Nudges Work (and practitioners know exactly how well)


At TDL, our goal is to make behavioral science accessible to the masses. This article is part of a series on cutting edge research that has the potential to create positive social impact. While the research is inherently specific, we believe that the insights gleaned from each piece in this series are relevant to behavioral science practitioners in many different fields. As a socially conscious applied research firm, we are always looking for ways to translate science into impact. If you would like to chat with us about a potential collaboration, feel free to contact us


Behavioral science is opening up new lines of inquiry across all sorts of areas, both in academia and in the public sector. One of the places that behavioral science has been most effective is in public policy. Most famously represented by the Behavioral Insights Team, behavioral scientists are changing the content and delivery of policy all over the globe. As a socially-conscious applied research firm, TDL is interested in using empathy, technology, and design-thinking to promote better outcomes in many aspects of society, from health to education to the economic empowerment of disadvantaged groups. To amplify these impacts even further, we reach out to experts currently conducting research in areas that engage behavioral science in the pursuit of socially conscious goals.

With this in mind, The Decision Lab touched base with Elizabeth Linos and Stefano DellaVigna, two prominent academics who study economics, public policy and behavioral science. 

Dr. Elizabeth Linos is an assistant professor of public policy at UC Berkeley. Her research lies at the intersection of public management and behavioral science, which involves using tools from behavioral science to improve government service delivery. She was formerly the VP and Head of Research and Evaluation at the Behavioral Insights Team in North America, where she worked with government agencies in the US and the UK to improve programs using behavioral science and to build capacity around rigorous evaluation. 

Dr. Stefano DellaVigna is the Daniel Koshland, Sr. Distinguished Professor of Economics and Professor of Business Administration at the University of California, Berkeley. He is the co-director of the Initiative for Behavioral Economics and Finance and a co-editor of the American Economics Review. He has studied the economics of the media, the design of model-based field experiments, the analysis of scientific journals, and reference-dependence for unemployed workers.

In their article, Dr. Linos and Dr. DellaVigna investigated the effectiveness of nudging based on data provided by two of the largest Nudge units in North America. 

A full version of the article is available here:


Nathan: How would you describe the focus of your research to a general audience?

Dr. Linos: The focus of my research is how to use what we know about how people actually behave, drawing on decades of research from psychology and economics, to improve how governments are able to deliver services. In my case, this means thinking about how to recruit, retain, and support government workers in delivering better services. It also means improving the way governments reach out to residents about programs and services for which they are eligible. 

Nathan: How did you bring those broad themes into a specific project?

Dr. Linos: In this project, we were hoping to better understand the impact of behavioral science units in governments. “Nudge Units,” as they are often called, have become very popular across the world with over 200 units in various countries dedicated to using behavioral science to improve government service delivery. These units have done something many academics have only dreamed of: they have normalized the use of rigorous evaluation (randomized controlled trials) at scale, by running hundreds of well-designed trials in policy areas ranging from education to the social safety net take-up to public transportation. Our goal was to understand what the average effect of these nudges are at scale, to better understand whether this nudge approach can meaningfully make a difference when taken outside of labs and individual academic studies to broad scale-up.

Nathan: Can you give us an overview of your experimental approach?

Dr. Linos: First, we analyzed hundreds of trials (that have not already been published) conducted by two of the largest “Nudge Units” in the US: the Office of Evaluation Sciences and the Behavioral Insights Team North America. These two units have worked at the federal, state, and local levels and gave us full access to every trial they have run since 2015. It’s noteworthy that this involved a remarkable case of transparency and academic documentation. We estimated the average effect of a nudge across all trials and then compared our results to existing meta-analyses of “nudge” trials that are published in the academic literature. The majority of the project explores why there is a large gap between the average effect of a nudge, when you look at Nudge Units, and the average effect of a nudge if you just look at recent meta-analyses. We considered various options: selective publication, the difference in the characteristics of trials, and differences in the characteristics of nudges.

Nathan: What were your findings?

Dr. Linos: First, we found that the average treatment effect of a nudge is statistically significant and positive: across all trials, nudges increase take-up by approximately 1.4 percentage points (or an 8% increase above a control group). If you were to look at recently published meta-analyses of academic papers, the average effect of a nudge would be over 8 percentage points. We find that we can completely close the gap between these two estimates when we consider selective publication in academic papers. That is, there are probably trials with null results or negative results that are conducted by academics that are not published, or even written up. This is an issue that is often called “the file drawer” problem. It leads to an overestimate of the average effect of a nudge in the academically published literature. We can also close the gap by about two thirds by considering differences in the types of nudges run by academics and those run by Nudge Units. Some of these differences go hand in hand with going to scale; for example, trials with in-person interventions are more effective but also much less likely at scale. 

Nathan: How do you think this is relevant to an applied setting (i.e. in business or public policy)?

Dr. Linos: These results are particularly relevant to an applied setting because they provide an optimistic but realistic estimate of what is possible with a nudge. On the one hand, we now have clear evidence that, on average, nudges conducted across a variety of settings and government agencies are effective compared to a well-defined comparison group. This is no small feat. As a reminder, the majority of policies and programs implemented in government are not rigorously evaluated at all, and we often see cases of programs that were thought to be effective but ended up being ineffective, once they were put under the scrutiny of rigorous evaluation. At the same time, the likely impact of any given nudge is probably smaller than what policy makers would predict, if they only looked at the academic literature. This means that businesses or policymakers may need to move beyond “nudges” to achieve a larger impact. Nudge Units themselves acknowledge this — many behavioral science teams and experts are already exploring how to use insights from behavioral science to design better policies, better legislations, and rethink programs as a whole. Nudges are just one small part of the toolbox. 

Nathan: Do you see future research stemming from your study? In what directions?

Dr. Linos: There are many additional questions this research spurs. First, in trying to understand the overall impact of Nudge Units on policy, it’s important to document what happens after a trial. That is, once we have evidence that something works better than the status quo, we want to know how quickly this knowledge spreads to other policy makers and how and when it gets implemented as the new status quo. We’re also interested in more deeply understanding the cost-benefit calculations that would quantify the exact value of a 1.4 percentage point increase in take-up. Last, we hope our research will spur an ongoing conversation about how to document and share results from all trials, irrespective of whether or not they are published in top academic journals. 

Texting Students Towards College Success


At TDL, our role is to translate science. This article is part of a series on cutting-edge research that has the potential to create positive social impact. While the research is inherently specific, we believe that the insights gleaned from each piece in this series are relevant to behavioral science practitioners in many different fields. As a socially conscious applied research firm, we are always looking for ways to translate science into impact. If you would like to chat with us about a potential collaboration, feel free to contact us.


Behavioral science insights can profoundly impact education outcomes — from improving curriculums to making more effective teachers. As a socially-conscious applied research firm, TDL is interested in using empathy, technology, and design-thinking to promote better outcomes in many aspects of society, from health to education to the economic empowerment of disadvantaged groups. To amplify these impacts even further, we reach out to experts currently conducting research in areas that engage behavioral science in the pursuit of socially conscious goals. 

In this spirit, we reached out to Dr. Ryan Yeung to understand how his work at in education policy harnesses behavioral science to improve society. 

Ryan Yeung is currently the Replication Evaluation Specialist in the Accelerated Study in Associate Programs (ASAP) unit at the City University of New York (CUNY). He has previously taught at Hunter College, SUNY-Brockport, and Rutgers University-Camden. He has a Ph.D. in Public Administration and Policy from Syracuse University. His main research interests are in education policy and public budgeting, and finance. 

For this project, Ryan and his co-author evaluated a program to increase enrolment and success in college through texting.

A full version of his project is available here:


Nathan: How would you describe the focus of your research in simple terms? 

Dr. Yeung: We examined how a group of New York City public schools used texting to increase enrollment and persistence in college. The text messages included tips and reminders to fill out the Free Application for Federal Student Aid as well as college tips. The text messaging program was theoretically and empirically grounded in the behavioral science research on nudges, inducements to people to make a specific choice that is in their best interest, for example, filling out the FAFSA after receiving a text message or going to a professor’s office hours.

Nathan: How would you explain your research question to the general public?

Dr. Yeung: We had several research questions in this study. 1) What impact did texting have on college enrollment and persistence? 2) What impact did texting in high school have on college enrollment and persistence? 3) How did response rates to text messages affect enrollment and persistence? 4) How can we improve response rates if we find response rates do indeed affect enrollment and persistence?

Nathan: What did you think you’d find, and why? 

Dr. Yeung: There’s a long literature that suggests text messaging programs have had positive impacts on postsecondary outcomes, so we were expecting to find a positive effect of the program itself on college enrollment and persistence. We weren’t sure if the timing of the messages (if they started in college or in high school) had much of an effect. If you think the messages do matter, it is natural to think that responses to the messages matter as well. Responses indicate the students have read the message and felt the need to ask for help.

Nathan: What rough process did you follow? 

Dr. Yeung: We created three matched groups, students who received text messages in high school and college, students who received messages only in colleges, and students who did not receive messages at all. These groups were matched to create similar groups of students based on demographics and academic achievement in each group. We then compared how each group performed in terms of college enrollment. We used the same matching procedure to estimate the effect of response rates on enrollment and persistence. Last but not least we ran a multiple regression to predict how different types of messages and demographics affected response rates.

Nathan: What did you end up finding out?

Dr. Yeung: Our multiple regression results suggest that although when the texting began did not appear to matter, the texting program increased enrollment and persistence in college. In addition, response rates were positively associated with enrollment and persistence. Finally, we find that the content of the messages themselves as well as individual characteristics affected response rates.

Nathan: How do you think this is relevant to an applied setting (i.e. in business or public policy)? 

Dr. Yeung: Text messaging programs do appear to have positive effects on postsecondary outcomes at a reasonably low-cost, but there are things that can be done to improve their efficacy. Personalization matters. Students were very likely to respond to messages wishing them a happy birthday. Perhaps there are other ways to personalize the messages in the future. Timing matters. You want to send messages that are relevant for students at the specific point in the school year during which they are sent (like during FAFSA time). We also recommend sending out messages early afternoon or on a weekend when students are most likely to see them.

Nathan: Do you see future research stemming from your study? In what directions?

Dr. Yeung: We acknowledge that our study was not a randomized experiment and is instead a quasi-experimental design. Although the results used matched data, we cannot completely rule out potential bias from unobserved factors. An ideal scenario would be an experiment where students are randomly assigned to receive or not receive texts and randomly assigned to when they receive texts (while in high school v. post-high school). The content of the messages may also be randomized to get a handle on the influence of various types of messages. This would have been the best method for controlling for unobservables. Perhaps a future study could use such an approach. Randomization of the content of the messages including personalization would also be helpful to isolate the effects of message types on college outcomes.

Fighting the Looming Antibiotic Crisis


At TDL, our role is to translate science. This article is part of a series on cutting edge research that has the potential to create positive social impact. While the research is inherently specific, we believe that the insights gleaned from each piece in this series are relevant to behavioral science practitioners in many different fields. As a socially conscious applied research firm, we are always looking for ways to translate science into impact. If you would like to chat with us about a potential collaboration, feel free to contact us.


Behavioral science insights can profoundly impact health outcomes — from encouraging prosocial handwashing behaviors during a pandemic to increasing the number of individuals who sign up for health insurance. As a socially-conscious applied research firm, TDL is interested in using empathy, technology, and design-thinking to promote better outcomes in many aspects of society, from health to education to the economic empowerment of disadvantaged groups. To amplify these impacts even further, we reach out to experts currently conducting research in areas that engage behavioral science in the pursuit of socially conscious goals. 

In this spirit, we reached out to Dr. Patricia Cummings to understand how her work at the forefront of antibiotic resistance prevention harnesses behavioral science to improve society. 

Dr. Patricia Cummings is the Director of the Department of Epidemiology Research and Evaluation at Eisenhower Medical Center in Rancho Mirage, California. She received a PhD in Epidemiology from the University of California Los Angeles (UCLA) and a Master’s in Public Health (MPH) with a concentration in Epidemiology and Biostatistics, from the University of Southern California (USC). 

Her academic interests and work have included research and evaluation studies related to behavioral economics and food choices, chronic and infectious diseases, health disparities, and aging-related diseases. Currently, she focuses mainly on infectious diseases and ways to prevent antibiotic resistance.

In this article, Dr. Cummings and her colleagues track the over-prescription of antibiotics and test strategies for reducing this problem. 

A full version of the paper is available here:


Nathan: How would you define the focus of your research to a general audience?

Dr. Cummings: Antibiotic resistance is when an organism that would ordinarily be sensitive to an antibiotic becomes resistant to it. In other words, it is the ability of bacteria to change in a way that reduces the effectiveness of drugs. Antibiotic resistance is an urgent public health threat globally, affecting all countries, regardless of socioeconomic status. One of the primary drivers contributing to the emergence and persistence of this threat is antibiotic misuse. In the United States, a majority of antibiotics are prescribed unnecessarily during flu season for likely viral infections. Antibiotics do not work against viruses, they only work against bacteria. So, our research aims to identify and mitigate the factors that contribute to this issue, so that we can ultimately prevent and reduce the threat of antibiotic resistance.

Nathan: How do those themes narrow into a specific project?

Dr. Cummings: Inappropriate antibiotic prescribing may be influenced by a number of factors, like physician characteristics (e.g., knowledge, memory, training, number of years in practice), but it can also be influenced by external forces, such as a patient putting pressure on a physician to prescribe them an antibiotic when it is not needed. Our research question for this first study was, “are behavioral science interventions effective in reducing unnecessary antibiotic prescribing among physicians in our community hospital setting?”

Nathan: How was your project designed?

Dr. Cummings: One of the behavioral science interventions we implemented was called “peer comparison.” We ranked physicians against their peers (other physicians in the group) by giving them their prescribing data. We also emailed a blinded ranking list to all the physicians in the group, so they could see where they ranked against their peers. We did this by assigning them a random number, so they were not identified in the public email, but could see how well they were doing compared to their peers. We hypothesized that the peer comparison strategy would work best or result in the greatest decrease in unnecessary antibiotic prescribing since providers tend to be inherently competitive in nature. We also believed that by giving them their data, it empowered them to do better (i.e., social desirability) while creating a competition-like environment among their peers. The other interventions we implemented were patient and staff education, and public commitment. These were supportive, but not as effective as the peer comparison strategy.

Nathan: What rough process did you follow?

Dr. Cummings: We were lucky we didn’t need to reinvent the wheel – we adapted a process from existing studies that have been published on this topic. Specifically, we used work published by Dr. Larissa May at the University of California Davis who has done research in this area, as well as Dr. Daniella Meeker and Dr. Jason Doctor from the University of Southern California. Their work helped us to adapt and develop the interventions to our community hospital setting and population.

Nathan: What did you end up finding out?

Dr. Cummings: Our results showed there were significantly fewer inappropriate antibiotic prescriptions written during the intervention period than the pre-intervention period, resulting in a significant decrease in the rate of inappropriate antibiotic prescribing among physicians over a 6-month period. The strategies included in this intervention suggest that utilizing a behavioral science approach for antimicrobial stewardship may greatly reduce inappropriate antibiotic prescribing.

Nathan: How do you think this is relevant to an applied setting (i.e., in business or public policy)?

Dr. Cummings: Like many behaviors, antimicrobial prescribing is a complex behavior that is influenced by a combination of factors. Aside from individual provider factors, there are a number of outside influences that differ by context and setting. What the published evidence currently tells us is that some of these factors lead to better sustainability than others, such as peer comparison. Applying this to other sectors, such as fast food and beverage industries, we know that external factors (e.g., price, convenience, and advertising) have a heavy impact on what we choose to eat. Sometimes these external factors have a greater impact on our decision-making than individual-level factors. The most important takeaway here is to identify and learn what factors have the most influence and to leverage those factors’ effects on people’s behavior. In our case, we leveraged those factors to improve quality and patient safety, as well as contribute to a global public health goal of reducing antibiotic resistance. 


The AI Governance Challenge

Nathan: Do you see future research stemming from your study? In what direction?

Dr. Cummings: Yes, we are about to launch the next phase of this study, which will examine physician-level characteristics that may influence the likelihood of prescribing antibiotics (e.g., age, training, number of years in practice) and we will look at how these factors contribute to the sustainability of the intervention over time. This work will be important to inform the sustainability of these interventions and will provide much-needed data for other community hospitals looking to implement similar interventions.

Tackling Conspiracy Theories Amid COVID-19

Bill Gates, 5G, microchips, global control, coronavirus, human creation. These words may not seem related but they are significant for many conspiracy theorists. However, beliefs alone cannot hurt anyone, right? In fact, they are not as benign as they seem and can lead to several problems, especially in times that are as chaotic and indecisive as the present.

But how can people believe something without any valid proof? It is valid to be skeptical of any particular event, but it is another thing to be hyper skeptical and overinterpret evidence.11 As Carl Sagan, astronomer and science communicator, once proclaimed: “the extraordinary must certainly be pursued. But extraordinary claims require extraordinary evidence.” 

Indeed, that is the rule to follow, but often the opposite happens. Things of great magnitude are affirmed with total conviction, which only reflects inconsistency and contradiction because, of course, these “extraordinary claims” are not supported by “extraordinary evidence”. Sadly, many people see extraordinary claims as absolute truths, which proliferates misinformation. This can have serious consequences.

Faced with a pandemic that requires large-scale behavioral changes and threatens considerable psychological strain, social and behavioral sciences emerge to address this challenge.18  By understanding the psychology behind conspiracy theories and knowing their effects, it is feasible to dismantle them and align people’s conduct with public health recommendations. 

A long history

Conspiracy theories are not unique to our time. They have prevailed for many centuries, spreading more intensely in times of crisis.19 The content of a series of letters sent to the New York Times and Chicago Tribune between 1890 and 2010 reveal that the highest peaks for conspiratorial belief were located in the height of the second industrial revolution and the beginning of the Cold War. If we go further back to AD 64, where Nero, the Roman emperor, intentionally and unjustifiably blamed Christians for burning Rome, conspiracy theories caused many to be sacrificed or burned alive.19  For their part, people who believe in conspiracy theories are not satisfied with official explanations for global phenomena; instead, they prefer to think that malevolent organizations or individuals want to take control of the world.

Social psychology has been unraveling the mysteries behind the different types of irrational thoughts through biases and heuristics.8 The study of conspiracy theories has developed alongside this discovery of irrational behavior. Research led by  Karem  Douglas,3  a professor of social psychology at the University of Kent, suggests that conspiracy theories satisfy certain psychological needs, and not necessarily conscious ones. Namely, the need to understand things, a desire to have control over situations, and the need for a positive self-image.

Other studies have found correlations with personality factors such as meanness, mistrust, openness to experience, and Machiavellian behavior, which is a focus on self-interest at the cost of manipulating others.9 Evidence has even been found to show that people who believe in conspiracy theories are more likely to accept or engage in daily criminal activities.7

In fact, research suggests that conspiracy theories are more harmful than good.3 These can encourage the rejection of conventional medicine, scientific consensus, and democratic organization.3 The combination of these effects makes perfect sense in contrast to the current misinformation “infodemic”17 plagued by villains and miracle cures.

To be in control

Living in uncertainty can be incredibly stressful; consequently, psychologists have ventured to perform different types of experiments to test how we engage with uncertain circumstances.1  Finding yourself in the midst of a pandemic because of an unknown virus that is claiming the lives of hundreds of thousands of people and has shaken the world economy only exacerbates the uncertainty of our lives. Human beings facing this type of situation have used different biases to try to have their inner world controlled and stable.

Research suggests that people dissatisfied with small-scale explanations may need to explain large-scale events with causes of a similar magnitude.10 When institutional explanations feel unsatisfactory, they may create a version of events that accords with the magnitude of the problem.

A perceived lack of control could explain conspiracy theories as an opportunity to reject official reports and allow someone to exert control over information.5  In turn, this lack of control could activate the adaptive ability to see illusory patterns around us to reduce uncertainty.20 People who have the greatest propensity to see these types of patterns are those who have more deep-rooted beliefs about conspiracy theories.21

I know things that “they” don’t

At some point or other, most of us have wanted to stand out from the crowd. This motivation to feel different from others —unequally distributed among people—  is clearly reflected in day-to-day decision-making, whether buying a unique article of clothing that very few can get13 or perhaps  “acquiring” an unusual belief.9  This last aspect is where conspiracy theories fit perfectly so that people can demonstrate their uniqueness by seeing these beliefs as unique and original possessions.9  What could be more “unique” than thinking that we are lab rats for the Bill and Melinda Gates Foundation or that 5G antennas spread the current virus?

Research carried out by Lantian et al.9 argues that people who have a greater need for uniqueness are more likely to believe in such theories, which, by their own nature, have the characteristics of being unconventional and potentially scarce information. The secret plots in which conspiracy theories are developed make people feel special as it allows them to self-identify themselves as more informed people about what is happening in the world.

By relying so much on this information, you can get to the point where you believe yourself to be more knowledgeable than the real experts about the event itself.9 With this in mind, it is not surprising that theorists react in disbelief or mockery in response to official sources of information. In fact, conspiracy theories diminish confidence in government and scientific institutions.4, 6

“They” are the culprits

Conspiracy theories fulfill the same function as deities in ancient times. When Deities were blamed for unfortunate events and myths that explained these events spread It seems that conspiracy theories now fulfill that role, despite our scientific advancement.19 It also happens in politics, when  “incompetent rulers” are blamed for everything. However, blaming others, in this case, “villains with macabre plans” and we as “the victims”, can also easily lead to disengaging from any kind of responsibility.2  It is the perfect excuse: “I cannot infect someone with coronavirus by not wearing a mask; instead, they are the culprits for creating a biological weapon that is killing us.” Unfortunately, this is how it often falls when it comes to a lack of self-care to fight COVID-19.

In the current context, people begin to reject medical treatments or resist a future COVID-19 vaccine based on unfounded ideas (e.g. “the vaccine seeks to implant a chip”). Instead, they start to come up with alternative treatments that are ineffective or prohibited, and that could lead to lethal consequences (e.g. death from ingestion of chlorine dioxide).

Vulnerability, loneliness, and lack of power

One of the social reasons that Douglas suggests are behind conspiracy theories refers to groups in vulnerable conditions with an objectively low status: people in a state of poverty or belonging to a “losing” group.2,3 His studies suggest that these groups may take to conspiracy theories as a defensive response to displace other reasons for the disadvantaged state in which they find themselves.

Swami & Coles14 also found that people who believe in conspiracy theories are likely to have greater feelings of helplessness, social isolation, and anomie —subjective deviation from social norms. It is also observed that people who perceive themselves as misaligned have alternative explanations when rejecting official sources and seek to satisfy their need for belonging by going to conspiratorial groups or marginalized subcultures.12 This is how people who feel powerless with their reality and ignore the norms of society as unfair can quickly dive into conspiracy theories.12


The AI Governance Challenge

Biases and more biases

We prefer dispositional explanations a thousand times over situational ones; this is based on the natural tendency to think that behind the occurrence of an event there must be a reason: “accidents do not exist”.

Studies show that people who are more likely to believe in conspiracy theories tend to overestimate concurrent events, assigning them to an intentionality that does not exist or is unlikely, which is called the fundamental attribution error.2 We thus observe that conspiracy theories are guided more by people’s intentions rather than by an objective truth that one wants to discover, which is certainly also linked to lower levels of analytical thinking.15

The conjunction fallacy is also present when a concurrent event with several conditions involved is considered more credible than a general one. We consider it more representative, even if it does not make sense.16 This is relevant for conspiracy theories because the unlimited details of conspiracies make them seem more credible. From a conspiracy theorists point of view, Bill Gates is a billionaire, something that gives him great power, he “predicted” that there would be a pandemic in the future, the virus was born in China, Bill Gates has a laboratory in China, so Bill Gates must have created the coronavirus. Almost logical, isn’t it? But, misleading of course.

Within all this blindness, the relationship between the confirmation bias and conspiracy theories has also come to light, where conspiracy theorists unhesitatingly accept information that confirms their preconceptions and resist accepting information that refutes their ideas.2 This fact builds a giant wall to break down in the fight against conspiracy theories. People turn to pages that only reinforce their point of view and, in the end, make the conspiracy theory cycle hard to break.

What can we do?

Based on the study of the psychology of conspiracy theories, Lewandosky & Cook11 offer us practical strategies to face the whirlwind of these theories that are currently fueled by the uncertainty of the pandemic.

One of them is inoculation or precombustion. Just as a vaccine is inoculated to release a small amount of a virus and build immunity to it, in the same way, before conspiracy theories fully engulf a person, they can be warned of their danger by developing greater resilience to them. An inoculation that allows awareness of the inconsistency of a conspiracy theory could be enough to destroy it. To directly discredit a conspiracy theory, you can also choose to disprove it by providing precise evidence that disproves it or by using the logic behind it by exposing the faulty reasoning on which it is based.

Considering that people with reduced control of their reality or who are perceived as vulnerable are more likely to believe in conspiracy theories, the opposite effect can be generated if they are “cognitively empowered”. For example, they can be encouraged to think analytically instead of being carried away by their intuition or by motivating them to recall moments in their life where they have had control in such a way that a sense of control is generated. It is also possible to empower citizens by making them perceive that principles of procedural justice have been followed regarding decisions made by the government. Even if there are unfavorable results, people will be more accepting of them, rather than dropping onto a search for global villains.

However, for a conspiracist who does not go to external evidence and goes into an echo chamber of conspiracy theories, it is much more complicated. How to talk to them? The following matrix summarizes the strategies that can be followed:

Show empathyListen to them carefully and try to build a mutual understanding.
Affirm critical thinkingYou are dealing with people who consider themselves as critical thinkers, reaffirm that sentiment and take advantage of it! Guide them to analyze even more critically what they believe.
Avoid ridiculeRidiculing a conspiracy is almost a guarantee of total rejection. You do not intend to win an argument. Avoid it as much as possible. 
Use trusted messengersShowing ex-conspiracy members or respected characters refuting conspiracy theories will allow for a more positive evaluation and a more impactful and lasting mark of memory.


Disinformation and conspiracy theories can lead to troubling outcomes, hence, it is necessary to recognize them as a serious social problem. What better way is there to do so than understand their background, and work on effective strategies that any person or entity can put into practice.

One must be aware of the responsibility that comes with handling incorrect information, especially in social networks where it proliferates. There are people who share conspiracy theories, even if they do not actually believe in them, which is worrisome.

Finally, we want to emphasize that this article is not intended to discredit all kinds of conspiracies, because unfortunately there have been real conspiracies (e.g. the Watergate scandal), nor is anyone denying that there are people who can weave large-scale deceptions. The search for the truth is fair and necessary, but not with the wrong tools, full of imprecise thoughts and inconsistencies. Rather than looking for patterns where there are none, value healthy skepticism, evidence, and consistency,11 which will leave the doors open to find what is really out there.

Evidence And Values In Policy And Research

Brooke Struck, our research director, sits down with Nathan Collett, to discuss the nebulous intersection between evidence, facts and policy. We talk through:

  • The complexities of selecting research methods
  • The rise and fall of data science
  • The reasons why technocracy is not the solution to all our problems
  • Challenges at the root of democracy
  • How to communicate between polarized sides of a debate
  • The current political scene


Nathan: Thanks for sitting down with me. Can we start by introducing the way that people usually think about good policy?

Brooke: I think it is very widely held that policy ought to be this extremely rational process of simply taking the evidence, weighing it, seeing which direction it points in, and then just doing that thing. One of the obvious shortcomings of that kind of ideal is that while evidence can be really valuable in pointing us in the direction of how to achieve what we want, it doesn’t tell us what we ought to want.

Brooke: So the outcomes that we set for ourselves are not evidence-based or evidence-driven or anything like that, and they’re not supposed to be. The evidence only comes in once you have a motive. The first layer of kind of problematization around that ideal is that once you have an objective set, all you need to do is look at the evidence and it’ll tell you how best to get there. That first step of setting an objective, it’s not an evidence-relevant activity. Evidence can be used to identify instrumental means to reach a goal, but goal selection itself is an inherently normative, values-driven thing that the evidence just doesn’t drive.

Nathan: Right. And you can probably bring evidence in, to kind of shape the goal selection. If you’re looking at quality of life, there’s certain pieces of evidence or research that can kind of inform that evaluation, right?

Brooke: And this is where we start to get into a second layer of critique of that very, very kind of stringent, hyper-simplified evidence-based ideal of policy-making. That second thing is that even as we shape our values and our preferences, we keep in mind how effective things can be. Often, the narratives, not just between people, but even in the way that we conceptualize the outcomes that we want, are often strongly informed by the indicators we use.

Brooke: So for instance, when we talk about quality of life, often, the way that we think about quality of life will be strongly informed by the way that we assess it. I think everybody is pretty much on board with the idea that quality of life is something that should be promoted. What they don’t agree on is what makes a high-quality life. Even something as simple as quality versus quantity. Is an additional year at 10% less quality inherently more desirable than one less year, but all of the years between now and then being 10% higher quality? I think that there are big disagreements about that.

Nathan: How do you find answers to those problems without using evidence and research?

Brooke: That’s the thing. I don’t think that the harsh distinction between facts and values is one that we should get on board with. We should embrace this more complex kind of interactive relationship between facts and values, where even the way that we conceptualize our values will be informed by the types of facts, or evidence, that we are creating. And when I say creating, I don’t mean fabricating data. What I mean is we must choose a measurement protocol in order to create data. And in making those methodological choices, we’re doing just that. We have an active role to play in how the evidence is created.

Nathan: That reminds me of what I have read by Jürgen Habermas

Brooke: Oh, of course. Critical theory is all about this, right? There are some really interesting things going on right now in critical data theory and feminist data theory about how the datafication of the world is not kind of a neutral medium through which we view the world of experience. But actually, these media themselves actually have a perspective. They are a specific prism through which we view the world in which we live.

Nathan: I wonder, and TDL just published a recent article that mentions this sort of neoliberal idea that companies would just be better off if they could kind of cut through all their biases and hire the best people. Do you think that’s kind of an oversimplification where you’re not going to recognize the best people precisely because of the kind of structure in which we’re making those assessments about who’s valuable and who’s not?

Brooke: Yeah. I think that the oversimplification in that instance comes from the term “best”. Along which dimensions are certain people the best? If there is a very clear and unproblematic way to define that, then I agree that we can probably get this argument off the ground that all we need to do is de-bias the process and then we’re golden. But actually trying to define who the best candidate will be is an extremely difficult and fraught process. And in fact, I think that some of the most interesting things happen specifically when we get into productive arguments and productive disagreements about what “being the best” means in terms of hiring, in terms of fit, these kinds of things.

Nathan: Do you think evidence has to do with how we determine the best or is that fully something that’s normative and the evidence is selecting someone once we’ve decided what our goals are? How does an interaction between evidence and values play out in that kind of specific context?

Brooke: Here I think is a good opportunity to kind of pull open that complexity, that interactivity between norms and evidence. We might say, “Okay, well, I want to define ‘best’ along five dimensions, A, B, C, D, and E.” I can only do that by drawing from this kind of lexicon of measurable stuff that is out there. So, that lexicon, that armory of tools, which we can go and pick up in building our normative definition, is where the evidence is that informs values, that values are intimately connected to our ways of building evidence. Maybe it’s not evidence that builds up our normative values. It’s methodologies.

Nathan: So it’s the way that we collected our data that matters. 

Brooke: That’s right. So, methodologies are maybe a place that we should focus more on this discussion, that methodologies are intimately connected to both values because we need some way to concretize our values and to our evidence because we need some way to collect our evidence. In the absence of methodology, we will really struggle to define our values or to concretize them, and we’ll be completely at a loss for how to collect the information to try to identify who best fits these normative descriptions, like who’s the best candidate for this job position.

Nathan: So, let’s go one level deeper. Where do we determine our methodologies from? I mean, I know you said values. But in a concrete sense, if you’re going to conduct a survey, a lot of this comes from past experience, right? To use the language of your paper, what are some incision points, or touchpoints, where one can actually intervene and change the process?

Brooke: One of the most valuable interventions in terms of identifying what evidence will be relevant for a problem, helping them to collect that evidence, helping them to process it both in kind of a very technical data science type of way but also in this much softer, decision-making institutional process kind of way as well. One of the things that’s most important in that work is keeping visibility on this whole kind of cascade or flow from the type of outcome that you want, which guides the type of evidence you identify as being relevant, and in turn, influences the type of methodology that you select to go to collect evidence. Finally, the analysis of what you collect leads to the decision-making process that is informed by that analysis.

Brooke: For me, I think the incision point is really about keeping visibility on that whole pipeline. That’s something that in a lot of contexts breaks down. So the functions that I just talked about in that whole cascade, institutionally are often divided into very siloed ecosystems within an organization. The person who’s responsible for creating the data probably doesn’t have great visibility on the institutional process that is going to be informed by that data later on down the line. I think in some ways data scientists have been in a privileged position to do that. Specifically because for a number of years data scientists kind of escaped or eluded very concrete descriptions of what the role was. They were more or less labeled as these unicorns or wizards who just did everything that touched on data within an institution or an organization. 

Brooke: And in that respect, because they eluded that kind of description, because they eluded that pigeonholing, they were also allowed to have the freedom to walk all over the boundaries that constrained most people in an organization. In having a role that was allowed to just meander liberally across all of these boundaries, we created that kind of interconnection between the silos that allowed better decision making to take place. In that respect, good data scientists were good data scientists and really enabled organizations to make better decisions not just because of their technical mastery but also because of the unique role that institutions allowed them to occupy that normally they don’t allow anyone else to occupy, just like a transversal cut across the organization.

Nathan: Just kind of going against that sort of traditional Adam Smith sort of specification of labor thing.

Brooke: Exactly.

Nathan: Do you think the success of data scientists is a critique of that idea of specification of labor? Do you think that’s something that needs to be revised just when we look at organizational behavior and you have this CEO that’s really disconnected from the kind of nitty-gritty data collection sort of thing. Is that a problem given this sort of framework?

Brooke: Yeah, I think that it is. I think that it becomes a problem when our methodologies start to evolve very rapidly. Basically, as long as your methodologies are evolving very, very slowly, you don’t run into these challenges where a CEO, for instance, ends up with some kind of data report on their desk which might just be a one-page executive summary of insights. You don’t end up in a situation where a CEO can have that kind of product arrive on their desk and the entire process leading up to that product will be opaque.

Brooke: If the methodologies are the same ones we’ve been using for 50 years, the specialization of labor becomes less problematic. Because when the CEO ends up in their role, there may not be that much difference between the way that the process was happening when they had their hands kind of deep in the muck and the way that the process is happening now that they’re more hands-off in that kind of day-to-day execution.

Brooke: As methodologies start to evolve more rapidly, that type of system breaks down because the people who have taken this step back and are taking a wider strategic lens on what’s going on cease to have a good transparent visibility on what the day-to-day operation looks like at the ground level.

Nathan: One of my colleagues in computer science talked about this in the context of his field. Basically, if you don’t move up from being a coder within 5 or 10 years after getting your degree, your skills are obsolete, so you have to move into management before you kind of just stop being useful to the company. I think that’s a total example of the disconnect between the people running the show and the people on the ground using new tools that superiors don’t know how to use. 


The AI Governance Challenge

Brooke: Yeah, and this raises the point that in some ways is mind-blowing and cool, but in another way kind of seems bent out, which is that the size of an organization can be adaptive or maladaptive to a certain circumstance. Larger organizations, because they tend to ossify and silo are better adapted to situations of slow evolution because the siloization is less of a problem when the evolution is slow.

Brooke: That’s why smaller organizations tend to be better at innovation because the siloization hasn’t set in yet, and so you can maintain that kind of visibility that is very adaptive for situations of rapid evolution.

Nathan: Earlier, you were talking about data scientists and their unique opportunity to have had their hands in every kind of pocket of their organizations. You were talking in the past tense. Was that intentional? Why?

Brooke: My sense is that a bit of the sheen has come off of data science now. People have seen it enough to start to understand a bit more concretely what the skillset is, what these people are capable of and this kind of thing. So, you start to see job descriptions that are more detailed and more specified as people start to understand what the role can and cannot entail. There’s something of this sheen or mystique of the novel that is lost once you start to understand how the thing works. The thing that makes magic magical is that you don’t understand. For anyone out there who enjoys magic and sleights of hand and this kind of thing, never, never, never ask how it’s done because it’ll stop being fun.

Nathan: That’s a really interesting segue. One of the topics that you have written about has to do with myth and the kind of role that myth plays both in ancient religion and also in modern political circumstances. I want to make a shift towards a discussion of rhetoric and policy and the kind of role that evidence plays in building policy. Make that link for us.

Brooke: Going back to the complexity that we were talking about before, the complex relationship between evidence and values, which passes through this conduit of methodology, the rhetoric that we see right now around the push for evidence-based decision making suggests that decision making currently is not or is insufficiently driven by evidence, which is immediately politicized. Decisions right now are not made based on no evidence. They are made based on evidence that certain groups of people critique because of the methodology through which it is gathered. Qualitative evidence is still evidence. As a politician, going out and speaking to the members of your community that you represent, going out and speaking to your constituents and hearing their stories, that is evidence gathering.

Brooke: Essentially, when we say we want decisions to be based on evidence, what we’re implicitly claiming is that going out and talking to constituents is not evidence, that that’s not legitimate, that it’s not important. So what we’re doing is devaluing certain kinds of evidence in favor of others. We need to be aware of that. We need to be aware of the fact that when we try to make these hard cuts between this is evidence and this is not, what we’re doing is we’re making a political statement about what kinds of methodologies and whose types of perspectives are allowed to count.

Nathan: What are some examples where that’s been effective lately?

Brooke: One of my favorite examples, it’s not a particularly recent one, but it’s one that for me is very evocative. The US Supreme Court made a ruling at one point about abortion saying that they needed to take an evidence-based approach. The challenge there is that there are broadly two perspectives that are in tension with each other when it comes to discussions about abortion. One is about health and well-being and the other is about the sanctity of life. More or less, the health and well-being question is the one that we use science to get traction on. The sanctity of life question is just not a scientific question.

Nathan: How so?

Brooke: Whether it is right or whether it is wrong to end a life is not a scientific question. There’s no experiment that you could go out and run that would gather evidence, analyze it, and come to the conclusion that it is wrong to do X or Y because that’s just not the kind of conclusion that empirical evidence leads to.

Brooke: What we end up with in this Supreme Court situation is the discourse about needing to take an evidence-based approach, which inherently favors the side of the discussion that is looking at evidence because their concern is an empirical concern. Whereas the other side of the discussion where the concern is more about religion, about values, about our sense of identity, the empirical elements of that are not really in question. I mean, there are some empirical discussions that happen at the margins there, around trying to identify as of which stage…

Nathan: A heartbeat and what not…

Brooke: Exactly. Trying to lay down some empirical moorings for things like when does life begin. But really, those questions, in my interpretation of the argument, are not the foreground. They’re not really what’s at the heart of the position of people who are advocating against abortion.

Nathan: It strikes me as kind of a post hoc way of coming at it. You have your opinions and you have your norms that are set through religion, and culture, and tradition, and whatnot. Then, you’re kind of stepping up to play at that technocratic, evidence level of the discussion.

Brooke: Yeah, there is something that goes on there. And I mean, not all of that is disingenuous, right? I mean, I think that there are people who are genuinely interested in this question of at what moment does life begin because they feel that those explanations help them to kind of deepen their understanding of their own values, their own views on the sanctity of life. I don’t think that that has to be a just disingenuous pursuit. But, certainly, the way that those kinds of discourses are framed, like this Supreme Court decision, explicitly making reference to wanting to take an evidence-based process, that’s kind of the covert way that certain political perspectives get smuggled into discussions while trying to do it in this apparently even-handed way that says, “We just want to look at the evidence?”

Nathan: So then how do we look at something like public opinion as something that’s been devalued in terms of a form of evidence. We have an election coming up in the states in a couple months. Is that something that should be reintegrated? How do you choose what evidence to take up in a climate like this?

Brooke: I’d like to push the question one step further and say we should be considering all kinds of evidence. The difficult, and interesting, and fun stuff happens when we ask, “How do these different lines of evidence fit together?” It’s not about your evidence versus my evidence. That, I think, leads us into all kinds of problems. It’s certainly not the type of approach that’s going to help us out of the current situation of polarization that we find ourselves in. But, trying to understand how facts and values fit together, even as I openly criticize that distinction, how these different lines of evidence fit together is the really problematic and difficult question. The opposition of “do I use this evidence” or “do I not,” is kind of a symptom of not knowing how to fit them together. That if I don’t know how to fit them together, the easiest response is to say, “Well, I choose one and I put the other aside” because I know how to work with one thing. If two things are incommensurable, then I just choose the one thing I’m going to work with and I don’t go through the hard work of figuring out how they fit together.

Nathan: Would you disagree with the statement that kind of all evidence is political or at least all evidence comes from somewhere?

Brooke: I definitely support the view that all evidence comes from somewhere. Somewhere, methodological decisions need to be made about how we collect evidence and how we process it to reach a conclusion.

Nathan: It’s a pretty serious challenge to an objective world view.

Brooke: Absolutely.

Nathan: How do we come together with people that are selecting evidence in a very different way from us? If objectivity is something that is constructed, where is the common ground? And maybe that’s a big question about how you can go from facts to norms and vice versa as wel actuallyl. But in terms of good governance or a good society, I mean, how do you get there?

Brooke: When I hear you ask that question, if I kind of distill the essence of that, abstract away from some of the concrete details, what you’re asking me is how do we find the right process or what is the right process to use. In such situations, my approach is always to kind of throw the boom over to the other side and tack into the wind and say, “Well, what’s the problem that we’re looking to solve?”

Brooke: I’m not sure if we agree on what the problem is that we’re looking to solve. I think that some people really do look at this situation and say, “The problem is that we’re not paying attention. We’re not paying enough attention to science.” I think a lot of the people who hold that view and a lot of the people who are of that position are themselves scientists who are rolling out this extremely self-serving argument that, essentially, they want to have a more important seat at the table, whether they recognize it or not. And I think it’s hilarious that often those scientists are the same ones who say that science is completely apolitical and that they don’t want to be politically involved. But at the same time, they are covertly and potentially even unawaredly advocating for their own political power.

Nathan: So there is another question here about the role of expertise. So, sure, science is never apolitical and we make all these choices prior to kind of delivering our kind of objective conclusions that are constructed through the scientific process. But, what about the difference between something that has been conducted in a randomly controlled trial versus something that is a hoax? Or is a myth, is something that has been kind of culturally assimilated in a way that’s maybe not as strict.

Brooke: Let’s start out by identifying that what you’re putting forward here is the circle that we have been looking to square since at least the time of Aristotle, almost certainly earlier. This is one of the fundamental tensions at the heart of democracy. How do we balance these two incommensurable measures of the quality of all people and the expertise of some, which is inherently non-evil?”

Brooke: I’m going to bring that up as a little precursor to whatever answer I’m going to offer just to manage expectations that there isn’t a super satisfying answer to this. But also, I’m going to defend myself that there are some pretty smart people who have come before me who have also not had super satisfying answers to this. I think my unsatisfying answer will go along the following lines. We probably should never get too comfortable with whatever balance we strike between the quality and expertise, that there are important elements from both of those kinds of anchor points that we need to conserve, and we’ll need them in different measures in different times.

Brooke: The most important thing that we can do is remain vigilant when we feel that we’ve reached a situation of imbalance. I think that the shift, that the rise of populism in Europe, in North America, and elsewhere, is a very clear indicator to me that we have lost a comfortable equilibrium. Populism arises when large swaths of the, let’s say, normal society feel that they have been left out of their own democracy, that they don’t have a voice, that they don’t have opportunities to build a good life, that the interests that drive the political discourse, and political decision making and economic decision making do not protect their interests, do not respect their importance. That is the space in which populism arises or the space in which populism can arise where the populace stands up and says, “I’m one of you, the real people. And we together need to take back the economy, take back the political scene, take back whatever center of power that we feel has been ripped away from us.”

Nathan: This is really interesting. While you were talking about that, I was thinking about two different ways that you could get that marginalized group. You can have one that’s actually been oppressed within their society and has no voice to say, “No, we are here as well.” Then, there’s the “let’s take it back.” There’s the “what we once had.” How much does it matter which group you’re appealing to? Does it matter that people feel a sense of entitlement or just want equality?

Brooke: Let me draw a distinction between just kind of entitlement in a material sense and entitlement in a more political psychological sense. Even in a population where the economy is struggling, where people are in a situation of subsistence living, people can still feel a sense of inclusion. They can feel that their political cast is kind of doing as good a job as is possible under the circumstances and that they really do have the best interest of the people in mind.

Brooke: I think that’s a good illustration of the division that I’m trying to draw here between material entitlement and kind of symbolic entitlement that the people who are scratching out this subsistence living may be totally comfortable with very little feeling of material entitlement. They don’t say, “I deserve better. I should have more. I should be getting paid more money. I should have a bigger house. It shouldn’t be so difficult for me to ensure that there’s a square meal in front of every one of my family members three times a day.”

Brooke: Even in circumstances of material challenge, we can still feel this kind of symbolic Identitarian entitlement that I have a right from my perspective to be heard for my rights, to be respected, for my dignity to be recognized in governance of the place where I live. And that, in many instances, can be enough.

Brooke: The flip side can also happen. When I say it can also happen, what I mean is it is also happening, that there are large swaths of society in North America and in Europe who actually are having their material entitlements quite, quite met. And it’s the symbolic breakdown that is driving them towards these feelings of disenfranchisement. And that’s something that I think we didn’t really have our eye on enough during the 20th century, that a lot of the way that we’ve told narratives about the historical arch from the First World War to the Second World War, a lot of that has focused on the material pressures that were exerted on Germany through the Treaty of Versailles and this completely unsustainable economic situation being driven by the reparation payments that were demanded by that treaty and this kind of thing and that that material breakdown is what led to the possibility for Nazism to rise. 

Brooke: The lesson that we’ve taken away from that is we need to assure a kind of minimum material comfort for people, otherwise they will take these kinds of extremist positions. What we’re learning now is that actually just maintaining that kind of minimum material comfort is not enough. There needs to be a minimum level of kind of recognition, a sense of dignity, a sense of my voice being heard, my perspective being validated and included that also needs to be protected. Because if there’s a breakdown on that side of the ledger, it also opens up a space for people to take very seriously some quite extreme views.

Nathan: I think this is a pretty big realization for people pursuing social justice. But, I think we have a problem where if the most disadvantaged members of society are in that kind of cast where their material entitlement maybe isn’t met but that’s not necessarily a problem, it’s going to be a lot harder to advocate for their advancement compared the conservative argument of, no, things should go back to how they were because my psychological needs aren’t being met anymore. And so that it’s so much harder to drive progress with people that don’t expect more than they already have right now compared to kind of clawing back at something that feels like it once was.

Brooke: The article that I wrote that kind of sparked this whole conversation really picks up from an interesting piece of research that was conducted in the last couple of years looking at the determinants of who shifted their vote from Democrat in 2012 to Republican in 2016. One of the major narratives that came out of the 2016 election was this narrative around people being economically left behind. The Rust Belt, all of these jobs have gone away. They’ve either been automated or they’ve been shipped overseas. And people are feeling that the economy has left them behind and that they can no longer provide for their families and these kinds of things.

Brooke: That piece of research that I built my article on in some important ways actually demonstrated something quite different, that it wasn’t a sense of being left behind economically that seems to be the strong predictor of shifting from Democrat in 2012 to Republican in 2016. Rather, the strongest drivers that they found were a sense of identity being left behind, namely a worry about the role of the United States on the world stage, that people were concerned that-

Nathan: Which I might add is a, I don’t like using the word psychological, but can only be a psychological assessment. Most voters just aren’t foreign policy experts, right?

Brooke: Right.

Nathan: It’s a feeling that’s been represented and that’s been transmitted to them somehow. So, that’s an excellent example of where all data comes from somewhere. And in this case, it’s really coming from somewhere unknown.

Brooke: So, I’m going to push back against this idea that it’s a psychological construct, that it’s a sense of identity. It’s a narrative about who we are. It’s as much about figuring out what evidence is relevant as about the interpretation of that evidence. There are some very real metrics according to which the United States is no longer leading the world. Economic growth is clearly one of them. The meteoric rise of China in the last 15 years, let’s say, that’s just something that the United States has not kept up with. And I don’t feel that that’s a problematic or contentious statement to make.

Nathan: No, it’s right.

Brooke: Except that I’ve kind of smuggled in which indicator I take to be the important one, right?

Nathan: Yes.

Brooke: Not that economic growth actually is for me personally the important indicator, but I think it’s one that has been bandied about as kind of like, “Well, look at this way in which America is stumbling. We’re not leading the world anymore.” Those kinds of narratives, those kinds of data narratives, are not hard to find. If you’re looking for a line of evidence according to which some country, be it the United States or any other is or is not leading the world, you can go out and find one. You just need to go and be as unscrupulous as you need to be in order to go find an indicator, but you can find it.

Brooke: The thing that I think is more challenging is that this idea that America will be number one and that that’s kind of its manifest destiny is one that hasn’t been kind of critically examined very much.

Brooke: Along which dimensions do we expect America to be first in the world? And if the answer is all of them, I’m going to push back pretty hard and say, “But what about the dimensions that we think are bad ones?” What about deforestation? Do we want America to be number one in deforestation in the world? Probably not. So the fact that I can pull out these examples that are trivial and obviously ones that we think are kind of bad examples means that there is a line somewhere. And I hope that that highlights the value of getting into discussions of where we think that line should be. There are going to be cases that are obviously ones where any country wants to be leading and ones where any country doesn’t want to be leading. All the interesting stuff happens in the gray area where we decide whether something is a priority for us or not.

Brooke: Economic growth is something that is so widely used internationally that it goes down as smooth as a spoonful of sugar. But is that something that we as a society feel is unproblematic? If we had to trade it off against something else, would we? Would we make a trade off for quality of life, acknowledging that it is a fraught thing to measure? Would we choose economic growth over quality of life? I’ve been nothing short of astounded to hear some voices in the United States saying, “If COVID’s coming to get me and I’m old, I would rather die than to have the economy shut down to protect the health of the population of this country. That to me is a pretty shocking thing.

Nathan: It is interesting to me that those arguments for “I would die before I let our economy stumble,” often when pressed, they fall back on, “Oh, well, if the economy stumbles, more people are going to suffer. And ultimately, we’re going to be better off if we don’t shut down.” I think we touched earlier on the kind of idea of, during the abortion discussion, how evidence is brought in ad hoc to kind of prop up something that’s not a place where evidence has a role. That strikes me as one of these two where the value judgment’s been made. They’ve made that prioritization, and they put quality of life down, and now they’re sort of kind of coming back to it. 

Brooke: We need to be more kind of open, transparent, and critical about what evidence we actually care about. We have such a habit of just throwing out statistics, it’s my statistic against yours, that what we never do is sit down and have a hard discussion about, “Why do I care about this statistic compared to another?” The fact that we don’t have those discussions often leads us to these circumstances where we have to choose between listening to somebody and not listening to them based on whether their preferred statistic either agrees with or disagrees with my preferred statistic, whether we’re pointing in the direction of the same action that we’re advocating for.

Brooke: If we do the hard work, dig into the nitty-gritty of “why would I care about this statistic versus some other,” two things can happen. First of all, we can engage more meaningfully with people than we currently do, which is this very polarized binary way of either we agree or we don’t agree. There’s much more nuance once we get into the nitty-gritty of why we should care about various things.

Brooke: The other is that through that exploration in itself, we can actually come up with solutions that we hadn’t identified before because we hadn’t taken the time to kind of unpack it and figure out what we care about. Those kinds of explorations are really, really rich. There’s a lot of space for creativity in there to come up with not just a realization about which statistics we care about, and why, and how to fit them together, but also what solutions might be available to push us forward on the dimensions that we care about the most. 

Nathan: That’s a wonderful place to stop. Thank you for taking the time to discuss this with me. It has been really insightful and I look forward to the next time. 

The Ultimate Temptation: Dropping Out Of School

In a study by The Decision Lab, nearly half of all respondents disagreed that someone who graduated from high school would contribute as much to the economy as someone with a postgraduate degree. How can we position more students to graduate from high school and enjoy the associated economic, social, and cultural benefits of doing so?


Failing to finish secondary school isn’t without consequences: $630,000 lost in potential earnings, 90% job ineligibility, elevated rates of heart disease and diabetes, increased risk of mental illness, welfare reliance, and incarceration, to name a few.1 These aren’t limited to a small minority. In the United States, for example, 1 in 8 students never graduate from high school—a statistic that has remained largely unchanged since 1990.2 In Quebec, one in five girls and one in three boys leave high school before graduation.3 These findings are similar (and at times worse) in developing countries. In Jharkhand, India, a state with the highest school dropout rate in the country, 70% of students leave school at the primary or secondary level.4

This unfortunate combination of a large number of school dropouts and the immense consequences these individuals face calls for concerted action. Accordingly, this article explores a number of key questions: What is the demographic profile of students who drop out? What are the explanations—psychological and otherwise—for them doing so? Have interventions been devised to address this problem? Given the immense scope for improvement, what other solutions could be proposed?  

The Statistical Story  

The statistical story can be examined in two segments: (1) who drops out, (2) and why do they drop out. To frame this discussion, we focus particularly on North America, citing examples from the US and Canada.  

Who drops out? 

The propensity to drop out of school is accentuated by low socioeconomic and ethnic minority status, conditions that often overlap. Out of the 1.3 million students who left school in 2013 in the US, more than half were students of color, and most were low-income. In fact, “low-income students fail to graduate at five times the rate of middle-income families and six times that of higher-income youth”.5

A few explanations have been offered for this disparity.  One is that familial poverty coincides with several adverse conditions or stressors—food insecurity and malnutrition, absent or incarcerated parents, homelessness or poor living conditions, and domestic violence.6 These stressors detrimentally affect students’ capacity for learning, often contributing to poor academic performance, anti-educational attitudes, and low scholastic effort, which are all precursors to dropping out.7

Another explanation is that of geographical distribution and school funding inequalities. Ethnic minorities often congregate in certain areas, areas that are often marred by poverty and crime. For example, a 2000 report entitled, High School Dropout, Race-Ethnicity, and Social Background found that African-American high school students mostly reside in the South, Hispanic students in the West, and white students are distributed throughout.8 Given the high levels of poverty and crime in these regions of the South and the West, and the reliance on district education funding, there can be severe constraints on the resources and opportunities at the schools where ethnic minorities are mostly found. Contributing to a poor educational experience, these constraints can place students at risk of dropping out. 

Why do they drop out? 

The “push, pull, and falling out” framework dictates three conditions that can contribute to a student dropping out of school. 

The first, push, is when the student experiences a push from the school that results in their decision to drop out.9 The student could have not followed school rules or policies, for example, cheating on a test or being disrespectful towards a teacher. Doing so and the subsequent consequences—detention, suspension, expulsion—could result in a student dropping out.

The second, pull, is when the student is pulled by factors within themselves or their own lives to drop out.10 This might include pregnancy, changing interests and opportunities, financial concerns, illness, or familial factors. 

The final is falling out, whereby a student gradually becomes disengaged and disaffected with the idea of completing school.11 As opposed to push and pull factors that have identifiable causes, there may be no particular reason in falling out cases.  

Psychological Interventions To Prevent Drop Out  

Most school-dropout-prevention interventions focus on providing either psychological or academic support to students. The former might include counseling and increasing teacher support, while the latter might include providing specialized tutoring or altering expectations about student performance.

One intervention is aptly called Check & Connect. Applied throughout Kindergarten to Grade 12, this intervention is largely psychological—providing increased teacher support to at-risk students and their families.12 The researchers use a concept called “persistence plus” whereby they demonstrate to students that there is an adult at their school who believes in them, is available to them, and will motivate them to learn, complete homework, maintain regular attendance, and ultimately, stay in and succeed at school.13


The AI Governance Challenge

Having been applied for over two decades, the intervention has been successful in increasing students’ academic engagement. Why? “Student engagement across the school years depends on the degree to which there is a match between the student’s characteristics and the school environment so that the student is able to handle the academic and behavioral demands of school”.14 By creating close bonds between students and teachers, the intervention assists in cultivating a match between the student and the school, and as challenges arise, effectively addressing these. 

Similarly, academic support interventions have been successful. Benjamin Bloom, an American educational psychologist, considered the benefits of one common form of academic support: tutoring. He found that one-on-one teaching, in which tutors tailor their methodology for each student, is very beneficial. “According to Oakes & Lipton, 80% of Bloom’s experimental students achieved at a level reached by only 20% of the students in typical classrooms”.15 This finding is significant given that students who fail Math or English in 8th grade are 75% more likely to drop out of high school.16

However, a feasibility problem arises. Providing individualized tutoring for all students, or even just those at-risk of dropping out is not practical given budgetary/monetary limitations in the public school system.

Emulating this idea is the concept of motivational interviewing. In this intervention, which combines socioemotional and academic support, students regularly meet with a school staff member. In regular meetings, occurring every two weeks, they discuss strategies for increasing a student’s motivation to participate in and finish school.17 “Motivational interviewing is an empathetic, student-focused, collaborative, and directive behavior change strategy that works to increase students’ motivation”.18 Over time, the increase in students’ intrinsic motivation can reduce their risk of dropping out.

While these interventions can be effective, they are impractical to standardize and scale. So, what’s next? What could be some other options that offer better chances of system-wide improvements?  

Potential Solutions 

Potential solutions can be broadly split into two areas. The first is to extend current interventions, and the second is new solutions not yet in active practice.  

1. Extending current interventions

Researchers at Stanford University devised an intervention to reduce suspension rates among middle-school students.19 Previously cited as a push factor, suspension can contribute to a student’s decision to drop out of school. Consequently, reducing suspension rates and the associated misbehavior may discourage dropping out.  The intervention encouraged teachers to adopt an empathic as opposed to a punitive mindset when engaging with students who had misbehaved. To do so, teachers read an article, explained how empathy could be used to ensure classroom control, and subsequently noted how they would discipline students in three hypothetical scenarios. Conducted in five middle schools in the US, the intervention increased respect for teachers and good behavior among students, cutting student suspension rates in half: from 9.6% to 4.8%.19

Adopting this intervention, through administering widespread empathic response training, could be successful in reducing school dropout. For one, it reduces the push factors experienced by students. It also creates a sense of psychological safety. Their teacher—an authority figure—is responding to their misbehavior with understanding and offering opportunities for improvement. This can encourage positive attitudes about schooling and hence increase completion. 

2. Devising new ideas

Our previous exploration of successful interventions—individualized tutoring, Check & Connect, and motivational interviewing—revealed strategies that improve outcomes but are difficult to scale. Providing students with individualized guidance, giving them and their families school support, and engaging in regular check-ins, all prove key. However, the primary limitation is that this is resource-intensive. So, how can these strategies be delivered in an accessible manner?

One idea is that of group-based interventions. Rather than interacting with at-risk students on an individual basis, it could be done in a group format. Mirroring group therapy for depressive disorders, this would connect students with others who are experiencing similar issues and facilitate a shared avenue for resolving these issues. There is, of course, the risk that this creates comradery among at-risk students and enforces a mutual desire to leave school. 

Another idea is peer-to-peer counseling. Pairing students skeptical of school (and at risk of dropping out) with those with foremost educational goals could facilitate an important dialogue. Sharing perspectives across such disparate viewpoints would, at the very least, provide an indication that another attitude, experience, and outcome is possible. Doing this between students also facilitates free discussion—there is no obligation to feign understanding or withhold genuine questions.  

Looking forward

Graduating from high school is critical to achieving outcomes along key dimensions such as employment and earnings, physical and mental health, and incarceration. However, throughout the developed world—Scandinavian nations being the notable exception—high school dropout rates remain high. This inflicts an immense social, economic, and individual cost. Interventions have been effective in reducing school dropout rates in the local ecosystems where they have been implemented; however, they are limited in feasibility to apply at scale. Applying key lessons from these interventions in more accessible formats could create scalable and impactful interventions. Ultimately, our education system must adapt to the vicissitudes of students. It must be generous, inspiring, and enlightening.  

How To Motivate Volunteers With Behavioral Science

Julian Hazell, an associate at The Decision Lab, sat down with Jayden Rae, a senior consultant with expertise in non-profit coordination, to discuss some of the following topics: 

  • Why people like volunteering
  • Why behavioral insights might be necessary to understand why people volunteer
  • The benefits of altruism
  • Motivation for volunteers and challenges to the rational actor model of behavior
  • How we can make a difference in the topics that matter to us
  • Concerns surrounding voluntourism
  • How to convert passion into action


Julian: Thank you for taking the time to sit down and chat. To start, how do you define volunteering?

Jayden: Volunteering usually refers to an action that is intended to help others. In this context, I’m referring to formal volunteering arrangements. For instance, people who are volunteering for non-profit or charitable organizations, not other forms of unpaid work such as looking after a family member An individual performs this act without any expectation of monetary compensation or gain. 

Jayden: Volunteer activities necessarily involve some level of freewill – the volunteer actively wants to engage in the behaviour. You usually need to be in a privileged position to be able to volunteer, as you’re giving up time you could otherwise spend on paid work or leisure activities. Sometimes, people will be doing unpaid labor under the veneer of volunteering, yet that isn’t a product of free will. 

Julian: People sometimes have a perception that neoclassical economics says that we are only self interested and only care about maximizing our utility. Do you find that to be convincing?

Jayden: I think classical economic models can explain some things, but often they can’t explain  human behavior. As we now understand quite well with some teachings from behavioral science, individuals are not rational actors. 

Jayden: And there are many reasons why we may not act in a way that economists might predict. A big reason for this is that we’re a social species that forms complex communities. We never make decisions that are truly independent – they’re shaped and informed by our environment. 

Jayden: I think volunteering is a good example of acting in a way that economists may not predict – namely, that people respond to monetary incentives. A lot of the reasons why people might engage in unpaid labor are associated with the social role they play in their community. Also, intrinsic and altruistic motivations might provide utility to individuals in a way that economists wouldn’t necessarily predict (as those who volunteer are not being strictly monetarily compensated), so it can be tricky to measure using typical neoclassical models. 

Julian: So you think people can derive utility from altruism itself?

Jayden: Neoclassical economics “selfish” model of utility does not allow for the fact that most of us in society care about the well-being of others. However, you can imagine revised models where individuals derive value from prosocial behavior. In fact, there are things called “benevolent utility functions.”  I’ll get to some of the reasons why later, but there are many benefits that people receive as a result of volunteering that can incentivize them to engage in it. 

Julian: Okay, interesting. So many people in our current day and age feel really burned out from being overworked, or from academic obligations as well. What are some reasons that, in spite of our busy schedules, we should make more of an effort to volunteer?

Jayden: There are a ton of reasons why volunteering offers positive direct and other spillover effects to the person volunteering. The first main outcome of volunteering is increased happiness and life satisfaction, which can arise when you have the opportunity of aligning your actions to your values. Not everyone can find a paid job that fulfills their interests in contributing to a greater social good and playing a role in their community. For example, you may care a lot about the environment but work for an accounting firm. Beyond taking action in your individual life to minimize your carbon footprint, you could volunteer for a local conversation group. 

Jayden: Volunteering then can become an outlet to fulfill your social function in a way that will actually make you happier.This has been proven empirically in psychological studies that have demonstrated that the act of volunteering in itself can make you happier in the long run. For instance, a metastudy published by the University of Exeter found that volunteering had favourable effects on depression, life satisfaction, wellbeing. 


The AI Governance Challenge

Jayden: It has also been shown that even volunteering can actually make you physically healthier.  So of course the question is, are people who volunteer just generally happier and healthier, and this is a reason why they volunteer? One study from the University of Michigan involving over 3000 individuals controlled for self-selection effects and still found that over time volunteering could actually lead to healthier outcomes and decreased premature mortality rates. Mortality risk was reduced even more for each hour older adults volunteered per month.

Jayden: So I think there are a lot of reasons for volunteering based on those results. But of course, volunteering, as I mentioned earlier, is a privilege. It requires some level of free time. And people who are going to be filling volunteer roles must have the financial and physical freedom to be able to do so.

Julian: Okay. So what are some important behavioral motivators for getting people to actually volunteer?

Jayden: People often ask: Is volunteering simply a product of altruism? It’s important to emphasize that not all volunteering is altruistic and not all of altruism is volunteering. But altruism definitely plays a role, and there’s a really strong, positive association between altruism and volunteering. 

Jayden: One of the first reasons that behavioral science offers to explain why people volunteer is due to something called a values function. In simple terms, a value function means that individuals are drawn to participate in volunteer work to express and act on values that are important to them. 

Jayden: If we’re really driven towards a particular cause, such as working with people with disabilities, and it’s something that we feel is important and it makes us feel happy and fulfilled, we will want to  seek  out a role where we can do good with those populations. 

Jayden: The second reason people volunteer is due to something called an understanding function. This suggests that people will volunteer to increase their knowledge or skills in a particular area. For instance, someone might volunteer for an organic farm because they’re interested in developing skills around organic farming or gardening.

Jayden: The third reason is called an engagement function, essentially, we seek out activities which connect us with others. This makes sense because volunteer opportunities are typically quite social in nature. You get to meet people who you wouldn’t otherwise interact with, and that can enhance your own psychological development. 

Jayden: The fourth reason is career related. There’s a social norm around volunteering, such that it is something that is a good thing to do. A lot of employers and schools actually look for volunteer experience.

Jayden: I think those are some of the main reasons why people might volunteer. Of course, there may be other ones as well.

Julian: So if I’ve found something that I’m passionate about and I want to make a difference in, how can I be sure that volunteering is actually the best way of doing so? What about just donating money?

Jayden: Yeah, I think that’s quite an interesting question. The mere prospect of “making a difference” at the individual level requires a lot of self reflection about what your best levers of change are. If you don’t have enough time to actually make a real commitment and contribute to an organization, it probably makes more sense to give money. But if you’re interested in realizing some of the benefits of volunteering, a study at the University of Exeter showed that people who actually volunteered their time were more likely to feel better about the action than people that gave money. It had longer term effects on indicators like wellbeing and self-efficacy.

Jayden: One thing that’s kind of interesting to observe is that usually it’s not really ‘efficient’ in an economic sense for someone to volunteer their time. So for example, if a consultant is making $100 an hour in their usual job and they dedicate an hour of their time to volunteer, it probably would have been more valuable for them to donate, say, $50 to the charity in terms of the actual opportunity cost. There should be some level of alignment around what  skills  you  can offer, what the organization needs, and the role you play as a volunteer.  And also, are you able to measure the impact of that? That is a key consideration to keep in mind. 

Julian: What behavioral science insights can organizations use to generate social change and encourage volunteership?

Jayden: This connects closely to the reasons I previously outlined about why people volunteer in the first place. Organizations should understand why their current volunteers are attracted to working with them so that they can optimize the benefits and advertise those to potential future volunteers. So for example, if people who volunteer are most interested in developing a specific skill, the best thing they can do is advertise opportunities around skill development to provide ample opportunities for volunteers to develop their own capacity and self-advocacy. To the extent volunteer organizations can promote prosocial opportunities for volunteers to connect with the community and one another, it can foster intrinsic motivation.

Julian: Let’s pivot a bit to talk about voluntourism. So voluntourism is the act of going abroad to volunteer in different countries, which are often low-income. Some people have criticized this quite extensively. I just wanted to get your opinion on this: How can someone be sure that they’re actually volunteering both effectively and ethically, and that their efforts are actually wanted by the people that they are helping?

Jayden: I’ll start off by being generous and suggest that many people end up volunteering due to their own naivité. Often, voluntourism trips are branded in a way that makes it appear as though there is a clear cause for which you can directly have a positive impact. So, I believe.  there’s often a lot of willful ignorance on the part of the participant. 

Jayden: To avoid situations where harm is done  as a result of any sort of volunteer,  I think individuals must  be critical and really deliberate about why they’re volunteering in the first place and where they’re going to be volunteering. I would suggest a couple of questions people should ask themselves before they engage in any volunteer activity, whether it’s voluntourism or otherwise. 

Jayden: First, look at the governance structure of the organization. Who is running the organization, who comprises the board of directors, how is the organization spending its money? Usually organizations will release a financial statement every year if they’re registered, and even taking a quick look at that gives you a useful glimpse into how they’re managing their income and resources.

Jayden: Second, deeply consider what causes or issues are driving you to volunteer in the first place, and whether the organization you’re considering volunteering with actually aligns to your values in a way that is authentic. 

Jayden: Third, consider what you have to offer. Do you have specific skills that will benefit the organization you’re looking to work with? If you’re going to an orphanage where you’re going to be reading a book for two hours, is that really the most important skill you can offer and is that something that this organization would need? And will you be taking the job of someone who is local, who would otherwise be earning income from what you’re completing through your volunteer involvement. 

Jayden: Fourth, look for organizations that are taking an evidence-based approach by measuring the impact of their actions over time. If you’re volunteering for an organization that has no mechanism to quantitatively measure their direct impact on the community, the spillover effects, or the negative outcomes of their work, that’s definitely a red flag and it’s something you should be looking out for early on.

Julian: Let’s say someone is really passionate about a specific issue or a specific cause but they don’t know who to reach out to, and they’re worried that the first organization they would find might not be super effective. What is a way that people can take an evidence based approach to finding out where they’re going to be the most effective in a cause that they care about?

Jayden: This is precisely why it’s so important to do your background research. Look at the financial statements of the organization, look at their past impact reports and see what types of activities they’re conducting in certain focus areas.  If you have the opportunity, it’s always a good idea to talk to someone within the organization in an informal way to inquire about the role they see volunteers playing in their organization, about the relationships they have with other organizations, and what they identify as the key barriers or challenges to fulfilling their mandate. Often, this informal interaction piece is crucial to determine how the organization is actually operating, and what frustrations you might experience as a volunteer. 

Julian: What would you say to those who feel that they are just a small fish in a big pond and who feel that they are passionate about a big problem, but they are not necessarily big enough or influential enough to solve it themselves? How would you address feelings of paralysis, when we feel overwhelmed at the size of the problem and have no idea where to begin?

Jayden: I worry that the small fish big pond sort of thinking can lead us down the path of empathy. Believing that the individual has no impact on system level outcomes can become an excuse to not act at all. 

Jayden: I would offer a different type of thinking. Perhaps, when an individual dedicates their talents, time, and energy, there could be a cascading effect when they merge forces with similarly passionate individuals. Real change tends to happen when a group of people come together. It’s about community-level mobilization, which necessitates the initial individual-level action. 

Julian: What are some final tips you have for our readers who are passionate about different areas and actually want to make an impact? What are some tangible things that they can take away from this and go out and do right away?

Jayden: There are a couple of key things. I think the first one is to do a self assessment of the types of skills that you can offer and find the ones that give you a lot of fulfillment and are also helpful for the organization you care about. So for example, if you love to write, research organizations that might benefit from having someone who works on the publication or communication side. If you’re trained as an accountant, you could do wonders for a small charity that is otherwise having a hard time managing it’s bookkeeping. Sustainable and impactful volunteer relationships should be reciprocal. 

Jayden: The second one is to consider the various levels of change available to you. It doesn’t necessarily have to be volunteering for multiple hours a week. It could be making an annual charitable donation for example (some apps like RoundUp make this so easy you’ll probably forget about it), or it might be volunteering for a political campaign. There are a lot of ways that you can contribute to civil society and engage in civic life that isn’t necessarily working at a soup kitchen. 

The Science Behind Curiosity

When I was little, maybe 7 or 8 years old, I remember reading a funny story. It was called ‘The White Elephant’. It goes like this:

Long long ago, in a faraway land, lived a hardworking gardener and his wife. One night, the gardener was walking back home from work, when he heard a rustle in the bushes. He hid and peered out from the bushes, and to his utter surprise, found a white elephant. The elephant silently grazed the grass for some time, and then took off flying towards the sky just before dawn.

Out of curiosity, the gardener ran and hung on to the elephant’s tail. A few moments later, he found himself in a strange garden. Everything in it was enormous — tomatoes, apples, cucumbers, you name it. The gardener gleefully started picking up some fruits and vegetables to take home. A few hours later, the white elephant reappeared and started going back to earth. Our man hung on to the tail and hurried home with the loot.

He told his wife about the adventure. She was unable to contain her excitement, so she told her friends about it. Somewhat in the spirit of fake news, the story spread and everyone demanded to go on this adventure. On a selected day, the whole village hid behind bushes. When the gardener ran to catch the elephant’s tail, the villagers followed him and formed a long chain, hanging on to each other’s legs.

It was all going well and the group was well on its way towards the garden until the last person could not contain their excitement any longer and asked the person above exactly how big a watermelon in this garden could be. The question was passed on and reached the gardener. He got annoyed and shouted, “they are THIS big”, opening up his hands wide. They all fell down.

I remember laughing a lot at this story. But unfortunately, as a behavioral scientist, I am now questioning the details. Why did the gardener have to hang on to the elephant’s tail? What made his story spread like fake news? Why was everyone dying to go on this adventure?

But it’s not just about this story. Why do we watch binge-watch cliffhangers? Why do we care if 2 random strangers ended up marrying each other on reality shows? Why does it matter if an iPhone X can be blended? (Spoiler alert: Yes – it can be blended into a fine powder)

To put it simply, we are all what Herbert Simon referred to in the title of his talk at Carnegie Mellon in 1992 – “The Cat that curiosity could not kill”.1

The curious case of curiosity

For many years, curiosity has baffled psychologists and philosophers alike. This is not surprising in itself, given that our curiosity goes as far back as the story of Eve taking a bite of the apple she was forbidden from touching and Pandora opening the box she was told not to open.

One of the earliest attempts at uncovering the underlying reasons behind curiosity was by William James in his 1890 publication, The Principles of Psychology.2 He identified two types of curiosity: one driven by emotions, and one driven by scientific wonder. Following this, several other attempts were made, and we now have a somewhat clearer understanding of what gets us curious and how we can sustain this.

Daniel Berlyne introduced us to the world of curiosity aroused by external stimuli with characteristics of novelty, uncertainty, and conflict. He also hypothesized on the level of stimulation — if the stimulation is too low, there will be no reason to explore, and if it is too high, it will result in anxiety.3 The right balance of stimulation is needed for exploration. This has been captured as the “zone of curiosity” by H.I. Day, a colleague of Berlyne.4


The AI Governance Challenge

Taking this hypothesis further, George Loewenstein synthesized several theories and produced what is now known as the “information gap theory”.5 This states that for curiosity to be aroused, a small amount of knowledge about the subject serves as a priming dose. The gap that emerges between what’s known and what’s unknown is what drives motivation for curiosity. Based on this, we arrive at a simple framework to think about curiosity and how we can use it in product design and marketing.

Getting the curiosity mix right

Imagine it this way. There are two cliffs, separated by a gap. So, three things need to happen for someone to jump from one cliff to another over the gap:

  • People need to reach the edge of the first cliff
  • People need to feel curious enough to want to know what’s on the other cliff
  • People need help to jump to the second cliff.

Now, apply this analogy to make people curious about a subject, make them want to know more, and then give them enough information to help them cover the gap they have about the subject.

Let’s look at each of these steps.

Step 1: Bringing people to the gap – Grabbing attention

Our brain is designed to notice when things change. The best way to get people’s attention is to break a pattern and bring about a change that gets noticed. Surprise is a reaction that is triggered when a pattern gets broken and our existing schemas of knowledge fail. Our brain demands attention to help understand why the failure happened so we can prepare for future events.6

So the right way to use surprise to get attention is to break a pattern, cause a schema to fail, and then fix the schema by giving information. If there is no fix, people lose interest. An interesting study that uses this methodology was done by Bruce Whittlesea and Lisa Williams. They tasked people to look at 4 words and react.7 These words were:





PHRAUG and TAYBLE look unusual but sound familiar. They break a schema, but people can immediately put two and two together and arrive at an explanation: frog and table. BARDLE and HENSION cause a frown because they look familiar but cannot be placed in any existing schema. They cannot be “solved”.

A similar theme is observed in advertising. A good ad is one that piques interest and then lets the audience figure a solution. As opposed to an advertisement that piques interest and does not fix the problem satisfactorily. Consider this famous advertisement by CANAL+.

The reveal of the bear at minute 0:21 is the surprise and this builds up till minute 1:07 when they show why the bear is playing the role of a director. If the reveal at 1:07 had been a non-consequential joke, such as a generic statement that said bears love movies as much as humans, this advertisement would have failed.

Step 2: Convert the abyss to a gap

So you have grabbed attention. Now, what can you do to keep people on the edge? The key here is to ensure the gap does not seem like an abyss to the person.

Loewenstein gives an interesting context to this with an analogy.5 Say person A knows the state capitals of 47 out of 50 states, while person B knows the state capitals of 17 out of 50 states. Who do you think would be more curious to close the gap and reach 50 quicker? The people who know 47 state capitals focus on what they don’t know. So, Person A has a curiosity to close that gap and know more, as opposed to person B, who is proud of their knowledge of 17 capitals.

This plays out in many forms around us. For instance, celebrity gossip is popular because we know a lot about being human and we understand a lot of emotions, but we don’t know the exact details of being Kim Kardashian. When she gives us a sneak peek into her life, she is filling those gaps.

So the key is to give some knowledge that forms a basis and then fill the gaps. Have you ever noticed that reality shows such as American Idol often show a background context about the participant before revealing their talent? It is generally quite dramatic with tears and backstories that seem straight out of movies. In the absence of that, the audience is viewing the participants as just singers. The background context fills up the knowledge bucket just enough to make them curious to know how these people would perform.

Step 3: Making sure the gap isn’t too small

Once a person has reached the edge and is waiting for the reveal, the size of the gap becomes important. If the gap is made too small with the knowledge we are providing, crossing the gap becomes a small step. In other words, a huge build-up followed by a small reveal will cause disappointment.

If you were playing a video game and the hints are step by step guides that keep telling you at every point exactly what you should do, you would lose interest in the game very fast.

The trick is to give just the right amount of information to keep people engaged. They should know enough to wonder about the next reveal and start forming theories in their mind. No one enjoys a mystery thriller book that makes the killer too obvious. The fun is in giving people clues and making them curious enough to solve the mystery while reading.

So there you go. Curiosity never killed the cat. It only made the cat more engaged with its environment, provided the gap was just the right amount. Now if you will excuse me, I must go find out if an Amazon Echo can be blended in a blending machine.

The Unconscious Process Behind Financial Instability


This article is part of a series on cutting edge research that has the potential to create positive social impact. While the research is inherently specific, we believe that the insights gleaned from each piece in this series are relevant to behavioral science practitioners in many different fields. At TDL, we are always looking for ways to translate science into impact. If you would like to chat with us about a potential collaboration, feel free to contact us.


It seems as though we live in a world of boundless financial speculation, from volatile stock markets to unpredictable crashes that damage our faith in assured progress and safety. Professor Selim Aren is a researcher in business administration who believes that unconscious drivers may be responsible for the madness that plays out in today’s financial markets. He has pioneered a niche area of study at the crossroads of experimental psychology and finance to determine some of the causes of this complex part of our lives. 

The Decision Lab is a social enterprise that aims to democratize behavioral science. We aspire to share this essential insight with a wide audience, with the hope of reaching the ears of critical decision-makers. With this goal in mind, we reached out to Selim to connect his important work with a broader audience. Too often, research does not naturally reach the people that need its luminance the most. This piece is part of a series that aims to bridge that gap.

Nathan: How would you describe the focus of your research in simple terms?

Selim: Throughout history, many financial bubbles have been observed, such as the tulip balloon and the South Sea Company. Although people think that these will not be repeated in modern times, the dot com and hedge fund financial bubbles of the early 2000s refute this. In our research, we investigate why people do not learn from these events. Our main idea was that unconscious processes can have an impact on people’s decisions. Therefore, “phantasy” based on human infancy was the focus of our study.

Nathan: What does phantasy mean?

Selim: Phantasy is different from fantasy. Fantasy is conscious. However, phantasy refers to unconscious dreams. Sometimes we develop our own phantasy without realizing it ourselves. Sometimes others develop our phantasy for us. Many people want to get rich quickly. They believe in the existence of financial investments with no risk but with very high returns. This belief exists not only in individual investors but also in financial professionals. They also believe, perhaps unconsciously, in such a dream. This belief often develops through narratives of genius investors that seemingly can do no wrong. We tried to understand and explain the factors that develop these unconscious dreams of success without risk.

Nathan: What did you expect to be behind this phenomenon?

Selim: The main purpose of this research was to find phantasy determinants. Just because phantasy is an unconscious process does not mean it cannot be manipulated. If people’s phantasy leads them to poor financial decisions, we can protect people from this mistake by manipulating phantasy. If phantasy was unconscious, there is likely another unconscious process driving it. With this in mind, we began with a literature review. The prominent concepts were states of mind, group feeling, and narratives. However, there were no empirical studies showing the relationship between these concepts and phantasy. This excited us very much. There were theoretical relationships, but they had not been empirically modeled and tested. We also knew that herd behavior is typical of financial bubble periods. At that time we thought that this too could be effective in phantasy development. This is how we designed our model, developed scale our variables, and tested them.

Nathan: How did you test your model?

Selim: The first thing we did was to read and understand the literature very well, because phantasy has been the subject of very limited study in the field of finance. However, this was not very different in the general literature. Studies were theoretical or interview-based, which was not suitable for generalizations. If we could develop a scale, we thought that a large number of people could be surveyed. For this, we developed scales for both phantasy and phantasy determinants. Because there was no scale for them either in the financial literature or in the general literature. Also, since our focus is on financial bubbles and human madness, we based our scale on financial concepts. Then we had to model the relationship of phantasy determinants to each other and to phantasy. This was also a difficult and first-time process. Ultimately, we achieved this and tested our model.

Nathan: What did you end up finding out?

Selim: This research exposed some phantasy determinants and their relationship to each other. We found that narratives, and divided mind and group feelings are strongly related to each other. The typical behavior of the financial bubble periods was herd behavior. However, this did not have a unique structure. There could be informed and uninformed herd behavior. It was the informed herd behavior that increased phantasy. There is even a moderation effect in the relationship between informed herd behavior, narrative and group feeling to phantasy. This shows us the strong impact of using similar sources of information and performing similar analyzes. It also serves as a guide for manipulating phantasy. Formation of financial bubbles can be controlled if sources of information are correctly directed by public authorities.


The AI Governance Challenge

Nathan: How do you think this is relevant to an applied setting?

Selim: It is very important that the phantasy effect, which we consider as the determinant of financial bubbles, is also accepted by public authorities. In this way, they will know not only how to manage conscious processes but also to take actions to manage unconscious processes. Also, having knowledge about the determinants of this unconscious process is very useful for their understanding of the process. Because our knowledge up to now has been aimed at providing conscious explanations of human madness, which causes policy to be implemented only against  conscious processes. Thanks to the findings of this research, the awareness of unconscious processes can also be developed and policies related to this can be developed.

Nathan:What do you think some exciting directions are for research stemming from your study? 

Selim: We are excited to have created awareness of the impact of unconscious processes and phantasy on financial decisions. It is also pleasing that we have developed the first financial scales for these concepts. We believe that the findings of this study will be a guide for future studies. As the empirical studies that follow this research increase, different relationships can be found and generalizations can be made.