by your gut feelings.


The AI Governance Challenge

In sum, while deliberation, analysis, and calculation can be useful for decisions where there is plenty of quality information, analytical tools, and time, it is not the sole measure
of quality decisions. Intuition is not only helpful but sometimes offers better decisions, precisely because it is fast, disregards some less relevant information and reflects the fundamental uncertainty surrounding many of our decisions. “The brain is a democracy,”[16] with different capacities contributing to our wellbeing and that of those we care about. For all our decisions, we ought to use all of our capacities to make them the best way we can.

Want to learn more about Decision Fish? Got personal finance questions? Want to get in line to try our app? Work at a company that does not yet offer financial wellness tools to employees? Please, let us know at!

How Does Society Influence One’s Behavior?

Behavioral economics and nudge theory gets a bad reputation. Sometimes vilified as dark marketing, government interference, or self-serving paternalism, fears arise around the notion that such interventions infringe on individual rights.

Yet, the fact is that we very rarely make choices in isolation of outside influences.

We are social beings, and thus, our choices are made in the context of social connections, personal relationships, and physical environments — all of which will have been influenced by other people.

Indeed, the very concept of behaviorally-influenced public policies, and the extent to which these can be effective, demonstrates how individuals respond to outside agency. Behavioral economics harnesses these human insights, and works on the premise that — both to help people individually and to have a positive impact on the widest number of people — individuals’ behavior can be influenced without restricting their liberties.

For example, when it comes to taxes on alcohol or sugar consumption, some argue that their body is their own, and thus they should be left to make their own decisions. To be sure, public health policy aims to provide the individuals with the utmost freedom in cases where the negative consequences of their behaviors can be internalized. However, it also holds that if there are externalities, or public costs, to these behaviors (as there often are), the government is justified in campaigning to reduce the incidence of such behaviors. Thus, it is not only that social forces influence our behaviors, but that, in turn, our behaviors impact societal outcomes.

The UK’s Behavioral Insights Team (BIT) use the framework MINDSPACE to aid the application of behavioral science to the policymaking process. They argue that the ideas captured in the mnemonic are ‘nine of the most robust influences on our behavior.’ These are as follows:

Whether it is acting in line with social norms, seeking ways to act that make us look good to others, or relying on category and perception to form our opinions of those with whom we interact, it is clear that these social components have an outsized impact on our individual behaviors. Let’s look at a few examples:

Messenger – interaction with others

In another classic example, the UK’s BIT worked with HMRC (the UK’s tax collection agency) to increase tax payments by sending out reminder letters stating that most people in the recipient’s area had paid their tax. The impetus for this intervention is the simple insight that no one wants to be the naughty individual in their community, and that reframing tax payment as not only a legal obligation but a social norm would increase compliance. The letters emphasizing the positive social norms produced a 15% higher response rate than the standard letter, and it has been estimated that if the approach was taken across the country, it could help to collect £160m extra tax revenues per year.

Norms – peer pressure

In another classic example, the UK’s BIT worked with HMRC (the UK’s tax collection agency) to increase tax payments by sending out reminder letters stating that most people in the recipient’s area had paid their tax. The impetus for this intervention is the simple insight that no one wants to be the naughty individual in their community, and that reframing tax payment as not only a legal obligation but a social norm would increase compliance. The letters emphasizing the positive social norms produced a 15% higher response rate than the standard letter, and it has been estimated that if the approach was taken across the country, it could help to collect £160m extra tax revenues per year.

Commitment – a public declaration

Similarly, facilitating the creation of social norms is part of UNICEF’s approach to challenging the practice of female genital mutilation (FGM) in some African villages. According to the NGO Tostan, a key factor leading to the abandonment of the practice was addressing collective rather than individual behaviors. Public condemnation of FGM and declarations against it were found to have a significant symbolic value.


The AI Governance Challenge

Defaults – tipping point

The Social Cognitive Networks Academic Research Center (SCNARC) analysed vast quantities of data to identify the tipping point at which a marginal belief becomes the majority opinion. Their estimate suggests that at least 10% of people have to hold an opinion for it to have a chance of being adopted more widely.  Thus, they argue, a small group can create change — so long as they are committed and consistent in their belief. Perhaps the most effective way to achieve widespread modification of behavior is to reach those 10%. If word of mouth is the best form of advertising, obvious and clear actions could be the best form of encouraging social change.

What these results all suggest is that, though we like to think of our choices as our own, in fact, they are often profoundly impacted by the choices and views of our peers. In that way, John Donne was right — no man is an island. Especially when it comes to behavior.

Can Peer Pressure Save Lives?

This article originally appeared in [] and belongs to the creators.

When we hear that someone succumbed to peer pressure or conformed to group expectations, we are inclined to think about it in negative terms. We imagine a young person smoking his first cigarette or an adult parroting the consensus of her community. We know that these social forces can cause people to act in ways that are harmful to themselves and others; but every day we are discovering more ways that they can be harnessed to solve problems in health, education and other areas. This is crucial. For decades, development organizations have spent billions of dollars developing medicines, installing wells, or building clinics or schools that people have not fully used, if they have used them at all.

Providing the right tools to fix a problem is only part of a solution, and often the easy part. Changing behavior is much tougher. Consider water. Impure water can cause diarrhea, which kills 760,000 children under 5 each year. To curb transmission of waterborne diseases, many governments and donors focus on building wells and other water sources, but one big problem is that water is often recontaminated when people transport and store it. There is a relatively simple solution to this problem: chlorine.

Social norms have so much sway that it’s possible to get people to change their behavior simply by telling them what the norm is.

It’s not expensive. In Kenya, for instance, the cost of chlorine for a family of five is about one cent per day. Despite the fact that it would save many lives, and reduce illness, most people do not use it to treat their water. One organization, Innovations for Poverty Action (IPA), based in New Haven, Conn., applied behavioral science to the problem. They developed a new chlorine dispenser with a convenient delivery system and a valve calibrated to release a set dose, making it simple to treat a 20-liter container of water.

But they also went further: They installed the dispensers at communal water sources, where neighbors could see one another using it, and feel pressure to follow suit. They enlisted a community member to be a “promoter,” whose job is to refill the chlorine tank each month, teach the community about the importance of chlorine, and report problems to the local health ministry.

The combination of a convenient, free device and social pressure to use it changed people’s behavior. In a randomized control trial, IPA found that two years after installing the dispenser, 61 percent of sampled households had chlorine in their water, compared to less than 15 percent of households in the control group.

“If you accept the basic framework that we make decisions to maximize our happiness, there are two parts that incorporate other people,” said Dean Karlan, a Yale economics professor who is the founder of IPA. “One part is that our happiness isn’t just a function of what we eat, drink and consume: it’s also our image to others, and our reputation. The second way that people influence decisions is through their information networks. I get information from friends, and that information will affect the decisions I make. [Many public health] interventions are using those levers: They’re using peers to send information.”

“You need opinion leaders in a community to do something, which gets other people to mimic that behavior,” added Jeremy Hand, who ran IPA’s safe water program. “The other driver is the idea of peer pressure: if you know that you’re being observed, and the community accepts this behavior as healthy, that peer pressure factor can be a big driver of adoption.”

In many parts of the developing world, behavioral economists are attacking problems including poverty, malnutrition and familial violence, by applying these basic insights — particularly the realization that a primary force governing how we behave is how other people behave. We imitate those we respect. We turn to trustworthy sources for information. We conform to what’s considered normal. And when we feel that someone is watching, we’re more likely to do the right thing — whether it’s putting the trash in the bin or avoiding that parking spot reserved for disabled people.

When a problem is inextricably linked to behavior change, it’s essential to make the solution both convenient to practice, and something that can be socially reinforced.

Social norms have so much sway that it’s possible to get people to change their behavior simply by telling them what the norm is. This has been demonstrated by Opower, a Virginia-based company that has gotten its customers to cut their energy consumption by roughly two percent simply by telling people how their bills compare to their neighbors. Similarly, Tina Rosenberg has reported in this column about how this approach has been used to combat binge drinking at Northern Illinois University. Students there drank heavily at parties — but they believed that binge drinking was more widespread than it really was. School officials built a campaign around the message that “most students drink moderately.” It cut binge drinking by nearly half over 10 years. It is one of many examples showing that social norm campaigns work.

Positive social pressure has been a central part of the effort to eradicate Guinea worm disease, which is transmitted through dirty, stagnant water that incapacitates its victims. (It’s a heinous disease: Waterborne larvae mate and grow inside a person’s abdomen, reaching as long as three feet, before emerging from the body through a lesion on the skin, causing excruciating pain.)

The only way to prevent Guinea worm disease is to convince people to stop drinking contaminated water. Health workers figured out part of that challenge when they devised an inexpensive, cloth pipe filter that they distributed free throughout Sudan and other parts of Africa. But they struck epidemiological gold with a simple behavioral tweak: adding nylon cords to the pipes, so that people could wear them around their necks. Volunteers spread the message that contaminating water is an unneighborly act. Local leaders began wearing the filters, which over time became a symbol of good judgment and respect for the community’s health, according to Dr. Donald Hopkins, the vice president of health programs at The Carter Center. Based in part on these efforts, Guinea worm disease is close to being eradicated.


The AI Governance Challenge

When a problem is inextricably linked to behavior change, it’s essential to make the solution both convenient to practice, and something that can be socially reinforced. In Liberia, for example, 60 percent of women are pregnant by age 19. How do you effectively teach young people about protected sex and contraceptives so that it changes their actions? That’s a problem that Population Services International has been struggling with for years, said Reid Moorsmith, its representative for Liberia.

One shift it has made is delivering its training programs inside youth clubs, and holding “Clinic Celebration Days,” where H.I.V. testing and contraceptives are provided on the spot. There is a significant difference in bringing the clinic to the participants. Not only is it simple, but young people can watch their peers, friends and family members choosing to get tested or obtain contraceptives. According to Moorsmith, 59 percent of the 6,300 women who had participated in Clinic Celebration Days were getting family planning services, compared to 19 percent nationwide.

Many health solutions are simple — or, at least, they seem like they should be. Breast-feeding, for instance, helps build immunity against childhood killers like diarrhea and pneumonia, and it doesn’t cost anything. But while the practice is common, in many developing countries women don’t breast-feed exclusively for six months, which the World Health Organization recommends. Mothers will often feed their babies dried milk mixed with water, or just water — which is frequently contaminated.

Historically, health officials have tried to inform, or sometimes, scare people into adopting a healthy behavior. While it’s important for people to understand risks and causes of illnesses, when it comes to changing behavior, it is often more effective to lead with a message that is clear and aspirational.

So, in Bangladesh, health workers from Save the Children drew attention to happy, fat babies  and their breast-feeding mothers  to encourage other women to nurse during their child’s first six months. The health workers organized birthday parties for 6-month-old babies who had been exclusively breast-fed, and invited the whole community so they could showcase the health benefits and teach other mothers about infant nutrition. Researchers believe that when mothers see the other women breast-feeding, they’re more likely to follow suit. Save the Children reports that after five years of this program, the rate of exclusive breast-feeding had increased from 29 percent to 64 percent.

Positive social pressure can be applied in countless other ways — to increase rates of vaccination, get people to shift to clean cook stoves or to encourage them to educate daughters. When people make these choices, they’re acting on deep human impulses: to be accepted and liked by others, to imitate those we respect, and to connect and fit in with our peers. The stakes are high. Governments and development programs have spent billions of dollars to make lifesaving interventions available around the world. But often people have passed them by. By using the knowledge we have today about what it takes to change behavior, we stand to save many lives.

Why Road Safety And Bank Queues Are Long-Lost Cousins

What if someone told you that you could be safer on the road, without you or your millions of fellow drivers (abysmal as you believe their skills often are) ever moving a finger?

Sweden’s road policy, ‘Vision Zero,’ is structured on the premise that policies should be ‘human-centered’, instead of using classical designs that may penalise the driver without sufficient thought for road safety.

Their creative methods include redesigning roads to minimise risky road ‘overtaking,’ and hiking the frequency of cameras alongside highways to deter speeding (rather than generate revenue) [1].

In traditional economics, people behave with the logical precision of modern day robots and the manners of Victorian saga heroes. Today, however, like in the Swedish case, policymakers and private players are cottoning on to the reality that people are often imperfectly rational (if not entirely irrational), and factoring this into policy design.

That prompts our next question: what do roads and bank queues have in common?

Choice Architecture and Bank Queues  

A few years ago, a bank in Oman asked me to examine the problem of heavy consumer traffic and congestion at some of its branches. Oman, home to 4.15 million people (roughly 56% of which are expatriate residents), makes for fascinating studies on consumer behavior, given its diverse sociocultural makeup [2].

We began with selected bank branches under our microscope: these branches were gasping under the brunt of heavier consumer traffic than they were built to face. Early-stage analysis pointed us in the direction of cash withdrawals and deposits of small denominations.

Imagine an individual needing the equivalent of 50 dollars in cash. She could avail herself of a speedy ATM withdrawal, or wait her turn in a long branch queue to withdraw cash over the counter.

What’s more, she had to conduct an online transaction worth 50 dollars, she could use a mobile or laptop — reasonably ubiquitous devices, particularly for economically able consumers.  

Despite these options, vast hordes of consumers preferred queuing for over-the-counter withdrawals. It seemed like the time these consumers lost whilst queuing was less than the benefit they felt from interacting with a red-blooded human being.

Interestingly, the Omani nations who preferred over-the-counter transactions for smaller amounts (despite hailing from different income segments and ethnic groups), possessed one common trait: they were largely over 40.

Phone-savvy millennials, however, had a greater affinity for ATMs and mobile banking: perhaps unsurprising in a market with a smartphone penetration rate that trumps the global average by 12%.

“What Does Behavioral Science Have to Do With it?”

As our team familiarised ourselves with these behavioral patterns, we found ourselves asking: how can we encourage customers to use ATMs or eChannels (eBanking/ Mobile banking), instead of waiting in the queues (which, let’s face it, was rather unpleasant in our view)?

If successful in our endeavour, we could reduce queuing times, alleviate branch congestion, and improve ‘customer experience’ in one fell swoop. Right?

We shall address whether our theory was correct later. For now, here are the factors we took into account when weighing this decision.

First, we considered the type of move that would be termed ‘neoclassical’ in economics circles: slapping a fee on customers withdrawing from counters. In other words, the same customer who wanted to withdraw the equivalent of 50 dollars would now find it more expensive to do this over the counter, less expensive over a mobile, and free at the bank’s own ATM machines.

When doing this, we asked ourselves a few questions:

  • On Revenues: What happens if we successfully move people away from the branch, but overall, make less money from the fees we charge customers? How do we construct a fee based on the customer’s willingness to pay (WTP), such that we know the exact price at which it is no longer beneficial to them to transact over the counter?
  • On Framing: Can we frame the move in a way that makes customers more likely to behave in a specific manner?
  • On Fairness: Charging someone at a counter is like taxing them for using the service. Yet, they are customers after all, and ought to be entitled to the service. Is this fair?
  • On Customer Experience: What would make the customer happier: interacting with a human or a machine? Are we placing too much of a cost on lost time, and too little of a benefit on the happiness amounting from human interaction?

Second, we considered a move that was less overtly intrusive. What if we had a ‘meeter-greeter’: a red-blooded human at the branch door, redirecting customers to an ATM, if needed? Would the fact that this move is less intrusive and still offer the customer ‘freedom of choice’ make it fairer?

And finally, we considered moves from the seemingly daft (handing out ice cold popsicles during Oman’s sweltering hot summers, whilst standing next to outdoor ATMs, to induce customers to use them) — to the seemingly mundane (branch redesign).

Have you ever wandered into a bank, seen an ATM stare you in the face before you reached the service counter, and been drawn straight to the machine? If so, you were just ensnared by the magical clutches of behavioral science.


The AI Governance Challenge

On Fairness

It would be massively unfair to have this discussion and not broach the subject of fairness. Many believe that what they call ‘libertarian paternalism’ is permissive in both the private and public spheres, as long as there is no compromise with respect to either consumer welfare or freedom of choice.

I still find myself struggling with the question. To me, for instance, standing in a queue is even less stimulating than watching paint dry. Yet, is it acceptable for an organisation to manipulate my behavior, even if directed at mutually beneficial ends?

Richard Thaler, in a 2015 article, gives us a 3-step guide on this front:

Nudges should always be transparent, offer an ‘opt-out’ alternative, and improve the welfare of the target population [3].

This seems like immensely sage advice: yet, some of my quandaries remain. The main one being- how inherently acceptable is the idea of paternalism by a private actor with its own motivations, even when consumer welfare is in play?

I am sorry that I proffer no definitive answer.

All I can do is fervently hope that I have made things a little more interesting for the next time you find your car halting at a red light, or groaning in a bank queue.

Can Nudge Theory Be Applied To Public Health?

This article originally appeared in [] and belongs to the creators.

Behavior change prompted by nudge theory alone is not going to solve complex public health problems such as obesity and chronic disease, according to UK public health leader Professor Mike Kelly.

Professor Kelly, a former Director of the Centre for Public Health at the National Institute for Health and Care Excellence (NICE), spoke at a recent public forum jointly hosted by The Australian Prevention Partnership Centre and the Hospital Alliance for Research Collaboration.

The focus of his talk was the British experience in using nudge theory for public health interventions. Nudge theory became popular after the publication of an economics text, Nudge: Improving Decisions About Health, Wealth, and Happiness, by Richard H Thaler and Cass R Sunstein.

Professor Kelly,  who is an Honorary Senior Visiting Fellow at the University of Cambridge’s of Public Health, said nudge theory was based on the idea that about 80% of human behavior is automatic, with people responding to cues in the environment, sometimes known as choice architecture.  An example was placing a fruit bowl on the front counter in a school canteen to encourage children to buy more fruit. An unhealthy nudge might be placing sweets near supermarket checkouts to encourage children to pester their parents to purchase.

Watch an interview with Professor Mike Kelly:

Nudge theory and public health interventions

Professor Kelly said public health interventions traditionally focused on the other 20% of behavior change, the more deliberate type of decision making in which we are expected to take in public health messages, interpret and process them and then change our behavior, such as stopping smoking or exercising more.

While nudge could apply to microenvironments, such as the school canteen, it also had a macro application in terms of system changes to prevent chronic disease, which is the focus of The Australian Prevention Partnership Centre.

“If you are going to think about nudging in a broader, scaled-up way, you have to think systems-wide change, with the focus on the nature of the built environment, on architecture, planning, transport systems, food systems and so on.”

Professor Kelly said the jury was still out on whether nudge theory could work in public health. “We need more research, especially in terms of physical activity.”

He sees nudge as just one part of the armoury needed to tackle obesity and chronic disease.

“If we’re looking to change obesity and improve rates of physical activity, we need that multi-pronged, multi-layered approach that involves everybody, from governments to individuals,” he said. “All have a role to play, some of which will be about nudges and automatic responses to things in the environment, but some of which will still rely on state regulation.”

Behavioral complexity

He said behavior and changing behavior are complex.

“Humans do things that make their lives easier, including taking short cuts in their thinking and actions and if you can harness those towards healthy ends, all well and good. But that is not the be all and end all of behavior change.

“Tobacco and the reduction in levels of smoking in Australia, Great Britain, the US and Canada have been a consequence of multi-layered and multi-level strategies, including nudge, that have been integrated with each other. This has taken decades and it may take just as long with obesity and chronic disease.”


The AI Governance Challenge

Nudge and ethics

Professor Kelly said the ethics of nudge theory had been questioned by some in public health who claimed it manipulated people to healthy choices without them knowing it.

“The counter argument is that people are being nudged to buy sugary drinks and fattening food in an obesogenic environment that nudges them to eat at every opportunity and no one’s consent has been sought for that, so that’s just as unethical.”

This Is Your Brain On Money

If there is one thing behavioral research has taught us, it is that human behavior is not always rational. Our judgement and decision making skills are fallible, and based on context, can fluctuate. Take, for example, how our brain understands money.

Mental accounting, which plays an instrumental role in helping us make financial decisions, explains how we treat money differently (place it under different categories) depending upon its source and its intended use. A classic explanation of the concept uses the example of movie tickets:

Imagine you just arrived at a theatre and as you reach into your pocket to pull out the $10 ticket you purchased in advance, you discover that it’s missing. Would you fork over another $10 to see the movie? Compare that to a second scenario in which you did not buy the ticket in advance, but when you arrive at the theatre, you discover you had lost a $10 bill on the way. Would you still buy a movie ticket?

Kahneman & Tversky (1983) used this hypothetical in their research, and although in both cases the amount of money lost equals $10, more people (88%) were willing to buy a ticket in the latter case, compared to the former. The 44% who were willing to replace the missing ticket in the first scenario felt the cost of watching the movie had doubled, since it was drawing from the money (mentally) allocated for movies, which was not the case with the lost cash.

Mental accounting also explains why small windfall gains, like a $50 lottery win, or gift cash from a friend, are more likely to be spent easily, as they are considered unexpected gains, rather than regular income. By creating mental accounts, we essentially ignore the fact that money is fungible, i.e. that all money is the same, and interchangeable.

The phenomenon also helps us understand why we tend to treat credit card payments differently from cash payments. First, credit cards “decouple” the purchase from the payment, by separating the two and delaying the payment to a later point in time. Second, they makes individual costs less salient; a $50 purchase on a $1000 bill has less impact than a $50 purchase by itself. This has its basis in loss aversion – that is, in our mind, losses are more salient than gains, and we usually seek ways to make them less noticeable, which is why credit cards are useful.

Mental accounting is just the tip of the iceberg; there are several other forces shaping our financial decisions. One key game-changer in the psychology of money is the mode of payment. In this article, we will explore two common forms of payments – cash and credit cards. We will also look at how concepts like opportunity cost, pain of paying, and others, factor into our decisions under the two contexts.

Spending with Cash vs Credit Card

Researchers at MIT (Prelec & Simester, 2001) studied willingness to pay (WTP) when using both credit cards and cash. They set up an auction for tickets to sporting events, and restricted the form of payment to either cash or card for each participant. Interestingly, participants using credit cards were prepared to pay almost twice more than those who were paying cash were willing to. That is, their WTP was almost double. These results have been corroborated in a number of studies (Raghubir & Srivastava, 2008; Finkelstein, 2009); put simply, we are ready to spend more when using credit cards. This spending behavior can be explained by the pain of paying – the moral tax (emotional distress) we experience when money is spent. Using a combination of brain imaging techniques, priming, and placebos in a series of experiments, Nina Mazar and colleagues (2016) found that pain of paying is not merely a metaphorical concept, but individuals truly experience psychological pain when making monetary purchases. They found evidence that anticipating paying with money (making the decision to purchase) did indeed activate pain processing regions in the brain, albeit those were associated with higher-order, affective pain, and not somatosensory (i.e., physical) pain. When participants were primed for affective pain, their WTP decreased, further confirming the affective nature of pain of paying.

Credits cards encourage spending because they reduce pain of paying (this is linked to loss aversion). When we hand over a ten dollar bill at a store, we see the money going away. Such transparency of payment is absent in credit cards. For one, we do not observe the money disappearing. Moreover, credit cards are always returned to the owner, further reinforcing the notion that we are not losing money. When researchers (Shah, Eisenkraft, Bettman, & Chartrand, 2015) attenuated the pain of paying, they observed a decrease in loss aversion, and increase in risky financial decisions. By making money less tangible, cards reduce the pain of paying, thus encouraging spending.

The potential future opportunities we forgo (the cost) when we expend a resource (money or time) is known as opportunity cost. We don’t always consider opportunity costs when it is crucial to do so (like when we are saving to buy a car, but spend money on an expensive handbag instead), but sooner or later, these alternate realities become apparent. Cash transactions prompt us to contemplate the consequences of financial decisions.  When you only have $10 in your wallet, the decision to make one purchase would significantly affect the ability to make another purchase, making the opportunity cost of your purchase more salient. Credit cards don’t make opportunity costs prominent, hence increasing the tendency to spend without sufficiently weighing the upshots.

Psychological Factors of Spending

Although at a glance, cash may appear to be the best form of payment for responsible spending, there are other psychological factors that must be considered.

Previous research has established that we tend to overvalue our own possessions – a bias called the endowment effect. It appears that the mode of payment when purchasing can influence how much we value an item. In a study examining how “connected” we are to items bought using different modes of payment, participants, after purchasing a discounted coffee mug, were asked to quote a price to sell it back (Shah, Eisenkraft, Bettman, & Chartrand, 2015). Those who had paid with cash demanded almost $3 more than those who used credit, suggesting that we place more value (endowment effect), and are more connected, to purchases made with cash.  

In a study yielding somewhat surprising results (Bagchi & Block, 2011), researchers studied whether pain of paying had any effect on impulse buying for immediate consumption. They manipulated payment mechanisms, difficulty of earning money, and level of “indulgence” (calories) in food items in a series of experiments. When the pain of payment was higher (i.e., when they used cash), participants indulged more, suggesting an attempt to offset the imputed cost of paying with cash.

Such studies provide insights into the complexities behind financial decision-making, beyond simple behavioral economic principles.

Leveraging Psychological Biases for Better Financial Decision Making

Behavioral research has worked to uncover the different ways in which we think about money. This knowledge can be used to ‘hack’ our behaviors to bring about financial and mental well-being.

Mental Accounting:
The World Bank ran a study in Kenya to increase savings for health expenses, in which they provided residents with lockable metal boxes, a key, and a passbook to record saving goals and deposits. They found that, besides goal setting and increased security, the act of labelling the money for a specific use inhibited using the money towards any other purpose – coined the labelling effect. By exploiting mental accounting, where money belonging to different categories are perceived differently, the researchers succeeded in increasing savings. Similarly, partitioning money for different, specific purposes can motivate saving behavior.

Pain of Payment:
Pain of payment varies with mode and time of payment. By manipulating these factors, depending upon the context, we can maximise financial well-being and satisfaction. In a survey estimating enjoyment of visiting a health club when comparing a monthly fixed fee to hourly payments, respondents judged that the monthly fixed fee would lead to higher enjoyment (Prelec, 2009). Pre-payment for a continued experience can reduce repeated reminders of the cost, which isn’t the case in a pay-as-you-go system. Payment by credit card would further reduce the pain of paying in this scenario. For general spending with credit cards, setting up text notifications with information on amount spent and remaining balance can make the pain of paying and opportunity cost more salient, thus facilitating better financial decisions.

Friction Costs:
Simply put, friction costs are the effects of barriers to decision making. For example, although we want to save money, once the cash is in hand, it is difficult to calculate opportunity costs (financial trade-offs, i.e., choosing between a coffee at Starbucks, vs. one at Dunkin’ Donuts, or a homemade one), thus posing a difficult barrier to saving, as the temptation to spend is greater than the mental energy required to save. Friction costs can be reduced by setting up automatic savings before receiving the money (pre-commitment). This can almost double savings (Beasley, De La Rosa, & Berman, 2017), because you’ve never seen the money in the first place, hence reducing temptation, and the loss aversion associated with having to allocate this newly-received cash to a distant outcome (i.e., money saved in the future). On the other hand, escalating friction costs can also increase account deposits. During a study with Latino Community Credit Union members (Beasley et al., 2017), those seeking to cash their pay checks were required to fill out a check cashing slip, which included a suggestion about depositing some amount of the check into an account (introducing friction into the process). This small message resulted in an average deposit of $167, which has far-reaching implications for long term account deposits.


The AI Governance Challenge

The Power of Technology:
A new app in the market – Qapital, applies behavioral insights to increase savings. The app works by making opportunity costs salient, thus encouraging savings and responsible spending. According to the “Guilty Pleasure Rule” of the app, every time you indulge in a guilty pleasure, you also save a certain amount of money. Another feature rounds up the change after every purchase, which the app then automatically puts towards saving. With several such features loaded into the app, it serves as a handy reminder to make prudent financial decisions.

The above examples offer a glimpse into how small changes in the system, or useful tools like mobile apps, can help us make better choices. Research into financial decision-making and the psychology of money is a growing subfield in behavioral economics, with innovative applications aplenty.

Going Forward…

In a world that is getting increasingly creative with triggers for spending, it makes paradoxical sense that we would turn to our own biases to help circumvent the ever-growing culture of consumerism.