The Penn Center for Social Norms and Behavioral Dynamics has launched a new Working Paper Series to showcase pre-publication versions of academic articles, book chapters, or reviews by our Center members and affiliates. With this new series, we seek to provide open access to high-quality publications from various disciplines and from our team of affiliates around the globe, across multiple universities and research organizations.

Our Working Paper series will also provide our researchers and affiliates the opportunity to share preliminary research findings with the broader research community, encourage discussion, and facilitate feedback from researchers working on related topics. Most papers will be freely available through the SSRN website. To access our published papers, books and reports please visit our main Research page.

19. In science we (should) trust: Expectations and compliance across nine countries during the COVID-19 pandemic

Cristina Bicchieri, Enrique Fatas, Abraham Aldama, Andrés Casas, Ishwari Deshpande, Mariagiulia Lauro, Cristina Parilli, Max Spohn, Paula Pereira, and Ruiling Wen

The magnitude and nature of the COVID-19 pandemic prevents public health policies from relying on coercive enforcement. Practicing social distancing, wearing masks and staying at home becomes voluntary and conditional on the behavior of others. We present the results of a large-scale survey experiment in nine countries with representative samples of the population. We find that both empirical expectations (what others do) and normative expectations (what others approve of) play a significant role in compliance, beyond the effect of any other individual or group characteristic. In our vignette experiment, respondents evaluate the likelihood of compliance with social distancing and staying at home of someone similar to them in a hypothetical scenario. When empirical and normative expectations of individuals are high, respondents’ evaluation of the vignette’s character’s compliance likelihood goes up by 55% (relative to the low expectations condition). Similar results are obtained when looking at self-reported compliance among those with high expectations. Our results are moderated by individuals’ trust in government and trust in science. Holding expectations high, the effect of trusting science is substantial and significant in our vignette experiment (22% increase in compliance likelihood), and even larger in self-reported compliance (76% and 127% increase before and after the lockdown). By contrast, trusting the government only generates modest effects. At the aggregate level, the country-level trust in science, and not in government, becomes a strong predictor of compliance.

This paper was published in PLoS ONE, 16(6):

18. Deviant or Wrong? The Effects of Norm Information on the Efficacy of Punishment

Cristina Bicchieri, Eugen Dimant, and Erte Xiao

Research examining the effect of weak punishment on conformity indicates that punishment can backfire and lead to suboptimal social outcomes. We examine whether this effect is due to a lack of perceived legitimacy of rule enforcement, which would enable agents to justify selfish behavior. We address the question of legitimacy by shedding light upon the importance of social norms and their interplay with weak punishment in the context of a trust game. Across six conditions, we systematically vary the combination of the existence of weak punishment and norm information. Norm information may refer either to what most others do (empirical) or to what most others deem appropriate (normative). We show that in isolation, neither weak punishment nor empirical/normative information increase prosocial, reciprocal behavior. We instead find that reciprocity significantly increases when normative information and weak punishment are combined, but only when compliance is relatively cheap. When compliance is more costly, we find that the combination of punishment and generic empirical information about others’ conformity can have detrimental effects. In additional experiments, we show that this negative effect can be attributed to the punishment being perceived as unjustified, at least in some individuals. Our results have important implications for researchers and practitioners alike.

17. The Impact of Transfers on Citizen Cooperation: Theory and a Lab-in-the-Field Test in Mexico

Abraham Aldama

Governments commonly use aid and social programs to garner citizen support during conflict, garner electoral support in democracies, and thwart collective action in autocracies. I explore whether monetary transfers increase citizens’ cooperation with an authority. To do so, I develop a game-theoretic model of citizen contribution to a public good from which an authority bene fits. I test the model’s implications by setting up a lab-in-the- field in Mexico. The model and the empirical test are designed to analyze variations in both of citizens’ intrinsic and extrinsic motivations to cooperate with the authority. The results show that citizens reduce their cooperation rates when an authority can spend money but does not. The results are consistent with a reduction in citizens’ intrinsic motivation when the authority actively chooses not to spend.

16. Intentionality Matters for Third-Party Punishment but not Compensation in Trust Games

Cristina Bicchieri and Marta Maras

We investigate how the intentionality of investors or trustees’ actions affects third party compensation and punishment interventions after a trust game. Investors and trustees are randomly assigned to conditions where they either make intentional choices or their choices are made by a random machine. Overall, we find that lack of reciprocity is punished more than lack of trust, and third parties exhibit strong preferences for compensation over punishment. We find that only the punishment choice is affected by the intentionality of parties’ actions, whereas compensation occurs in all conditions, whether lack of trust or reciprocity has been intentional or unintentional

15. The Power of Narratives in Social Norm Interventions: A Study of the Civic Culture Initiatives of Antanas Mockus in Bogotá, Colombia

Paulius Yamin, Saadi Lahlou, Santiago Ortega, Andrés Sáenz

Interventions that seek to transform social norms and behaviors in ways that benefit collective life are becoming more and more popular among policy and development practitioners around the world. The potential of these interventions to create behavioral changes in their target population is often determined by the narratives that participants and other stakeholders create and share around them. In this paper, we focus on a now classic intervention (the mime-artist initiative by Antanas Mockus in Bogotá, Colombia) to illustrate how narratives are essential to analyze how different stakeholders understand and make sense of social norm interventions. To do this, we collected narratives from 117 citizens, 80 press articles spanning over 24 years, and some of the original designers of the intervention. Using automated textual analysis and manual coding of Narrative Policy Framework categories, we analysed the content and structure of their narratives. Our findings describe the main characteristics that structure the narratives of citizens and the press in this successful case, and which pertain to (i) the main themes, (ii) the main characters and their agency, and (iii) the audience and their agency. We discuss the research and practical implications of these characteristics, with a focus on how narratives support the impact that the intervention had by allowing stakeholders to make sense of it, and by promoting collective self-observation, reflexivity and collective action around it. By doing this, we argue that good behavioral interventions create stories, and we propose seven recommendations that, based on this experience, could inform the design of more effective interventions to achieve positive policy outcomes.

14. Social Motives for Sharing Conspiracy Theories

Zhiying (Bella) Ren, Eugen Dimant and Maurice Schweitzer

Why do people share conspiracy theories? Recent work suggests that people share misinformation because they are inattentive. We find that people also knowingly share misinformation to advance social motives. Across three preregistered studies (total N=1,560 Prolific workers), we investigate the social motives for sharing conspiracy theories. We find that people are willing to trade off accuracy to build social connections when making content sharing decisions. Moreover, even though people know that factual news are more accurate than conspiracy theories, they expect sharing conspiracy theories to generate higher social value than sharing factual news. Lastly, in an interactive multi-round content-sharing paradigm, we find that social feedback could change the social value people attach to sharing misinformation. Our findings substantially develop our understand of why and when individuals are most likely to share conspiracy theories. These findings also make important contributions to understanding and curbing the spread of misinformation.

13. Persuading Republicans and Democrats to Comply with Mask Wearing: An Intervention Tournament

Michele Gelfand, Ren Li, Eftychia Stamkou, Dylan Pieper, Emmy Denison, Jessica Fernandez, Virginia Choi, Jennifer Chatman, Joshua Jackson, Eugen Dimant

Many people practiced COVID-19-related safety measures in the first year of the pandemic, but Republicans were less likely to engage in behaviors such as wearing masks or face coverings than Democrats, suggesting radical disparities in health practices split along political fault lines. We developed an “intervention tournament” which aimed to identify the framings that would promote mask wearing among a representative sample of Republicans and Democrats in the U.S. (N = 4,931). Seven different conditions reflecting different moral values and factors specific to COVID-19—including protection from harm (self), protection from harm (community), patriotic duty, purity, reviving the economy, threat, and scientific evidence—were implemented to identify which framings would “win” in terms of promoting mask wearing compared to a baseline condition. We found that Republicans had significantly more negative attitudes toward masks, lower intentions to wear them, and were less likely to sign or share pledges on social media than Democrats, which was partially mediated by Republicans, compared to Democrats, perceiving that the threat of COVID-19 was lower. None of our framing conditions significantly affected Republicans’ or Democrats’ attitudes, intentions, or behaviors compared to the baseline condition, illustrating the difficulty in overcoming the strength of political polarization during COVID-19.

12. Paying Them to Hate US: The Effect of U.S. Military Aid on Anti-American Terrorism, 1968-2018

Eugen Dimant, Tim Krieger and Daniel Meierrieks

Does U.S. military aid make the United States safer? Or does it have unintended and adverse consequences for U.S. security? We provide causal estimates of the effect of U.S. military aid on anti-American terrorism for a sample of 174 countries between 1968–2018 by exploiting plausibly exogenous time variation in global levels of U.S. military aid associated with distinct aid programs and cross-national time-series variation in the relative importance of the various military aid programs for recipient countries. We find that higher levels of military aid lead to an increased likelihood of the recipient country to produce anti-American terrorism. For our preferred instrumental-variable specification, doubling U.S. military aid increases the risk of anti-American terrorism by 2.7 percentage points. Examining potential transmission channels, we find that more U.S. military aid leads to more corruption and exclusionary policies in recipient countries. We argue that U.S. military aid results in anti-American terrorism by undermining institutions in recipient countries and creating anti-American resentment among those parts of the population that do not have direct or indirect access to benefits arising from aid.

11. Observability and Social Image: On the Robustness and Fragility of Reciprocity

Garry Bolton, Eugen Dimant and Ulrich Schmidt

Theoretical and empirical findings suggest that individuals are sensitive to the observability of their actions and the downstream consequences of this observability. We connect three streams of literature (social preferences, behavior change, and social norms) to investigate the conditions for which these claims are valid. Across multiple high-powered studies, we examine the mechanisms through which observability of one’s actions affects pro-sociality, when and why it sometimes fails, and how to utilize social and economic incentives to enact behavior change. Our three main results are: (i) observability alone has very little positive effect and can even backfire; (ii) inequality aversion drives the observed backfiring of observability; (iii) increasing the salience of norms can mitigate unintended consequences and successfully increase pro-sociality. From a policy perspective, our results highlight the potential pitfalls of simple behavioral interventions.

10. Expelled from (Tax) Heaven: Empowering consumers to reduce corporate tax avoidance

Enrique Fatas, Antonio J Morales and Axel Sonntag

We analyse corporate tax avoidance in a stylized experimental Bertrand setting with homogenous products and symmetric firms and consumers. More specifically, we investigate how market size and information disclosure of firms’ tax avoidance behaviour could reduce corporate tax avoidance. We find that making corporate tax behaviour more transparent by imposing a tax rating, makes consumers actively and costly boycott firms that do not pay their taxes. Firms anticipate consumer boycotts and increase their tax payments accordingly. When rating disclosure is voluntary, the positive effect on corporate tax compliance vanishes in large markets

9. The Democratic Peace: An experimental test of a causal relation and of underlying mechanisms

Jordi Brandts, Catherine Eckel, Enrique Fatas and Shaun Hargreaves-Heap

Democracies go to war with each other less frequently than dictatorships do with each other. This is an established empirical regularity. However, it is not clear whether there is a causal link between democracy and peace. We use laboratory experiments to study whether there is a causal impact. We study the bellicosity of democracies compared with two types of dictatorships, inclusive and exclusive, where each society is composed of three members. We also analyze how bellicosity depends on the presence of the possibility of deliberation between the members of a society. Neither the ‘voting’ nor ‘inclusion’ aspect of democracy nor ‘deliberation’ in isolation has a positive causal impact on peace. However, when all three are combined, there is evidence that their combination produces less bellicosity than some kinds of dictatorship. It is the addition of deliberation that makes the crucial distinguishing difference for democracy in our experiment.

8. In a Transparent Gov We Trust

Natalia Borzino, Enrique Fatas and Emmanuel Peterle

We transparency and reputation in a stylized investment game in which two senders interact with a receiver in a linear network. We analyze how ex-ante information (compatible with receiver’s reputation building) and ex-post information (making the receiver’s decision transparent) make senders and receivers comply with an exogenous, fair and non-binding rule. In the reputation treatment, senders know the average amount returned by the receiver to both senders in the previous period. In the transparency treatment, senders are aware of the amount received back by the other sender in the network. Relative to a baseline treatment in which senders are only informed about their own actions and outcomes, ex-ante and ex-post information disclosure have very differently effects on trust. While reputation helps some receivers, aggregate trust does not significantly change in the reputation condition. Transparency generates a large and significant increase in trust (amounts sent are 27% above the ones chosen in reputation), being the effect largely driven by a sustained and large reduction of senders choosing not to trust at all (56% less in transparency than in reputation in the second half of the experiment). Across all treatments and roles, the non-binding rule significantly and positively shapes individual decisions.

7. How do variations in contextual dimensions impact interventions? Insights from a norm-centric behavior change sanitation intervention delivered during COVID-19 pandemic in peri-urban Tamil Nadu, India

Kavita Chauhan, Sania Ashraf, Upasak Das, Alex Shpenev and Cristina Bicchieri

Contextual factors include factors operating at various levels such as socio-cultural, political, and physical, which may influence of moderate the research process and outcomes. An understanding and assessment of context helps in framing research insights. In early 2020, we launched a cluster randomised trial study of a norm-centric behavioural intervention in peri-urban Tamil Nadu which aimed to improve toilet ownership, use and maintenance among households living in communities with high toilet coverage (>65%). Before the initiation of our study, political developments created social unrest which disrupted the study. Soon after the commencement of the study, the entire country was under a COVID 19 lockdown which delayed the implementation process.

This paper has now been submitted to a journal for review, so the manuscript is no longer available here. The published version will be uploaded in our Research page when available.

6. Against the Wind: A Lab-in-the-field Experiment with Victims and Non-victims of Conflict in Colombia

Enrique Fatas & Lina Restrepo-Plaza

We present a lab-in-the-field experiment run with victims and non-victims of conflict in Colombia. Participants are either victims of conflict, ex-combatants or members of vulnerable groups, sharing socioeconomic characteristics with victims with the exception of conflict exposure. Participants are paired and they make unconditional contributions to a public good, and then conditional on the contribution of their counterpart using the strategy method. By comparing decisions across pairs we elicit outgroup discrimination (towards ex-combatants) and discrimination towards the ‘other ingroup’ (non-victims for victims, and victims for non-victims). By comparing unconditional and conditional decisions, we distinguish between beliefs- and preferences-based discrimination. We also elicit participants’ expectations (empirical and normative) in every interaction and study the role played by social norms. Both victims and non-victims discriminate ex-combatants significantly more than they discriminate each other, being their discrimination preference-based. However, victims discriminate ex-combatants significantly less than non-victims. While victims do not discriminate non-victims, they are discriminated by non- victims, and their behavior is consistent with different social norms: ingroup discrimination of victims is supported by discriminatory norms held by non-victims. However, victims and non-victims hold the same discriminatory expectations towards ex-combatants and only non-victims conform to the norm.

5. Politicizing Mask-Wearing: Predicting the Success of Behavioral Interventions Among Republicans and Democrats

Eugen Dimant, Dylan Pieper, Elena Giulia Clemente, Anna Dreber and Michele Gelfand

Consortium Co-Authors: BCFG (Behavior Change for Good), Michael Hallsworth and Aline Holzwarth

Scientists and policymakers seek to choose effective interventions that promote preventative health measures. We evaluated whether academics, behavioral science practitioners, and laypeople (N = 1,034) were able to forecast the effectiveness of seven different messages compared to a baseline message for Republicans and Democrats separately. These messages were designed to nudge mask-wearing attitudes, intentions, and behaviors. When examining prediction error across political parties, forecasters predicted larger effects than those observed for Democrats compared to Republicans and made more accurate predictions for Republicans compared to Democrats. These results are partly driven by a lack of nudge effects on Democrats, as reported in Gelfand et al. (2021). Academics and practitioners made more accurate predictions compared to laypeople. Although forecasters’ predictions were correlated with the nudge interventions, all groups overpredicted the observed results. We discuss how accurate forecasts of behavioral intervention outcomes provide insight that can help save resources and increase the efficacy of interventions by updating the accuracy of forecasters’ beliefs.

4. Nudging Enforcers: How Norm Perceptions and Motives for Lying Shape Sanctions

Eugen Dimant & Tobias Gesche

The enforcement of social norms is the fabric of a functioning society. Through the lens of two experiments, we examine how motives for lying and norm perceptions steer enforcement. Our contribution is to investigate the extent to which norm breaches are sanctioned, how norm-nudges affect the observed punishment behavior, and how the enforcement is linked to norm perceptions. Using a representative U.S. sample, Experiment 1 provides robust evidence that norm-enforcement is not only sensitive to the extent of the observed transgression (= size of the lie) but also to its consequences (= whether the lie remedies or creates payoff inequalities). We also find norm enforcers to be sensitive to different norm-nudges that convey social information about actual lying behavior or its social disapproval. To explain the punishment patterns, Experiment 2 examines how norms are perceived across different transgressions and how norm-nudges change these perceptions. We observe a malleability of social norm perceptions: norm nudges are most effective when pre-existing norms are vague. Importantly, we also find that punishment patterns in the first experiment closely follow these norm perceptions. Our findings suggest that norm enforcement can be successfully nudged and thus represent an expedient alternative to standard incentive-based interventions.

3. Perceived Inequality and Policy Preferences

Abraham Aldama, Cristina Bicchieri and Jana Freundt

We study the hypothesis that perceived income inequality affects people’s fairness views and their support for redistribution. In a comprehensive and well-powered survey experiment with a representative sample of US Americans, we (1) asses a causal link between the perceived income distribution and political views and (2) test the role of possible moderators. In particular, we investigate how own income, partisanship, trust in government, perceived equality of opportunity and perceived autonomy can predict how strongly a person’s perception of inequality impacts her political views. We measure participants’ political views as their fairness judgments and their support for redistribution by public and private means. Additionally, we obtain a behavioral measure for the willingness to (privately) redistribute, a donation to a charity. We find precisely estimated null effects. There is no evidence for a causal effect of perceived inequality on political views or behavior and this finding holds for all subgroups. In the light of increasing economic inequality since the 1960s in the US that has not been accompanied by increased taxation of the rich, the question how US Americans (mis-)perceive this inequality and to what extent this translates into a demand for more redistribution has become an important policy question. Our study suggests that informing people about the extent of inequality in a society will not effectively alter their support for redistributive policies.

2. Toilet Use is a Descriptive Norm: The Influence of Social Expectations on Toilet Use in Bihar and Tamil Nadu, India

Erik Thulin, Alex Shpenev, Sania Ashraf, Upasak Das, Jinyi Kuang, Cristina Bicchieri

Open defecation is a global public health issue. We applied Bicchieri’s Social Norms Framework to diagnose this behavior and determine what type of interventions could be effective in the Indian context. We conducted a mixed method study in rural, urbanizing, and slum areas of Bihar and Tamil Nadu. Using data from randomly selected individuals (n=5052) we assessed toilet use, empirical expectations (beliefs about what other people do), normative expectations (beliefs about what other people think one should do), and the dependence of toilet use behavior on those expectations. We found that empirical expectations were a strong driver of toilet use, while normative expectations had negligible predictive value. Only a minority of respondents believed there were any negative sanctions for defecating in the open. Taken together, these findings indicate that toilet use is a descriptive norm. We therefore conclude that nudges in the form of information about similar other’s improved practices might be an effective behavior change strategy to improve toilet use.

1. Social Proximity and the Erosion of Norm Compliance

Cristina Bicchieri, Eugen Dimant, Simon Gächter and Daniele Nosenzo

We study how compliance with norms of pro-social behavior is influenced by peers’ compliance in a dynamic and non-strategic experimental setting. We show that social proximity among peers is a crucial determinant of the effect. Without social proximity, norm compliance erodes swiftly because participants only conform to observed norm violations while ignoring norm compliance. With social proximity, participants conform to both types of observed behaviors, thus halting the erosion of compliance. Our findings stress the importance of the broader social context for norm compliance and show that, even in the absence of social sanctions, norm compliance can be sustained in repeated interactions, provided there is group identification, as is the case in many natural and online environments.