Revista Española de Investigaciones Sociológicas REIS. Rev. Esp. Investig. Sociol undefined https://doi.org/10.5477/cis/reis 2026 0210-5233 1988-5903 Centro de Investigaciones Sociológicas España https://reis.cis.es To Enter or Not to Enter: Compliance with COVID-19-related Measures in the Spanish Bar-hopping Culture Salir de bares durante la COVID-19 en España: las metas individualesy el cumplimiento con las restricciones Marta Fraile and Salvador Parrado

Key wordsAttitudesCOVID-19Social Norm ComplianceConjoint ExperimentGoal Framing Theory

AbstractThis article examines how citizens prioritize different motivations for complying with anti-COVID measures, using Goal-Framing Theory (GFT), which posits a hierarchy of normative, gain, and hedonic goals. Using a conjoint experiment embedded in an online survey of a representative sample of 3291 Spanish respondents, the study analyzes people s preferences regarding entering a bar. Fieldwork was carried out during the Omicron wave of the virus in February 2022. The results indicate that the likelihood of entering a bar increases when the activity takes place outdoors, patrons wear masks when not consuming alcohol, and most individuals have received a booster dose. Fear of contracting the virus (risk perception) emerges as the main factor influencing the decision to enter a bar.

Palabras claveActitudes frente a la COVID‑19COVID‑19Cumplimiento de normas socialesExperimento conjointTeoría del encuadre de metas

ResumenEste artículo analiza cómo la ciudadanía prioriza distintas motivaciones para cumplir con las medidas anti‑COVID. Utilizando un experimento conjunto (conjoint experiment en inglés) incluido en una encuesta en línea representativa de la población española (N = 3291), examinamos las preferencias de las personas al serles propuesto el ejercicio de decidir si entrar en un bar. El trabajo de campo se realizó durante la ola de la variante ómicron (en febrero de 2022). Los resultados muestran que la probabilidad de entrar a un bar aumenta cuando el consumo se realiza al aire libre, cuando los clientes llevan mascarilla mientras no consumen y cuando la mayoría de la población ha recibido la dosis de refuerzo. De todas las motivaciones sugeridas, la más relevante es el miedo a contraer el virus.
Citation Fraile, Marta; Parrado, Salvador (2026). «To Enter or Not to Enter: Compliance with COVID-19-related Measures in the Spanish Bar-hopping Culture». Revista Española de Investigaciones Sociológicas, undefined:27-44.(doi:10.5477/cis/reis.27-44) Marta Fraile: IPP-CSIC | marta.fraile@csic.es Salvador Parrado: UNED | sparrado@poli.uned.es doi:10.5477/cis/reis.195. - However, survey-based studies have several limitations. According to Noone et al. (2021), in their review of 84 studies analyzing the determinants of compliance with social distancing measures during the COVID-19 pandemic, most articles lacked a well-developed theoretical framework, were subject to biases from non-representative samples or failed to identify specific causal mechanisms. Perhaps most importantly, these studies rely on self-reported behavior, which is often affected by social desirability bias due to social pressure and prevailing norms of compliance (Daoust et al., 2021; Hansen et al., 2022). Experimental designs are rare in this line of research, with only a few exceptions (Amat et al., 2020; Goldstein and Wiedemann, 2021; Hamidi and Zandiatashbar, 2021). These studies employ relatively simple designs, limiting comparisons between control and treatment groups to a single dimension. As a result, they cannot simultaneous analyze explanatory factors influencing compliance. This study contributes to the research in three ways. First, it employs a conjoint experimental design embedded in an online survey. Conjoint experiments permit the simultaneous estimation of multiple factors by presenting respondents with hypothetical choice scenarios between two options (Bansak et al., 2021). Second, the analytical design draws on social contagion theory from legal sociology to examine how individuals adhere to social norms based on the behavior of others (Bicchieri et al., 2020; Scherer and Cho, 2003). Finally, the study incorporates Goal-Framing Theory (Lindenberg and Steg, 2007) to include an additional set of relevant factors, such as hedonic, gain and norm-related goals. Using evidence from a nationally representative online survey designed by the authors, with fieldwork conducted during the Omicron wave in February 2022, the results suggest that people s preferences are influenced by both social considerations and risk perception. Fear of contracting the virus emerges as the primary motivation for entering (or avoiding) a bar when the setting appears unsafe. Profit-seeking, based on a cost benefit analysis, outweighs hedonistic motives and contributes to compliance with regulations. Finally, although regulations and restrictions varied across regional governments, the degree of compliance and citizens’ preferences were similarly influenced by social and risk-related considerations across all regions, except for mask-wearing when not consuming, which was more important in Galicia and less so in Asturias. The implications of these findings for the literature on norm compliance are discussed in the conclusions section of these findings. Introduction During the COVID-19 pandemic, citizens were required to comply with restrictive rules that limited their mobility and everyday activities. Adherence to these regulations required changes in daily routines and constrained the freedom to organize everyday life in many countries. In this context, a line of research emerged to analyze variations in citizens behavior and the reasons why people comply with measures adopted by governments in response to the pandemic. Building on this literature, this study addresses two questions: How did citizens prioritize their motivations for complying with COVID-19 regulations? Is there meaningful variation across regions? Previous studies have analyzed citizens propensity to comply with rules during the COVID-19 pandemic, such as social distancing, mask use, or mobility restrictions, using survey-based observational evidence from or multiple countries (Daoust et al., 2021; Kooistra et al., 2020; Harper et al., 2021; Clark et al., 2020; Reinders et al., 2020). However, survey-based studies have several limitations. According to Noone et al. (2021), in their review of 84 studies analyzing the determinants of compliance with social distancing measures during the COVID-19 pandemic, most articles lacked a well-developed theoretical framework, were subject to biases from non-representative samples or failed to identify specific causal mechanisms. Perhaps most importantly, these studies rely on self-reported behavior, which is often affected by social desirability bias due to social pressure and prevailing norms of compliance (Daoust et al., 2021; Hansen et al., 2022). Coleman (1994: 244 ff.) follows a different line of reasoning in his theory of social systems analysis. According to this author, norms are formed at the social (macro) level, even though they originate from a collection of individual (micro) actions aimed at achieving specific goals. For a norm to fully develop, a transition from the micro to the macro level is necessary. When individuals initial actions, driven by personal goals, produce more disadvantages than benefits due to credible sanctioning mechanisms, a social norm begins to form. This individualistic perspective is grounded in a cost benefit analysis. Our approach to understanding how citizens prioritize compliance with COVID-19 pandemic norms focuses on the goals guiding their decision-making. We draw on Goal-Framing Theory (Lindenberg, 2017; Lindenberg et al., 2013, 2021), which posits that individuals act according to the goals that are most salient to them. According to Lindenberg (2017), three main goals compete for dominance: hedonic, gain, and normative. The hedonic goal seeks to maximize feeling good in the moment , a self-centered focus that may distance individuals from broader social considerations. The gain goal involves a cost benefit calculation, in which compliance with norms depends on whether the potential sanctions or risks outweigh the benefits of non-compliance. We argue that risk aversion, and the fear of losses more than the desire for gains (Kahneman and Tversky, 1982; Tversky and Kahneman, 1992), can override hedonic goals and align behavior with normative principles. This argument is supported by the meta-analysis of Floyd et al. (2000), which applies Protection Motivation Theory. Their findings suggest that perceived threat severity and individuals vulnerability increase adaptive behaviors, including compliance with norms. Finally, the normative goal refers to acting appropriately , which aligns with the social-norms framework proposed by Bicchieri (2005). Acting appropriately involves beliefs about what people typically do and what is considered acceptable within a given social context. Using evidence from a nationally representative online survey designed by the authors, with fieldwork conducted during the Omicron wave in February 2022, the results suggest that people s preferences are influenced by both social considerations and risk perception. Fear of contracting the virus emerges as the primary motivation for entering (or avoiding) a bar when the setting appears unsafe. Profit-seeking, based on a cost benefit analysis, outweighs hedonistic motives and contributes to compliance with regulations. Finally, although regulations and restrictions varied across regional governments, the degree of compliance and citizens’ preferences were similarly influenced by social and risk-related considerations across all regions, except for mask-wearing when not consuming, which was more important in Galicia and less so in Asturias. The implications of these findings for the literature on norm compliance are discussed in the conclusions section of these findings. Conceptual framework This study builds on the concept of a social norm, understood as a shared standard of behavior accepted within society (Coleman, 1994; Burke and Young, 2011). Social norms are informal, generally unwritten rules that regulate, guide, and condition individuals actions to facilitate coexistence and the regular, harmonious functioning of the communities in which they live. According to Bicchieri et al. (2020: 6), a norm is a rule of conduct that prescribes or proscribes certain behavior for a specific group of people in a specific class of situations . Such a rule becomes a social norm when individuals believe that most people adhere to it and, as a result, consider it appropriate to follow. Both empirical and normative expectations influence whether others comply with the norm. Adherence to rules is also related to social contagion. Social contagion theory has been applied to the adoption of innovative practices by physicians (Burt, 1987), the diffusion of risk perceptions within communities (Scherer and Cho, 2003), and compliance with social norms (Bicchieri et al., 2020). Coleman (1994: 244 ff.) follows a different line of reasoning in his theory of social systems analysis. According to this author, norms are formed at the social (macro) level, even though they originate from a collection of individual (micro) actions aimed at achieving specific goals. For a norm to fully develop, a transition from the micro to the macro level is necessary. When individuals initial actions, driven by personal goals, produce more disadvantages than benefits due to credible sanctioning mechanisms, a social norm begins to form. This individualistic perspective is grounded in a cost benefit analysis. In the context of compliance, acting appropriately or adhering to the normative goal is particularly relevant, as it relates to the established norms governing social life. However, since the normative goal typically ranks below hedonic and gain goals in individuals goal hierarchy, it requires substantial support to remain salient (Six et al., 2021). Compliance becomes unstable when gain or hedonic goals prevail (Lindenberg et al., 2021). Following the principles of social contagion theory within the framework of Goal-Framing Theory (GFT), we argue that non-compliance with norms reflects the dominance of hedonic goals (e.g., avoiding the discomfort of wearing masks, attending crowded events, or visiting friends and family without restrictions). Risk aversion and fear of sanctions can be understood as expressions of gain goals, albeit framed negatively. Finally, normative goals involve doing what is appropriate for oneself and others, highlighting the social dimension of behavior in which sociability plays a central role. The following hypotheses have been derived from our conceptual framework: H1. Citizens are more likely to comply with COVID-19 norms when they perceive that others are also complying. H2. Citizens are more likely to comply with COVID-19 norms when they recognize the costs associated with non-compliance. H3. Citizens are more likely to comply with COVID-19 norms when they fear contracting the virus. A key issue regarding the centrality of normative goals concerns the legitimacy of rules. From this perspective, rules become social norms when individuals accept the authority that establishes them. However, governments may adopt varying regulatory approaches. Pierson (1993) proposed the concept of policy feedback, emphasizing the role of policies in shaping the behavior of both political elites and citizens. Subsequent studies have shown that policies influence citizens behavior (Mettler and Soss, 2004) and can have a transformative effect by highlighting the benefits of civic conduct (Soss and Schram, 2007). Nonetheless, governments often fail to elicit the desired behavior among citizens (Mettler, 2019). Building on the policy feedback framework, we argue that different policies generate distinct compliance preferences. Pierson (2000), when examining how supply-side dynamics shape demand-side outcomes, argues that institutional structures and existing policies produce mutually reinforcing feedback effects. Negative feedback suggests the opposite effect. Wlezien (1995: 981) introduces the concept of thermostatic preferences and argues that: Lindenberg argues that, at any given moment, one of these three goals becomes most salient for individuals. While individuals often prioritize hedonic (immediate pleasure) goals or pursue gain goals (based on cost benefit considerations), the social environment, shaped by the presence and actions of others, can serve as a normative guide for behavior. Thus, social norms may influence individuals to act beyond the pursuit of pleasure or personal gain (Lindenberg and Steg, 2007). The social environment is shaped through a process in which individuals develop behaviors, form expectations about others behavior and receive signals about the group’s predominant goal (Lindenberg, 2017; Bicchieri, 2005). A shared goal provides the basis for collective behavior, that is, adherence to social norms. In this sense, collective normative behavior can foster individual commitment to norms, overriding hedonic interests and the pursuit of personal gain. In the context of compliance, acting appropriately or adhering to the normative goal is particularly relevant, as it relates to the established norms governing social life. However, since the normative goal typically ranks below hedonic and gain goals in individuals goal hierarchy, it requires substantial support to remain salient (Six et al., 2021). Compliance becomes unstable when gain or hedonic goals prevail (Lindenberg et al., 2021). Following the principles of social contagion theory within the framework of Goal-Framing Theory (GFT), we argue that non-compliance with norms reflects the dominance of hedonic goals (e.g., avoiding the discomfort of wearing masks, attending crowded events, or visiting friends and family without restrictions). Risk aversion and fear of sanctions can be understood as expressions of gain goals, albeit framed negatively. Finally, normative goals involve doing what is appropriate for oneself and others, highlighting the social dimension of behavior in which sociability plays a central role. Spain was severely affected by the novel coronavirus during the first wave of the pandemic (March-May 2020). From a health perspective, the country experienced a high number of infections, hospitalizations, and deaths, resulting in one of the highest mortality rates in Europe. Economically, Spain was highly vulnerable due to the high share of the tourism and hospitality sector in the Gross Domestic Product (GDP). Consequently, the political debate over measures to balance public health and economic activity in these sectors quickly gained prominence in both the media and public opinion (Fraile and Méndez, 2021). After the first period of the state of emergency (March 14 to June 21, 2020), declared by the national government (Royal Decree 463/2020) across the entire territory, the decentralized nature of Spain s quasi-federal system returned to the forefront. Within the framework of minimum common-denominator agreements of the Intergovernmental Health Council, each regional government designed and implemented its own measures to contain the spread of COVID-19. Regional authorities had discretion over hours of operation, maximum allowed capacity in bars and restaurants, limits on the number of people allowed to gather in public places, and similar measures. In addition, regional and local authorities were responsible for enforcement. In parallel, regional and local authorities implemented economic measures to mitigate the pandemic s impact and encourage compliance with COVID-19 rules. In addition to grants and tax exemptions, the hospitality sector benefited from the tax-free use of public streets and sidewalks to set up tables and chairs, expanding business opportunities and offsetting occupancy restrictions in both indoor and outdoor areas. Despite the limited discretion granted to the regions, their responses varied, producing additional effects on citizens civic behavior, social interactions, and rule compliance. We adopt a comparative interregional design to examine whether meaningful differences exist in citizens’ preferences. When the actual temperature of policy differs from the preferred temperature, the public sends a signal to adjust the policy accordingly, and once sufficiently adjusted, the signal ceases. In the context of COVID-19 restrictions, we incorporate the role of government into the policy feedback cycle and examine its impact on compliance (Moynihan and Soss, 2014). Lindenberg (2017) further asserts that institutional systems significantly shape how overarching goals (preferably prosocial and normative) are established for individuals and how they behave to maintain appropriate conduct. Consistent with the positive feedback effect, we expect that governments with more permissive anti-COVID regimes—characterized by less stringent enforcement—will elicit a hedonistic response from individuals, with hedonic considerations taking precedence over normative ones. Conversely, in regimes with stricter policies and enforcement, normative goals are expected to dominate over hedonic and gain goals. This reasoning leads to our fourth hypothesis: H4. The collective salience of each goal differs across regional contexts. Research Design In this section, we first explain why Spain provides an appropriate context for testing our theoretical expectations regarding citizens preferences for anti-COVID-19 measures. This is followed by a subsection summarizing the experimental design and data. Spain: a laboratory for analyzing citizens preferences for complying with anti-COVID-19 measures Spain was severely affected by the novel coronavirus during the first wave of the pandemic (March-May 2020). From a health perspective, the country experienced a high number of infections, hospitalizations, and deaths, resulting in one of the highest mortality rates in Europe. Economically, Spain was highly vulnerable due to the high share of the tourism and hospitality sector in the Gross Domestic Product (GDP). Consequently, the political debate over measures to balance public health and economic activity in these sectors quickly gained prominence in both the media and public opinion (Fraile and Méndez, 2021). TABLe 1. Case selection

Autonomous Community

Political orientation

Type of government

Degree of stringency

Asturias

Left-wing

Single-party (Partido Socialista Obrero Español [PSOE])

High

Galicia

Right-wing

Single-party (Partido Popular [PP])

High

Madrid

Right-wing

Coalition (PP/Ciudadanos)

Low

Valencia

Left-wing

Coalition (PSOE/Compromís/Unidas Podemos [UP])

Medium
Source: Author s own creation. In addition to conducting a representative survey of the Spanish population, we focused on four autonomous communities (Asturias, Galicia, Madrid and Valencia) by oversampling respondents. These regions exhibited varying levels of stringency in the measures adopted and their enforcement (see Table 1). This variation was confirmed through interviews with local and regional authorities responsible for COVID-19 management during the second wave of the pandemic (Parrado et al., 2025). The Madrid region adopted more permissive measures and frequently deviated from those agreed upon by the intergovernmental health body established to manage the pandemic. Galicia and Asturias, by contrast, implemented relatively strict measures, often justified by sociodemographic factors, such as Galicia s relatively older population, and by alignment with the national government’s policy preferences (from the same political party in Asturias). Valencia also implemented strict measures, although interviewees noted that these were significantly relaxed to support the economy during the 2021-22 Christmas period, just prior to our survey’s fieldwork (Parrado et al., 2025). These four regions also differ in terms of government composition and ideological orientation. Following the policy feedback framework (Pierson, 2000), citizens in Madrid are expected to exhibit more hedonistic preferences. In regions with stricter policies, such as Asturias and Galicia, these motivations are expected to yield to normative goals. The survey used for the conjoint exercise required approximately 10 minutes to complete and was administered to a nationally representative sample of Spanish residents aged 18 and older (N = 3291). To examine whether regional variations in the stringency of COVID-19 measures are associated with differences in citizens preferences, the sample was also made representative for the four regions of interest: Asturias, Comunidad Valenciana, Galicia, and Madrid. The survey was administered through an opt-in access panel operated by Netquest, which compensated participants with vouchers redeemable for goods at Netquest’s online store. Fieldwork was conducted between February 11 and March 10, 2022. In addition to conjoint analysis, the survey collected information on citizens’ opinions, behaviors, and attitudes regarding anti-COVID-19 measures implemented in Spain and in their respective regions. The wording of the conjoint experiment is provided in the Appendix . Despite the limited discretion granted to the regions, their responses varied, producing additional effects on citizens civic behavior, social interactions, and rule compliance. We adopt a comparative interregional design to examine whether meaningful differences exist in citizens’ preferences. In addition to conducting a representative survey of the Spanish population, we focused on four autonomous communities (Asturias, Galicia, Madrid and Valencia) by oversampling respondents. These regions exhibited varying levels of stringency in the measures adopted and their enforcement (see Table 1). This variation was confirmed through interviews with local and regional authorities responsible for COVID-19 management during the second wave of the pandemic (Parrado et al., 2025). The Madrid region adopted more permissive measures and frequently deviated from those agreed upon by the intergovernmental health body established to manage the pandemic. Galicia and Asturias, by contrast, implemented relatively strict measures, often justified by sociodemographic factors, such as Galicia s relatively older population, and by alignment with the national government’s policy preferences (from the same political party in Asturias). Valencia also implemented strict measures, although interviewees noted that these were significantly relaxed to support the economy during the 2021-22 Christmas period, just prior to our survey’s fieldwork (Parrado et al., 2025). These four regions also differ in terms of government composition and ideological orientation. Following the policy feedback framework (Pierson, 2000), citizens in Madrid are expected to exhibit more hedonistic preferences. In regions with stricter policies, such as Asturias and Galicia, these motivations are expected to yield to normative goals. Conjoint experiments have gained popularity among social scientists because they allow the analysis of multidimensional preferences and the causal effects of different attributes on hypothetical choices (Bansak et al., 2021; Ganter, 2023; Hainmueller et al., 2014). This method is particularly suitable for assessing GFT as it can estimate the relative importance of specific values and attributes that citizens prioritize. As noted earlier, the normative goal in GFT competes with hedonic and gain goals for preeminence at the group level. This dominance of one goal over the others varies across contexts, reinforcing individual behavior and the salience of that goal. The conjoint experiment presented respondents with a scenario in which they could choose to enter a bar, with two alternatives described by four attributes. Each attribute could take one of two values, randomly assigned in each profile. Random assignment of attribute values allows for the causal identification of how each attribute and its value influence respondents decision to consume at the bar or restaurant. Table 2 summarizes the two possible values of each of the four attributes . The attributes and their values were chosen so that all possible combinations are logically possible and equally realistic . Following Hainmueller et al. (2014), this approach avoids excluding any combination that could occur in diverse hypothetical situations. Each respondent was asked to choose their preferred alternative across four rounds, yielding a total of 26 328 observations. By iteratively pairing the situations and analyzing the rankings selected by each respondent, we aimed to identify the most salient overarching goals and their alignment with the relevant norm. Each attribute in the conjoint experiment was linked to a specific overarching goal of GFT. Adapting Goal Framework Theory for the conjoint experiment The conjoint experiment, embedded in a web-based platform, simulated choices related to the hospitality industry (bars, pubs, and restaurants), which are common gathering places for Spaniards to meet friends and family. Notably, Spain has ranked first in Europe in the number of such establishments (among 26 EU countries) since 2018 . Consequently, all respondents could realistically encounter this type of situation, making it a relevant contextual factor. We adapted the tenets of the GFT to our conjoint experiment, which focused on decisions about entering a bar or restaurant following the fourth wave of COVID-19 (Omicron variant). During this period, entry to bars in most Spanish regions was permitted, albeit with restrictions. Key rules included wearing a mask at all times in bars except when drinking or eating, maintaining a social distance of 1.5 meters between chairs (rather than tables), and limiting maximum capacity based on local health risk levels. Some of these restrictions were adapted for a hypothetical scenario in the conjoint experiment. Conjoint experiments have gained popularity among social scientists because they allow the analysis of multidimensional preferences and the causal effects of different attributes on hypothetical choices (Bansak et al., 2021; Ganter, 2023; Hainmueller et al., 2014). This method is particularly suitable for assessing GFT as it can estimate the relative importance of specific values and attributes that citizens prioritize. As noted earlier, the normative goal in GFT competes with hedonic and gain goals for preeminence at the group level. This dominance of one goal over the others varies across contexts, reinforcing individual behavior and the salience of that goal. TABLe 2. Attributes and their values

Attribute

Values

Mask-wearing

Everyone wears a mask when not drinking.

Almost no one wears a mask.

Law enforcement

Police issue fines to those who do not comply.

Police do not show up to issue fines.

Area of the bar/restaurant where consumption takes place

Outdoors.

Indoors.

Booster vaccination coverage

35 % of the population.

70 % of the population.
Source: Closa et al. (2025), author s own creation. The third and fourth scenarios relate to risk perception. For the third attribute, respondents could choose to consume either indoors or outdoors in winter (February 2022). Given widespread public messaging about the risk of contracting COVID-19 indoors, this attribute captures the trade-off between the hedonic motivation to drink in the comfort of a bar and the preference for drinking outdoors. The fourth attribute provides information on vaccination status, specifically the coverage of a booster dose (typically the third dose). In one scenario, 35 % of the population had received the booster vaccination, while in the other, 70 % had. Risk perception in both cases relates to the gain goal (i.e., a cost-benefit evaluation) but it operates through a distinct mechanism. Here, compliance is not the central issue. Rather, what matters is the extent to which an individual s overriding goal is shaped by the perception of a risky situation, independent of the behavior of others. Risk perception refers to a subjective assessment of the health threat posed to individuals (Neuburger and Egger, 2021), which may lead them to forgo the gain goal to avoid potential losses. We included control variables that prior research has shown to influence compliance with anti-COVID-19 measures, including gender (Galasso et al., 2020; Mazza and Scipioni, 2022), age (Oude Groeniger et al., 2021; Wang et al., 2021), education (Nivette et al., 2021), political orientation (Goldstein and Wiedemann, 2021; Harper et al., 2021; Wang et al., 2021), and policy feedback from different governments (Moynihan and Soss, 2014; Pierson, 1993). Each respondent was asked to choose their preferred alternative across four rounds, yielding a total of 26 328 observations. By iteratively pairing the situations and analyzing the rankings selected by each respondent, we aimed to identify the most salient overarching goals and their alignment with the relevant norm. Each attribute in the conjoint experiment was linked to a specific overarching goal of GFT. For the first attribute (mask-wearing; see Table 2), there are two alternatives: everyone complies and wears the mask when required (reflecting the normative goal) versus almost no one wears the mask (representing the hedonic escape). The second attribute (enforcement through sanctions) corresponds to the gain goal and the cost-benefit calculation of the probability of receiving a fine. This attribute is presented in two scenarios: in one, the police sanction non-compliers, which may discourage prioritization of the gain goal; in the other, the police do not enforce fines, allowing the hedonic goal to predominate. The third and fourth scenarios relate to risk perception. For the third attribute, respondents could choose to consume either indoors or outdoors in winter (February 2022). Given widespread public messaging about the risk of contracting COVID-19 indoors, this attribute captures the trade-off between the hedonic motivation to drink in the comfort of a bar and the preference for drinking outdoors. The fourth attribute provides information on vaccination status, specifically the coverage of a booster dose (typically the third dose). In one scenario, 35 % of the population had received the booster vaccination, while in the other, 70 % had. Risk perception in both cases relates to the gain goal (i.e., a cost-benefit evaluation) but it operates through a distinct mechanism. Here, compliance is not the central issue. Rather, what matters is the extent to which an individual s overriding goal is shaped by the perception of a risky situation, independent of the behavior of others. Risk perception refers to a subjective assessment of the health threat posed to individuals (Neuburger and Egger, 2021), which may lead them to forgo the gain goal to avoid potential losses. FIGURe 1.  Average Marginal Component Effects (AMCEs) on respondents decisions to enter a bar or restaurant for consumption  Source: Closa et al. (2025), author s own creation based on the results of Table A1. Figure 1 summarizes the main results of participants preferences regarding whether to enter a bar or restaurant to consume. The figure presents estimates from a linear probability model with standard errors clustered at the individual level, accounting for the fact that each respondent completed four pairs of comparisons. Table A1 in the Appendix provides the estimated coefficients, which include the four conjoint attributes as independent variables, along with several control variables measured at the individual level: gender (coded 1 for women and 0 for men), age (in years), education (an ordinal variable ranging from 0 for elementary school or less; 1 for secondary school; 2 for high school; 3 for university; to 4 for master’s/PhD), and ideology (a scale ranging from 0 for extreme left to 10 for extreme right). In addition, the model incorporates fixed effects for region of residence (Asturias, Valencia, Galicia and Madrid versus the rest of Spain). These controls were included to assess whether systematic differences exist in attribute preferences across these characteristics. We also replicated the analysis without control variables, and the results remained robust, suggesting that the randomization process was effective. Figure 1 summarizes the average marginal component effects (AMCEs) of the four attributes (estimates based on the coefficients in Table A2) and indicates that citizens preferences are influenced by social and risk-related considerations. Among the four attributes examined, mask-wearing when not drinking or eating, and consuming outdoors emerge as the most relevant dimensions in explaining the preference to enter a venue. Specifically, the probability of deciding to enter a bar or restaurant increases by 23 percentage points when consumption takes place outdoors (compared to an indoor setting). Conversely, the likelihood decreases by 21 percentage points when individuals are not wearing masks when not drinking (compared with venues where masks are always worn). Results Figure 1 summarizes the main results of participants preferences regarding whether to enter a bar or restaurant to consume. The figure presents estimates from a linear probability model with standard errors clustered at the individual level, accounting for the fact that each respondent completed four pairs of comparisons. Table A1 in the Appendix provides the estimated coefficients, which include the four conjoint attributes as independent variables, along with several control variables measured at the individual level: gender (coded 1 for women and 0 for men), age (in years), education (an ordinal variable ranging from 0 for elementary school or less; 1 for secondary school; 2 for high school; 3 for university; to 4 for master’s/PhD), and ideology (a scale ranging from 0 for extreme left to 10 for extreme right). In addition, the model incorporates fixed effects for region of residence (Asturias, Valencia, Galicia and Madrid versus the rest of Spain). These controls were included to assess whether systematic differences exist in attribute preferences across these characteristics. We also replicated the analysis without control variables, and the results remained robust, suggesting that the randomization process was effective. FIGURe 2.  Average Marginal Component Effects (AMCEs) on respondents decisions to enter a bar or restaurant for consumption, by region in Spain  Source: Closa et al. (2025), author s own creation based on the results of Table A1. These findings suggest that citizens override their hedonic goals when they perceive that others are not adhering to the norm or when the situation poses a health risk. They prioritize the gain goal (minimizing risk) and the normative motivation to act appropriately in accordance with social norms. We hypothesized that the relevance of each attribute might vary by region of residence, reflecting a policy feedback effect. As mentioned above, citizens may adjust their preferences based on the types of policies implemented in their region. For example, we would expect citizens in Madrid to prioritize hedonic motives over social or risk-related considerations. To test this, we replicates the estimation presented in Figure 1, adding a set of interaction terms between each of the four attributes and region of residence (relative to the rest of Spain). Figure 2 presents the main results. Figure 1 summarizes the average marginal component effects (AMCEs) of the four attributes (estimates based on the coefficients in Table A2) and indicates that citizens preferences are influenced by social and risk-related considerations. Among the four attributes examined, mask-wearing when not drinking or eating, and consuming outdoors emerge as the most relevant dimensions in explaining the preference to enter a venue. Specifically, the probability of deciding to enter a bar or restaurant increases by 23 percentage points when consumption takes place outdoors (compared to an indoor setting). Conversely, the likelihood decreases by 21 percentage points when individuals are not wearing masks when not drinking (compared with venues where masks are always worn). The likelihood of entering the bar or restaurant also increases by 10 percentage points when respondents perceive that other patrons are largely protected (70 % of the population vaccinated) compared with a scenario in which only a minority (30 %) is vaccinated. Finally, the perceived risk of being sanctioned for noncompliance appears largely irrelevant, as the probability of consumption only decreases by approximately five percentage points when no enforcement is present. To explore how individuals prioritize their goals in the context of COVID-19 norms, we designed a conjoint experiment in which multiple attributes were combined to create two alternatives. One alternative reflects the hedonic/gain pathway and the other reflects the normative/prosocial one. Participants were presented with two hypothetical scenarios involving the possibility of entering a bar during the peak of the Omicron wave. Each scenario varied across key dimensions including mask use, higher and lower-risk contagion conditions (sitting indoors or outdoors), booster vaccination rates, and the potential for sanctions as a mechanism to enforce norm compliance. Among the motivations examined, fear of contracting the virus appears to carry the greatest weight in shaping the decision to enter (or avoid) a bar when the environment is perceived as unsafe, consistent with prior research. For instance, a large-scale study by Liberoth et al. (2021), surveying over 173 000 individuals across 48 countries, found that those who feared contracting a disease were more likely to adopt protective measures. Other studies have also highlighted the importance of perceived disease risk as a key determinant of behavioral change (Dryhurst et al., 2020; Harper et al., 2021; Webster et al., 2020). Risk perception is shaped by factors such as knowledge, visibility, trust, and voluntariness (Slovic, 1992). Limited knowledge about an issue or phenomenon tends to increase perceived risk. For example, unfamiliar diseases or pandemics with unpredictable effects are especially frightening because individual responses to the threat are uncertain. Low visibility of a hazard also heightens perceived risk. Therefore, it is not surprising that COVID-19 risk has been a key factor in motivating individuals to modify their behavior (e.g., entering a venue knowing that most attendees are vaccinated or preferring outdoor to indoor settings). According to social contagion theory (Scherer and Cho, 2003), risk perception can prompt similar behaviors among people in the same context. Consequently, individuals cost-benefit evaluations may influence others to comply with norms. Contrary to our expectations, however, Figure 2 indicates that citizens preferences are influenced similarly by social and risk-related considerations across all regions examined, with the exception of mask-wearing when not consuming. The coefficient for this attribute is higher in Galicia (25 percentage points) and lower in Asturias (18 percentage points), with no overlap in the confidence intervals. We also examined potential sources of heterogeneity in the effects of the four main attributes at the individual level, corresponding to our control variables (gender, education, age, and ideology). We found no meaningful differences between men and women, young and old, across education levels, or among individuals with different ideologies. Discussion of results and conclusions Building on social contagion theory, which emphasizes the importance of social norms in shaping behavior, we adapt Goal-Framing Theory (GFT) to a context (entering a bar) where sociability is in tension with COVID-19 regulations. According to GFT, the normative goal represents the fundamental motivation to preserve the common good, driven by a sense of obligation to act appropriately when others comply and when an established social norm exists (Bicchieri, 2005). To explore how individuals prioritize their goals in the context of COVID-19 norms, we designed a conjoint experiment in which multiple attributes were combined to create two alternatives. One alternative reflects the hedonic/gain pathway and the other reflects the normative/prosocial one. Participants were presented with two hypothetical scenarios involving the possibility of entering a bar during the peak of the Omicron wave. Each scenario varied across key dimensions including mask use, higher and lower-risk contagion conditions (sitting indoors or outdoors), booster vaccination rates, and the potential for sanctions as a mechanism to enforce norm compliance. Finally, and contrary to our expectations, the stricter policies implemented in some regions (Galicia and Asturias) did not significantly influence respondents decisions to select a more restrictive scenario. Prior studies have shown that the severity of measures can increase trust in government and encourage compliance with COVID-19 regulations (Chong et al., 2020). In our study, the hypothetical nature of the scenario (an experiment embedded within a survey) likely diminished the relevance of participants’ regional contexts. We conclude by acknowledging several limitations of this study. First, our scenario for testing GFT was conducted at a specific point in time and was limited to a single social situation. While this context is relevant, as it highlights the tension between hedonic motivation and compliance with norms, it introduces an inherent self-selection bias: many individuals avoided bars during the pandemic because they perceived them as unsafe. Unfortunately, we did not ask respondents about their likelihood of entering a bar in either scenario (the preferred option or the non-chosen one). Moreover, we cannot determine whether the results would have differed in a subsequent wave of the pandemic. Ideally, a panel study spanning the entire pandemic would have allowed us to control for temporal variations in severity and better understand how motivations evolved over time. Finally, our study does not address involuntary situations, such as commuting to work using public transportation. Second, although we attempted to identify different attributes for each scenario, this was not always possible. For example, we could not determine whether mask-wearing, or the absence of it reflected social pressure or a risk-related consideration. When conjoint experiments move beyond the domain such as consumer goods, common in marketing research, or the choice of a presidential candidate, typical in political science studies, constructing clearly contrasting scenarios becomes more challenging. Risk perception, however, represents a gain-related motivation that is not associated with sanctions. Consistent with our results, the presence of sanctions did not significantly influence compliance. This suggests that Coleman s (1994) argument, that social norms emerge when the costs of micro-level, goal-oriented individual actions (due to sanctions) outweigh the benefits, does not fully apply to the Spanish context studied here. The concept of prosociality (Bicchieri et al., 2020) is closely linked to compliance, and widespread mask use ranked highly in individual preferences, consistent with previous studies (Oosterhoff and Palmer, 2020). Lindenberg et al. (2021) emphasize that mask use requires sustained support over time. Given that the fieldwork for this study was conducted during the Omicron variant wave, nearly two years into the global COVID-19 pandemic, it is reasonable to assume that mask use had already become largely normalized, despite the fatigue associated with prolonged compliance. According to GFT, overarching goals shape how individuals process and respond to information (Lindenberg and Steg, 2007). Hedonic, gain, and normative goals may be activated simultaneously, but they are not always compatible. COVID-19 norms, such as social distancing and mask use, may conflict with the hedonic pleasure of socializing with friends and family in a bar. However, a disruptive factor, such as the pandemic outbreak, that shifts the gain-goal calculus and creates significant losses may tip the balance, aligning normative and gain goals while relegating hedonic motivations to the background. In this context, the government did not need to exert substantial effort to reconcile gain and normative goals, as GFT proponents might suggest, because the virus s impact was widely visible. Nevertheless, GFT remains useful in ranking and understanding different motivations when designing policies. Clark, Cory; Davila, Andrés; Regis, Maxime and Kraus, Sascha (2020). Predictors of COVID-19 Voluntary Compliance Behaviors: An International Investigation . Global Transitions, 2: 76-82. Closa, Carlos; Fraile, Marta; Parrado, Salvador and Pereira-Puga, Manuel (2022). Los españoles ante las medidas y recomendaciones contra la COVID . DIGITAL CSIC. Available at: https://digital.csic.es/handle/10261/275830, access January 26, 2026. Closa, Carlos; Fraile, Marta; Parrado, Salvador and Pereira-Puga, Manuel (2025). Designing compliance for COVID-19 Public Policies [DATASET] [Data set] . DIGITAL.CSIC. doi: 10.20350/DIGITALCSIC/17767 Coleman, James S. (1994). Foundations of social theory. Harvard University Press. Daoust, Jean-François; Nadeau, Richard; Dassonneville, Ruth; Lachapelle, Erick; Bélanger, Éric; Savoie, Justin and Linden, Clifton van der (2021). How to Survey Citizens Compliance with COVID-19 Public Health Measures: Evidence from Three Survey Experiments . Journal of Experimental Political Science, 8(3): 310-317. doi: 10.1017/XPS.2020.25 Dryhurst, Sarah; Schneider, Claudia R.; Kerr, John; Freeman, Alexandra L. J.; Recchia, Gabriel; Bles, Anne M. van der; Spiegelhalter, David and Linden, Sander van der (2020). Risk Perceptions of COVID-19 around the World . Journal of Risk Research, 23(7-8): 994-1006. doi: 10.1080/13669877.2020.1758193 Floyd, Donna L.; Prentice-Dunn, Steven and Rogers, Ronald W. (2000). A Meta-Analysis of Research on Protection Motivation Theory . Journal of Applied Social Psychology, 30(2): 407-429. doi: 10.1111/j.1559-1816.2000.tb02323.x Fraile, Marta and Méndez, Mónica (2021). La opinión pública durante la pandemia: ¿Más de lo mismo? . Panorama Social, 33: 177-192. Galasso, Vincenzo; Pons, Vincent; Profeta, Paola; Becher, Michael; Brouard, Sylvain and Foucault, Martial (2020). Gender Differences in COVID-19 Related Attitudes y Behavior: Evidence from a Panel Survey in Eight OECD Countries . National Bureau of Economic Research, 27359: w27359. doi: 10.3386/w27359 Ganter, Flavien (2023). Identification of Preferences in Forced-Choice Conjoint Experiments: Reassessing the Quantity of Interest . Political Analysis, 31(1): 98-112. Second, although we attempted to identify different attributes for each scenario, this was not always possible. For example, we could not determine whether mask-wearing, or the absence of it reflected social pressure or a risk-related consideration. When conjoint experiments move beyond the domain such as consumer goods, common in marketing research, or the choice of a presidential candidate, typical in political science studies, constructing clearly contrasting scenarios becomes more challenging. Finally, the external validity of our study may be limited to the Spanish context. While we do not claim that our findings generalize to other settings, we believe that this conjoint experimental design could be applied to other pandemic situations and to environments where bar socializing is common. Bibliography Amat, Francesc; Arenas, Andreu; Falcó-Gimeno, Albert and Muñoz, Jordi (2020). Pandemics meet democracy: Experimental evidence from the COVID-19 crisis in Spain. doi: 10.31235/osf.io/dkusw Bansak, Kirk; Hainmüller, Jens; Hopkins, Daniel J. and Yamamoto, Teppei (2021). Conjoint survey experiments. In: Druckman, James N. and Green, Donald P.  (eds.). Advances in experimental political science (vol. 19, pp. 19-41). New York: Cambridge University Press. Bicchieri, Cristina (2005). The grammar of society: The nature y dynamics of social norms. New York: Cambridge University Press. Bicchieri, Cristina; Dimant, Eugen; Gachter, Simon and Nosenzo, Daniele (2020). Observability, Social Proximity and the Erosion of Norm Compliance . SSRN Electronic Journal. doi: 10.2139/ssrn.3576289 Burke, Mary A. and Young, H. Peyton (2011). Social norms. In: Handbook of Social Economics (vol. 1, pp. 311-338). Elsevier. Burt, Ronald S. (1987). Social Contagion and Innovation: Cohesion versus Structural Equivalence . American Journal of Sociology, 92(6): 1287-1335. doi: 10.1086/228667 Chong, Yuen Y.; Chien, Wai T.; Cheng, Ho Y.; Chow, Ka; Kassianos, Angelo P.; Karekla, Maria and Gloster, Andrew (2020). The Role of Illness Perceptions, Coping, and Self-Efficacy on Adherence to Precautionary Measures for COVID-19 . International Journal of Environmental Research y Public Health, 17(18): 6540. doi: 10.3390/ijerph17186540 Lieberoth, Andreas; Lin, Shiang-Yi; Stöckli, Sabrina; Han, Hyemin; Kowal, Marta; Gelpi, Rebekah; Chrona, Stavroula; Tran, Thao P.; Jeftić, Alma; Rasmussen, Jasper; Cakal, Huseyin; Milfont, Taciano L.; Yamada, Yuki; Debove, Stephane; Gelpí, Rebekah; Flis, Ivan; Turc, Fidan; Sikka, Pilleriin; Mamede, Salomé; Tuominen, Jarno (2021). Stress and Worry in the 2020 Coronavirus Pandemic: Relationships to Trust and Compliance with Preventive Measures across 48 Countries in the COVIDiSTRESS Global Survey . Royal Society Open Science, 8(2): 200589. doi: 10.1098/rsos.200589 Lindenberg, Siegwart (2017). The Dependence of Human Cognitive and Motivational Processes on Institutional Systems. In: Jann, Ben and Przepiorka, Wojtek (eds.). Social dilemmas, enninstitutions and the evolution of cooperation (pp. 85-106). De Gruyter. doi: 10.1515/9783110472974-005 Lindenberg, Siegwart and Steg, Linda (2007). Normative, Gain and Hedonic Goal Frames Guiding Environmental Behavior . Journal of Social Issues, 63(1): 117-137. doi: 10.1111/j.1540-4560.2007.00499.x Lindenberg, Siegwart and Steg, Linda (2013). Goal-framing theory and norm-guided environmental behavior. In: Trijp, Hans C. M. van (ed.). Encouraging sustainable behavior (pp. 37 54). New York: Psychology Press. Lindenberg, Siegwart; Six, Frederique and Keizer, Kees (2021). Social contagion and goal framing: The sustainability of rule compliance. In: Sokol, Daniel D. and Rooij, Benjamin van (eds.). Cambridge Handbook of Compliance (pp. 422-437). Cambridge: Cambridge University Press. Mazza, Jacopo and Scipioni, Marco (2022). The Gender Gap in Support for Governments During the COVID Crisis . Journal of European Public Policy: 1-17. doi: 10.1080/13501763.2022.2027503 Mettler, Suzanne (2019). Making What Government Does Apparent to Citizens: Policy Feedback Effects, Their Limitations, and How They Might Be Facilitated . The ANNALS of the American Academy of Political y Social Science, 685(1): 30-46. doi: 10.1177/0002716219860108 Mettler, Suzanne and Soss, Joe (2004). The Consequences of Public Policy for Democratic Citizenship: Bridging Policy Studies and Mass Politics . Perspectives on Politics, 2(1): 55-73. doi: 10.1017/S1537592704000623 Moynihan, Donald P. and Soss, Joe (2014). Policy Feedback and the Politics of Administration . Public Administration Review, 74(3): 320-332. Neuburger, Larissa and Egger, Roman (2021). Travel Risk Perception and Travel Behaviour During the COVID-19 Pandemic 2020: A Case Study of the DACH Region . Current Issues in Tourism, 24(7): 1003-1016. Nivette, Amy; Ribeaud, Denis; Murray, Aja; Steinhoff, Annekatrin; Bechtiger, Laura; Hepp, Urs; Shanahan, Lilly and Eisner, Manuel (2021). Non-compliance with COVID-19-related Public Health Measures among Young Adults in Switzerland: Insights from a Longitudinal Cohort Study . Social Science y Medicine, 268: 113370. Goldstein, Daniel A. N. and Wiedemann, Johannes (2021). Trust Me, Mask Up: Experimental Evidence on Social Trust and Responsiveness to COVID-19 Mitigation Policies. Available at: https://ssrn.com/abstract=3835934 / doi: 10.2139/ssrn.3835934, access January 26, 2026. Hainmüller, Jens; Hopkins, Daniel J. and Yamamoto, Teppei (2014). Causal Inference in Conjoint Analysis: Understanding Multidimensional Choices via Stated Preference Experiments . Political Analysis, 22(1): 1-30. doi: 10.1093/pan/mpt024 Hamidi, Shima and Zandiatashbar, Ahoura (2021). Compact Development and Adherence to Stay-at-home Order During the COVID-19 Pandemic: A Longitudinal Investigation in the United States . Landscape y Urban Planning, 205: 103952. doi: 10.1016/j.landurbplan.2020.103952 Hansen, Pelle G.; Larsen, Erik G. and Gundersen, C. D. (2022). Reporting on One s Behavior: A Survey Experiment on the Nonvalidity of Self-reported COVID-19 Hygiene-relevant Routine Behaviors . Behavioural Public Policy, 6(1): 34-51. doi: 10.1017/bpp.2021.13 Harper, Craig A.; Satchell, Liam P.; Fido, Dean and Latzman, Robert D. (2021). Functional Fear Predicts Public Health Compliance in the COVID-19 Pandemic . International Journal of Mental Health y Addiction, 19(5): 1875-1888. doi: 10.1007/s11469-020-00281-5 Kahneman, Daniel and Tversky, Amos (1982). The Psychology of Preferences . Scientific American, 246(1): 160-173. Kooistra, Emmeke B.; Reinders, Chris; Kuiper, Malouke E.; Olthuis, Elke; Brownlee, Megan; Fine, Adam and Rooij, Benjamin van (2020). Mitigating COVID-19 in a Nationally Representative UK Sample: Personal Abilities and Obligation to Obey the Law Shape Compliance with Mitigation Measures . SSRN Electronic Journal. doi: 10.2139/ssrn.3598221 Lieberoth, Andreas; Lin, Shiang-Yi; Stöckli, Sabrina; Han, Hyemin; Kowal, Marta; Gelpi, Rebekah; Chrona, Stavroula; Tran, Thao P.; Jeftić, Alma; Rasmussen, Jasper; Cakal, Huseyin; Milfont, Taciano L.; Yamada, Yuki; Debove, Stephane; Gelpí, Rebekah; Flis, Ivan; Turc, Fidan; Sikka, Pilleriin; Mamede, Salomé; Tuominen, Jarno (2021). Stress and Worry in the 2020 Coronavirus Pandemic: Relationships to Trust and Compliance with Preventive Measures across 48 Countries in the COVIDiSTRESS Global Survey . Royal Society Open Science, 8(2): 200589. doi: 10.1098/rsos.200589 RECEPTION: December 20, 2024 REVIEW: May 19, 2025 ACCEPTANCE: October 16, 2025 Reinders, Chris; Kuiper, Malouke E.; Olthuis, Elke; Kooistra, Emmeke B.; Bruijn, Anne L. de; Brownlee, Megan; Fine, Adam and Rooij, Benjamin van (2020). Compliance in the 1.5 Meter Society: Longitudinal Analysis of Citizens Adherence to COVID-19 Mitigation Measures in a Representative Sample in the Netherlands . Amsterdam Law School Research Paper, 2020 33. Scherer, Clifford W. and Cho, Hichang (2003). A Social Network Contagion Theory of Risk Perception . Risk Analysis, 23(2): 261-267. doi: 10.1111/1539-6924.00306 Six, Frederique; Vadder, Steven; Glavina, Monika; Verhoest, Koen and Pepermans, Koen (2021). What Drives Compliance with COVID‐19 Measures over Time? Explaining Changing Impacts with Goal Framing Theory . Regulation y Governance, 17(1). doi: 10.1111/rego.12440 Slovic, Paul (1992). Perception of risk: Reflections on the psychometric paradigm. In: Golding, Dominic and Krimsky, Sheldon (eds.). Theories of Risk. New York: Praeger. Soss, Joe and Schram, Sanford F. (2007). A Public Transformed? Welfare Reform as Policy Feedback . American Political Science Review, 101(1): 111-127. doi: 10.1017/S0003055407070049 Tversky, Amos and Kahneman, Daniel (1992). Advances in prospect theory: Cumulative Representation of Uncertainty . Journal of Risk and Uncertainty, 5: 297-323. doi: 10.1007/BF00122574 Wang, Donna; Marmo-Roman, Suzanne; Krase, Kathryn and Phanord, Lusta (2021). Compliance with Preventative Measures During the COVID-19 Pandemic in the USA and Canada: Results from an Online Survey . Social Work in Health Care, 60(3): 240-255. doi: 10.1080/00981389.2020.1871157 Webster, Rebecca K.; Brooks, Samantha K.; Smith, Louise E.; Woodland, Lisa; Wessely, Simon and Rubin, G. James (2020). How to Improve Adherence with Quarantine: Rapid Review of the Evidence . Public Health, 182: 163-169. doi: 10.1016/j.puhe.2020.03.007 Wlezien, Christopher (1995). The Public as Thermostat: Dynamics of Preferences for Spending . American Journal of Political Science: 981-1000. doi: 10.2307/2111666 Nivette, Amy; Ribeaud, Denis; Murray, Aja; Steinhoff, Annekatrin; Bechtiger, Laura; Hepp, Urs; Shanahan, Lilly and Eisner, Manuel (2021). Non-compliance with COVID-19-related Public Health Measures among Young Adults in Switzerland: Insights from a Longitudinal Cohort Study . Social Science y Medicine, 268: 113370. Noone, Chris; Warner, Nikolett Z.; Byrne, Molly; Durand, Hannah; Lavoie, Kim L.; McGuire, Brian E.; McSharry, Jenny; Meade, Oonagh.; Morrissey, Eimear and Molloy, Gerard J. (2021). A Scoping Review of Research on the Determinants of Adherence to Social Distancing Measures During the COVID-19 Pandemic . Health Psychology Review, 15(3): 350-370. Oosterhoff, Benjamin and Palmer, Cara A. (2020). Psychological Correlates of News Monitoring, Social Distancing, Disinfecting, y Hoarding Behaviors among US Adolescents during the COVID-19 Pandemic [Preprint]. PsyArXiv. Oude Groeniger, J.; Noordzij, Kjell; Waal, Jeroen van der and de Koster, Willem (2021). Dutch COVID-19 Lockdown Measures Increased Trust in Government and Trust in Science: A Difference-in-differences Analysis . Social Science & Medicine, 275: 113819. Parrado, Salvador; Perna, Roberta; Closa, Carlos and Pereira Puga, Manuel (2025). Regulatory Approaches in the Face of a Pandemic: Assessing the Role of Ideology, Context, and Rule Design in Four Spanish Regions . Public Administration. doi: 10.1111/padm.70010 Pierson, Paul (1993). When Effect Becomes Cause: Policy Feedback and Political Change . World Politics, 45(4): 595-628. Pierson, Paul (2000). Increasing Returns, Path Dependence, and the Study of Politics . American Political Science Review, 94(2): 251-267. Real Decreto 463/2020, de 14 de marzo, por el que se declara el estado de alarma para la gestión de la situación de crisis sanitaria ocasionada por el COVID-19 (BOE-A-2020-3692). Boletín Oficial del Estado, núm. 67. Reinders, Chris; Kuiper, Malouke E.; Olthuis, Elke; Kooistra, Emmeke B.; Bruijn, Anne L. de; Brownlee, Megan; Fine, Adam and Rooij, Benjamin van (2020). Compliance in the 1.5 Meter Society: Longitudinal Analysis of Citizens Adherence to COVID-19 Mitigation Measures in a Representative Sample in the Netherlands . Amsterdam Law School Research Paper, 2020 33. Appendix Table A1. Main results Conjoint

Preferred situation

Mask wearing

0.210 ***

(0.007)

Outside consumption

0.238 ***

(0.007)

Penalty Fee

0.050 ***

(0.007)

70 % Vaccination rate

0.102 ***

(0.006)

Women

0.006 **

(0.002)

Age (in years)

0.003

(0.002)

Education

-0.001

(0.001)

Ideology

-0.001 *

(0.000)

Regions

Asturias

0.005

(0.004)

Valencia

0.000

(0.003)

Galicia

0.004

(0.003)

Madrid

-0.004

(0.003)

Intercept

0.456 ***

(0.003)

Number of observations

23408
*** p<0.001; ** p<0.01; * p<0.05. Source: Closa et al. (2025). LPM estimations. Wording of the conjoint design (original language: Spanish) Imagínate que nos encontramos en una situación en el pico de una ola con muchos casos de COVID, en la que un alto porcentaje de la población está vacunada con las dos dosis o equivalente y existen algunas restricciones de acceso a bares y restaurantes: la consumición en interiores está limitada y se exige la mascarilla cuando no se esté consumiendo. En este contexto, vamos a mostrarte 4 pares de situaciones. Para cada par, indica, por favor, en cuál de las dos situaciones accederías al local. Puede que no lo hagas en ningún caso o que lo hagas en ambos casos, pero, por favor, elige una de las dos situaciones. Posteriormente te preguntaremos con qué probabilidad entrarías en el local en esa situación. Categorías y atributos para combinar en las siguientes preguntas:

Categorías

Atributos

1

Mascarillas

Todos se la ponen cuando no consumenCasi nadie tiene puesta la mascarilla

2

La consumición debe hacerse

Dentro del local En la terraza

3

La policía

Pone multas a quienes incumplenNo aparece para poner multas

4

La tercera dosis de la vacuna

En el 35 % de la poblaciónEn el 70 % de la población
¿Qué situación prefieres para entrar al local a consumir? 1/4 [Ejemplo, así sucesivamente hasta 4 pares por participante]

#

Situación 1

Situación 2

1

Mascarillas

Todos los clientes se la ponen cuando no consumen

Casi nadie tiene puesta la mascarilla

2

La consumición debe hacerse

Dentro del local

En la terraza

3

La policía

Pone multas a quienes incumplen

No aparece para poner multas

4

La tercera dosis de la vacuna

En el 35 % de la población

En el 70 % de la población
¿Tu preferencia?