Let's (not) get together!: The role of social norms in social distancing during COVID-19

Martinez, Deborah; Parilli, Cristina; Scartascini, Carlos G.; Simpser, Alberto Working Paper Let's (not) get together!: The role of social norms in social distancing during COVID-19 IDB Working Paper Series, No. IDB-WP-1168 Provided in Cooperation with: Inter-American Development Bank (IDB), Washington, DC Suggested Citation: Martinez, Deborah; Parilli, Cristina; Scartascini, Carlos G.; Simpser, Alberto (2021) : Let's (not) get together!: The role of social norms in social distancing during COVID-19, IDB Working Paper Series, No. IDB-WP-1168, Inter-American Development Bank (IDB), Washington, DC, https://doi.org/10.18235/0003044 This Version is available at: https://hdl.handle.net/10419/237463 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/3.0/igo/legalcode �4'IDB Inter-American Development Bank IDB WORKING PAPER SERIES N° IDB-WP-1168 Let's (Not) Get Together! The Role of Social Norms in Social Distancing during COVID-19 Deborah Martinez Cristina Parilli Carlos Scartascini Alberto Simpser Inter-American Development Bank Department of Research and Chief Economist February 2021 Abstract While effective preventive measures against COVID-19 are now widely known, many individuals fail to adopt them. This paper provides experimental evidence about one potentially important driver of compliance with social distancing: social norms. We asked each of 23,000 survey respondents in Mexico to predict how a fictional person would behave when faced with the choice about whether or not to attend a friend’s birthday gathering. Every respondent was randomly assigned to one of four social norms conditions. Expecting that other people would attend the gathering and/or believing that other people approved of attending the gathering both increased the predicted probability that the fictional character would attend the gathering by 25%, in comparison with a scenario where other people were not expected to attend nor to approve of attending. Our results speak to the potential effects of communication campaigns and media coverage of, compliance with, and normative views about COVID-19 preventive measures. They also suggest that policies aimed at modifying social norms or making existing ones salient could impact compliance. JEL classifications: D91; D90; I12; I18 Keywords: COVID19, Social norms, Social distancing, Normative expectations, Empirical expectations, Compliance We are thankful to Mart´ın Ardanaz, Ivanna Valverde, Fernando Cafferata, Ana Mar´ıa Rojas M, Simeon Schachtele, Jorge Streb, Eugen Dimant and participants in the Impact Evaluation Meeting on COVID-19 for their comments. All remaining errors are our own. This project has been IRB approved by ITAM. The pre-experimental plan for the larger project has been filed under the name “Social and Behavioral Drivers of Individual Compliance with Preventive Measures during the COVID-19 Epidemic in Mexico.” The information and opinions presented herein are entirely those of the authors, and no endorsement by the Inter-American Development Bank, its Board of Executive Directors, or the countries they represent is expressed or implied. 1 1 Introduction Since the COVID-19 pandemic began in early 2020, much has been learned about how infection can be prevented. In particular, social distancing and avoiding indoor gatherings have emerged as some of the most powerful and effective preventive behaviors (WHO, 2020). Despite the strength of the evidence on the dangers of close social contact (Frieden & Lee, 2020; Aschwanden, 2020), many people continue to gather with friends and to participate in social events (Miles, 2020; Shotsky, 2020; Holcombe & del Valle, 2020), which has helped the virus to potentially spread even to the highest political circles (Liptak, 2020; Margolin & Bruggeman, 2020). If the pandemic is to be contained, it is crucial to understand what drives people to engage in behavior that is inconsistent with the available scientific evidence and public health guidelines (Van Bavel et al., 2020). The problem does not appear to be one of information or credibility, as survey evidence shows that most people agree that social gatherings ought to be avoided. As far back as May of 2020, 79.5% of survey respondents in the United States agreed that gatherings of 10 or more people should not be allowed (Center for Disease Control, 2020). In Mexico, the country where we conducted the present study, 82% of those surveyed in April of 2020 approved of the public health guidelines in place, which included restrictions on mass gatherings (Buend´ıa & Laredo, 2020). According to our own data, 73% of people recognize that gathering in enclosed spaces, such as restaurants, represents a high risk for contracting COVID-19. Still, about 43% recognize having visited friends and family in their homes during the previous week. In this article, we investigate the role of social norms on compliance with preventive behaviors—specifically with social distancing. We do so by conducting a survey experiment on more than 23,000 individuals in Mexico. The experiment consists of a vignette, described in the form of a story, depicting a fictional individual, Mariana, who has been invited to attend a friend’s birthday gathering and must decide whether or not to attend. This story portrays a situation that most Mexicans can relate to (birthday celebrations) and what the literature highlights to be individuals’ relevant reference network during the current pandemic (family and friends) (Goldberg et al., 2020). These social gatherings are also relevant because they have been shown to lead to superspreading events (Frieden & Lee, 2020; Aschwanden, 2020). The treatments randomly assign respondents to different social norms prompts, providing information on Mariana’s beliefs about: i) whether other invitees will attend the gathering (empirical expectations), and ii) whether other invitees approve of others’ attending the gathering (normative expectations). After being exposed to the social 2 norms prompt, respondents are asked to state whether they believe that Mariana will attend the gathering, and whether they believe that Mariana should attend the gathering. We find that the prompt about whether others are likely to attend has a strong effect on the respondent’s prediction as to whether Mariana will attend the gathering or not. These findings are in line with prior findings, in settings other than the current COVID-19 pandemic, that individuals tend to conform to what they perceive is the prevailing behavior (Asch, 1951; Cialdini et al., 1990, 1991; Bicchieri, 2006; Cialdini et al., 2006; Lapinski et al., 2017). Interestingly, we find no effect of any of the treatments on respondent predictions about what Mariana ought to do: the overwhelming majority believe she should not attend. 2 Theoretical Background It has long been argued that individual behavior is strongly influenced by what others do (descriptive norms) and what others approve of doing (prescriptive or injunctive norms) (Coleman, 1990; Parsons, 1991; Cialdini et al., 1991; Bicchieri, 2006; John et al., 2019). The literature accords different roles and effects to descriptive versus. injunctive norms (Bicchieri & Dimant, 2019). Descriptive norms indicate those cases in which you prefer to carry out an activity because you believe it meets your needs (unconditional preference) or because you expect others to do it (conditional preference). Injunctive norms indicate those cases in which you prefer to engage in an activity because you believe it is the right thing to do (unconditional preference), or because you expect others to engage in the activity and believe that others think that you should do so as well (conditional preference). In this latter case of conditional preference, choices and behaviors depend on both empirical expectations (what you believe others are doing) and normative expectations (what you believe others think you should do) (Bicchieri & Dimant, 2019). In our setup, a social norm is a rule that maps empirical and normative expectations onto behaviors. A social norm is followed by individuals in a population “on the condition that they believe that i) most people in their reference network conform to it (empirical expectation) and ii) that most people in their reference network believe they ought to conform to it (normative expectation)” (Bicchieri & Dimant, 2019, p.5). Both empirical and normative expectations have been shown to influence behavior. Policymakers, for example, have increasingly made use of social norms to nudge individuals in diverse contexts, with goals such as reducing medical prescriptions, increasing tax compliance, and reducing energy and water consumption (Coleman, 2007; Thaler & Sunstein, 3 2009; Allcott, 2011; Ferraro et al., 2011; Hallsworth et al., 2016; Bhanot, 2018), and social norms can also affect willingness to enforce and sanction violations (Schelling, 1960; Traxler & Winter, 2012; Acemoglu & Jackson, 2017). Social norms could be extremely relevant for explaining and affecting behaviors during the current pandemic (Van Bavel et al., 2020; Lunn et al., 2020a). Goldberg et al. (2020) and Smith et al. (2020) find that an individual’s perceptions about how many others abide by social distancing correlate with the individual’s propensity to social distance herself, and the effect of social norms can be stronger on individuals lacking a sense of duty (Bourgeois et al., 2020). As people seek to conform or to imitate the behavior of others (Asch, 1951), news coverage of celebrities or political leaders failing to abide by, or criticizing, preventive behaviors (Miller et al., 2020; Blunt, 2020) could in fact reduce public compliance with such behaviors, as they might be “normalizing” them in the eye of the public (Ashforth & Anand, 2003; Bicchieri, 2016; Lindstr¨om et al., 2018). However, norm-based interventions and media coverage of events showing compliance with preventive behaviors can potentially help (Jiang et al., 2021). Still, it is worth noting that norm-based messages might not have any differential effect on the understating of COVID-19 guidelines (Bilancini et al., 2020) and that norm nudges need to include more than informative messages to be effective (Hume et al., 2020). These findings make it even more important to investigate how and why social norms would change people’s compliance with preventive behaviors in order to further refine future interventions and massive communication efforts. Bicchieri et al. (2020) run a survey experiment similar to ours where normative and empirical expectations are randomly varied in a 2-by-2 schema, and respondents are then asked to predict the compliance of a fictional third party with social distancing. That study, like ours, finds that assignment to the condition with “high” normative and empirical expectations promoted compliance. However, our approaches differ in three important dimensions. First, instead of asking whether the third party would abide by social distancing in general, we confront the respondent with a very specific scenario: whether or not to attend the birthday party of a close friend. We believe that our approach is more concrete and therefore less prone to eliciting abstract responses colored by social desirability biases or demand effects. Second, instead of using a Likert scale we force a dichotomic yes/no response that mimics many social distancing choices: one can either attend a gathering or refrain from attending. Third, we elicit both predicted behavior and respondent normative views, which allows us to study whether any effects on (predicted) behavior might be underpinned by, or correlated with, effects on normative assessments. 4 Our paper builds on a recent but strong behavioral literature studying behaviors associated with the current COVID-19 pandemic that attempts to promote preventive behaviors and a more effective pandemic response (Van Bavel et al., 2020). Capraro & Barcelo (2020b) shows that individuals primed with “reasoning” messages are more willing to wear face masks than those primed to “rely on their emotions,” which points out that people’s compliance can be increased if they are not driven by emotions in their decision-making. Lunn et al. (2020b) shows that highlighting the risks associated with not following social distance have a larger effect than providing information. Everett et al. (2020) highlights that a “deontological” message, based on people’s duty to do the right thing for their families and friends, seems to be more effective than utilitarian or moral messaging. Along this line, Capraro & Barcelo (2020a), Heffner et al. (2020), and Jordan et al. (2020) findings are also consistent with the idea that prosocial motivation is effective in promoting intention to comply with preventive behaviors, particularly if they are able to develop individuals’ empathy towards those more vulnerable to being infected (Pfattheicher et al., 2020). These findings are relevant, as they allow us to understand how individuals perceive and act according to the consequences of their own personal actions on others. Thus, this lays the groundwork to go even further and also understand how individuals react when faced with the behavior of others—that is, how perceived social norms can change individuals’ behavior even if they were personally willing to comply with preventive measures due to prosocial motives. Can the perception of what others do and approve of change individuals’ intentions of complying with public health guidelines? Our study aims to contribute to the related literature and complement other similar studies conducted during the pandemic. 3 Methods 3.1 Participants Our survey experiment was part of a broader COVID-19-focused survey in Mexico, approved by the IRB of the Instituto Tecnol´ogico Aut´onomo de M´exico (ITAM) on July 1, 2020, under the name “Social and Behavioral Drivers of Individual Compliance with Preventive Measures during the COVID-19 Epidemic in Mexico” (memorandum letter of approval available upon request from the authors). The questionnaire was pre-tested on a small sample of colleagues and acquaintances, and subject to the IRB’s recommendations. Survey respondents were recruited through a Facebook ad campaign and a separate email campaign. The Facebook ad campaign targeted a general audience composed of individuals over 18 years of age living 5 of treatment branches T2 (high empirical, low normative) versus. T3 (low empirical, high normative) suggests that empirical expectations matter more than normative expectations, as claimed in Bicchieri & Xiao (2009). At the same time, the estimated effect of treatment T4 (high empirical, high normative) is smaller in magnitude than, and statistically different from, that of treatment T2. This is surprising, since one might expect that when normative and empirical expectations are aligned (T4), the effect on behavior should be larger—yet this is not what we find. We take our results on the mixed treatments (T2 and T3) as an indication that empirical and normative expectations may interact in ways that are poorly understood (perhaps some form of crowding out is at work) and merit further research. Our study design, of course, has limitations. First, it is not obvious that the intensity of treatment is comparable across arms: it could be that changes in the perceived empirical expectations are greater than a change in normative expectations. Second, our results ought to be interpreted in the context of the fact that Mariana is said, in the vignette, to generally comply with public health guidelines. Therefore, respondents may infer that Mariana may care more about what her friends like her do (T1 and T2) than those friends who do not think like her (T3 and T4). Lastly, our estimations are based on the perception of participants on how others (Mariana) would behave in this scenario. We therefore cannot assure that participants would act similarly if they found themselves in a similar position. Our findings contribute to the general research on the relationship of social norms with behavior and are relevant for the design of communication strategies in both the public and private sectors. Highlighting that others are not complying is likely to reduce compliance, and this could be an unintended byproduct of news coverage of noncompliance. Politicization of the guidelines, and active and public repudiations of norms, can also lead to further erosion of compliance. Additionally, targeting normative expectations—what people ought to be doing—will likely not suffice to induce the desired behaviors unless people also expect others to comply. Thus, information highlighting others’ compliance and targeting normative expectations at the same time are likely to play an essential role in any successful information campaign seeking to encourage individuals to adopt preventive behaviors. 12 References Acemoglu, D. & Jackson, M. (2017). Social norms and the enforcement of laws. Journal of the European Economic Association, 15(2), 245–295. Allcott, H. (2011). Social norms and energy conservation. Journal of Public Economics, 95(9-10), 1082–1095. Asch, S. E. (1951). Effects of group pressure upon the modification and distortion of judgments. In H. Guetzkow (Ed.), Groups, leadership and men; research in human relations (pp. 177––190). Carnegie Press. Aschwanden, C. (2020). How ‘superspreading’ events drive most covid-19 spread. Scientific American.https://www.scientificamerican.com/article/ how-superspreading-events-drive-most-covid-19-spread1 (accessed: 09.24.2020). Ashforth, B. E. & Anand, V. (2003). The normalization of corruption in organizations. Research in organizational behavior, 25, 1–52. Bhanot, S. P. (2018). Isolating the effect of injunctive norms on conservation behavior: New evidence from a field experiment in california. Organizational Behavior and Human Decision Processes. Bicchieri, C. (2006). The Grammar of Society: The Nature and Dynamics of Social Norms. Cambridge, UK: Cambridge University Press. Bicchieri, C. (2016). Norms in the wild: How to diagnose, measure, and change social norms. Oxford University Press. Bicchieri, C., Ashraf, S., Das, U., Delea, M., Kohler, H.-P., Kuang, J., McNally, P., Shpenev, A., & Thulin, E. (2018). Phase 2 project report. social networks and norms: Sanitation in bihar and tamil nadu, india. Penn Social Norms Group, 17. Bicchieri, C. & Dimant, E. (2019). Nudging with care: The risks and benefits of social information. Public choice, (pp. 1–22). Bicchieri, C., Fatas, E., Aldama, A., Casas, A., Deshpande, I., Lauro, M., Parilli, C., Spohn, M., Pereira, P., & Wen, R. (2020). Norms, trust, and compliance with public health measures. Unpublished manuscript. Bicchieri, C. & Xiao, E. (2009). Do the right thing: but only if others do so. Journal of Behavioral Decision Making, 22(2), 191–208. Bilancini, E., Boncinelli, L., Capraro, V., Celadin, T., & Di Paolo, R. (2020). The effect of norm-based messages on reading and understanding covid-19 pandemic response governmental rules. arXiv preprint arXiv:2005.03998. 13 Blunt, G. D. (2020). Face mask rules: do they really violate personal liberty? The Conversation.https://theconversation.com/ face-mask-rules-do-they-really-violate-personal-liberty-143634 (accesed: 10.22.2020). Bourgeois, L. F., Harell, A., & Stephenson, L. B. (2020). To follow or not to follow: Social norms and civic duty during a pandemic. Canadian Journal of Political Science/Revue canadienne de science politique, 53(2), 273–278. Buend´ıa & Laredo (2020). Encuesta nacional de opini´on p´ublica. https:// www.buendiaylaredo.com/publicaciones/466/2004_ENCUESTA_COVID.pdf (accessed: 08.24.2020). Capraro, V. & Barcelo, H. (2020a). The effect of messaging and gender on intentions to wear a face covering to slow down covid-19 transmission. arXiv preprint arXiv:2005.05467. Capraro, V. & Barcelo, H. (2020b). Priming reasoning increases intentions to wear a face covering to slow down covid-19 transmission. arXiv preprint arXiv:2006.11273. Center for Disease Control (2020). Public attitudes, behaviors, and beliefs related to covid19, stay-at-home orders, nonessential business closures, and public health guidance — united states, new york city, and los angeles, may 5–12, 2020. https://www.cdc.gov/ mmwr/volumes/69/wr/mm6924e1.htm (accessed: 09.10.2020). Cialdini, R. B., Demaine, L. J., Sagarin, B. J., Barrett, D. W., Rhoads, K., & Winter, P. L. (2006). Managing social norms for persuasive impact. Social influence, 1(1), 3–15. Cialdini, R. B., Kallgren, C. A., & Reno, R. R. (1991). A focus theory of normative conduct: A theoretical refinement and reevaluation of the role of norms in human behavior. In Advances in experimental social psychology, volume 24 (pp. 201–234). Elsevier. Cialdini, R. B., Reno, R. R., & Kallgren, C. A. (1990). A focus theory of normative conduct: recycling the concept of norms to reduce littering in public places. Journal of personality and social psychology, 58(6), 1015. Coleman, J. S. (1990). Foundations of Social Theory. Belknap. Coleman, S. (2007). The minnesota income tax compliance experiment: replication of the social norms experiment. Available at SSRN 1393292. Everett, J. A., Colombatto, C., Chituc, V., Brady, W. J., & Crockett, M. (2020). The effectiveness of moral messages on public health behavioral intentions during the covid-19 pandemic. Ferraro, P. J., Miranda, J. J., & Price, M. K. (2011). The persistence of treatment effects with norm-based policy instruments: evidence from a randomized environmental policy experiment. American Economic Review, 101(3), 318–22. 14 Frieden, T. R. & Lee, C. T. (2020). Identifying and interrupting superspreading events—implications for control of severe acute respiratory syndrome coronavirus 2. Emerging Infectious Disease, 26(6). Goldberg, M., Gustafson, A., Maibach, E., van der Linden, S., Ballew, M. T., Bergquist, P., Kotcher, J., Marlon, J. R., Rosenthal, S., & Leiserowitz, A. (2020). Social norms motivate covid-19 preventive behaviors. PsyArXiv preprints.https://doi.org/10.31234/osf. io/9whp4. Hallsworth, M., Chadborn, T., Sallis, A., Sanders, M., Berry, D., Greaves, F., Clements, L., & Davies, S. C. (2016). Provision of social norm feedback to high prescribers of antibiotics in general practice: a pragmatic national randomised controlled trial. The Lancet, 387(10029), 1743–1752. Heffner, J., Vives, M.-L., & FeldmanHall, O. (2020). Emotional responses to prosocial messages increase willingness to self-isolate during the covid-19 pandemic. Personality and Individual Differences, 170, 110420. Hogg, M. A., Abrams, D., Otten, S., & Hinkle, S. (2004). The social identity perspective: Intergroup relations, self-conception, and small groups. Small group research, 35(3), 246– 276. Holcombe, M. & del Valle, L. (2020). Back-to-back weekend parties at park near nyu raise coronavirus concerns. CNN.https://www.cnn.com/2020/09/14/us/ nyu-weekend-parties-coronavirus-concerns/index.html (accessed: 10.04.2020). Hume, S., John, P., Sanders, M., & Stockdale, E. (2020). Nudge in the time of coronavirus: The compliance to behavioural messages during crisis. Jiang, X., Hwang, J., Shah, D. V., Ghosh, S., & Brauer, M. (2021). News attention and social-distancing behavior amid covid-19: How media trust and social norms moderate a mediated relationship. Health Communication, (pp. 1–10). John, P., Sanders, M., & Wang, J. (2019). A panacea for improving citizen behaviors? introduction to the symposium on the use of social norms in public administration. Journal of Behavioral Public Administration, 2(2). Jordan, J., Yoeli, E., & Rand, D. (2020). Don’t get it or don’t spread it? comparing self-interested versus prosocially framed covid-19 prevention messaging. Lapinski, M. K., Zhuang, J., Koh, H., & Shi, J. (2017). Descriptive norms and involvement in health and environmental behaviors. Communication Research, 44(3), 367–387. Latkin, C., Donnell, D., Celentano, D. D., Aramrattna, A., Liu, T.-Y., Vongchak, T., Wiboonnatakul, K., Davis-Vogel, A., & Metzger, D. (2009). Relationships between social norms, social network characteristics, and hiv risk behaviors in thailand and the united states. Health Psychology, 28(3), 323. 15 Latkin, C. A. & Knowlton, A. R. (2015). Social network assessments and interventions for health behavior change: a critical review. Behavioral Medicine, 41(3), 90–97. Lindstr¨om, B., Jangard, S., Selbing, I., & Olsson, A. (2018). The role of a “common is moral” heuristic in the stability and change of moral norms. Journal of Experimental Psychology: General, 147(2), 228. Liptak, K. (2020). Inside one celebration that helped spread the virus across the us government. CNN.https://www.cnn.com/2020/10/03/politics/ trump-covid-amy-coney-barrett-event/index.html (accessed: 10.04.2020). Lunn, P. D., Belton, C. A., Lavin, C., McGowan, F. P., Timmons, S., & Robertson, D. A. (2020a). Using behavioral science to help fight the coronavirus. Journal of Behavioral Public Administration, 3(1). Lunn, P. D., Timmons, S., Belton, C. A., Barjakov´a, M., Julienne, H., & Lavin, C. (2020b). Motivating social distancing during the covid-19 pandemic: An online experiment. Social Science & Medicine, 265, 113478. Margolin, J. & Bruggeman, L. (2020). 34 people connected to white house, more than previously known, infected by coronavirus: Internal fema memo. ABC News.https://abcnews.go.com/Politics/ 34-people-connected-white-house-previously-infected-coronavirus/story? id=73487381 (accessed: 10.09.2020). Miles, F. (2020). Young americans are throwing massive coronavirus parties already nationwide. Fox News.https://www.foxnews.com/us/ young-americans-massive-coronavirus-parties-nationwide (accessed: 10.04.2020). Miller, Z., Colvin, J., & Madhani, A. (2020). Trump, still infectious, back at white house — without mask. APnews.https://apnews.com/article/ virus-outbreak-donald-trump-archive-d39bd670e8a280b6283abcdfc91d4794) (accessed: 10.04.2020). Parsons, T. (1991). The social system. Psychology Press. Pfattheicher, S., Nockur, L., B¨ohm, R., Sassenrath, C., & Petersen, M. B. (2020). The emotional path to action: Empathy promotes physical distancing and wearing of face masks during the covid-19 pandemic. Psychological Science, 31(11), 1363–1373. Rimal, R. N. & Real, K. (2005). How behaviors are influenced by perceived norms: A test of the theory of normative social behavior. Communication research, 32(3), 389–414. Schelling, T. C. (1960). The Strategy of Conflict. Harvard University Press. 16 Shotsky, A. (2020). Large parties at sdsu are a concern during covid pandemic. https:// www.cbs8.com/article/news/health/coronavirus/sdsu-parties-covid-concern/ 509-21cdbc1e-6f18-442c-a1a6-4218d83c866d (accessed: 10.04.2020). Smith, L. E., Amlˆot, R., Lambert, H., Oliver, I., Robin, C., Yardley, L., & Rubin, G. J. (2020). Factors associated with adherence to self-isolation and lockdown measures in the uk; a cross-sectional survey. medRxiv. Thaler, R. H. & Sunstein, C. R. (2009). Nudge: Improving decisions about health, wealth, and happiness. Penguin. Traxler, C. & Winter, J. (2012). Survey evidence on conditional norm enforcement. European Journal of Political Economy, 28, 390–398. Van Bavel, J. J., Baicker, K., Boggio, P. S., Capraro, V., Cichocka, A., Cikara, M., Crockett, M. J., Crum, A. J., Douglas, K. M., Druckman, J. N., et al. (2020). Using social and behavioural science to support covid-19 pandemic response. Nature Human Behaviour, (pp. 1–12). WHO (2020). Coronavirus disease (covid-19) advice for the public. https://www.who. int/emergencies/diseases/novel-coronavirus-2019/advice-for-public (accessed: 09.07.2020). 17 Figure 1: Treatment Effects: Joint Treatment and Controls Joint treatment Age Female Education Exposed Covid deathCovid Older 65 at home Had H1N1 Prob Infection Prob Hospital Went to party Visited family Risky inside Soc. Distance -.05 0 .05 .1 -.05 0 .05 .1 Note: 95% confidence interval Note: 95% confidence interval Will_attend Should_attend Coefficient Notes: This figure shows the coefficients for the joint treatment variable and the coefficients for the control variables. It corresponds to columns [2] and [6] in Table 3. 18 Figure 2: Treatment Effects 0 .05 .1 .15 Treatment effect T2(HN/LE) T3(LN/HE) T4(LN/LE) Will go Should go Note: 95% confidence interval Notes: This figure shows the treatment effects for the two dependent variables. They correspond to columns [2] and [6] in Table 3. 19 Table 2: Balance Table T1 Diff w.r.t. T1 (coeff and s.e.) p-value Wald test equality coefficients Sample Size (av and s.d.) T2 T3 T4 T2=T3=T4 T2=T3 T2=T4 T3=T4 [1] [2] [3] [4] [5] [6] [7] [8] [8] Age (group) 1.429 -0.011 -0.006 -0.000 0.585 0.610 0.301 0.598 22,896 (0.007) (0.010) (0.010) (0.010) 1.Female 0.660 0.003 0.004 -0.001 0.799 0.835 0.655 0.511 23,184 (0.006) (0.009) (0.009) (0.009) Education (group) 2.580 0.021* 0.008 0.018 0.548 0.293 0.803 0.427 22,925 (0.009) (0.012) (0.012) (0.012) 1.Exposed Covid 0.649 0.006 0.008 0.003 0.805 0.758 0.726 0.510 22,625 (0.006) (0.009) (0.009) (0.009) 1.Death Covid 0.576 -0.000 -0.000 0.002 0.958 0.994 0.803 0.796 23,184 (0.006) (0.009) (0.009) (0.009) 1.Older 65 0.265 -0.003 -0.002 -0.004 0.960 0.917 0.859 0.777 23,093 (0.006) (0.008) (0.008) (0.008) 1.Exposed H1N1 0.186 0.010 0.008 0.012* 0.832 0.730 0.796 0.546 23,184 (0.005) (0.007) (0.007) (0.007) Prob Infection 51.344 -0.062 0.206 0.098 0.879 0.613 0.765 0.839 22,964 (0.375) (0.532) (0.528) (0.534) Prob Hospital 45.429 0.080 -0.317 -0.320 0.621 0.397 0.398 0.993 22,988 (0.336) (0.474) (0.470) (0.476) 1.Attend Party 0.125 -0.006 -0.001 -0.002 0.701 0.429 0.521 0.885 23,087 (0.004) (0.006) (0.006) (0.006) 1.Visit 0.428 -0.008 0.005 -0.014 0.116 0.183 0.478 0.0411 23,085 (0.007) (0.009) (0.009) (0.009) 1.Risky Inside 0.734 0.005 -0.002 0.002 0.665 0.367 0.672 0.635 23,184 (0.006) (0.008) (0.008) (0.008) 1.Social Distance 0.360 0.008 -0.008 -0.004 0.189 0.080 0.184 0.681 23,098 (0.006) (0.009) (0.009) (0.009) Notes: Each row shows statistics for a different observable variable we have. Column [1] shows the sample average and the standard deviation in parenthesis for the control group -in this case, individuals in T1. Columns [2]-[4] shows the regression coefficient and the standard error in parenthesis corresponding to an OLS regression -observable is the dependent variable and the treatment variables are the independent ones. Standard errors are robust. *** p<0.01, ** p<0.05, * p<0.1 Columns [5]-[8] shows the p-value of a test of equality of coefficients. Column [9] shows the sample size for each regression. Variables [Age] and [Education] are tabulated according to ranges; as such they are categorical, with a higher category number referring to an older age and more years of education, respectively. 1.x refers to dummy variables. Source: Authors’ calculations 20 Table 3: Treatment effects Mariana will attend Mariana should attend (1) (2) (3) (4) (5) (6) (7) (8) T (T2+T3+T4) 0.073*** 0.076*** 0.076*** 0.077*** 0.001 0.001 0.002 0.002 (0.007) (0.007) (0.007) (0.007) (0.003) (0.003) (0.003) (0.003) Constant 0.264*** 0.321*** 0.381*** 0.339*** 0.033*** 0.107*** 0.142*** 0.129*** (0.006) (0.020) (0.066) (0.041) (0.002) (0.009) (0.033) (0.020) T2 0.098*** 0.100*** 0.100*** 0.101*** -0.002 -0.000 -0.000 0.001 (0.009) (0.009) (0.009) (0.009) (0.003) (0.003) (0.003) (0.003) T3 0.055*** 0.058*** 0.058*** 0.059*** 0.004 0.004 0.004 0.005 (0.009) (0.009) (0.009) (0.009) (0.003) (0.003) (0.003) (0.003) T4 0.067*** 0.069*** 0.070*** 0.071*** -0.000 0.000 0.000 0.001 (0.009) (0.009) (0.009) (0.009) (0.003) (0.003) (0.003) (0.003) Constant 0.264*** 0.322*** 0.378*** 0.338*** 0.033*** 0.107*** 0.143*** 0.129*** (0.006) (0.020) (0.067) (0.041) (0.002) (0.009) (0.033) (0.020) Observations 21,882 20,511 20,511 20,511 22,744 21,264 21,264 21,264 Controls No Yes Yes Yes No Yes Yes Yes Fixed Effects No No State Municipality No No State Municipality T2=T3=T4 0.000 0.000 0.000 0.000 0.189 0.412 0.394 0.437 T2=T3 0.000 0.000 0.000 0.000 0.076 0.229 0.208 0.220 T2=T4 0.001 0.001 0.001 0.001 0.675 0.935 0.887 0.808 T3=T4 0.192 0.228 0.202 0.198 0.180 0.263 0.266 0.328 Notes: The first block shows the results for the joint treatments. The second block for each treatment individually. Each row shows the regression coefficients and the standard error in parenthesis corresponding to an OLS regression. Dependent variables take the value 0-1 Standard errors are robust. *** p<0.01, ** p<0.05, * p<0.1. Controls include: sex, age, education, exposed to Covid, death to Covid, older than 65 at home, knows infected H1N1, belief about infection probability, belief about hospitalization probability, attends party, visits family, risk inside evaluation, and others practice social distancing. Source: Authors’ calculations 21