key: cord-0833515-mqcq4kwn
authors: Normand, Alice; Marot, Medhi; Darnon, Céline
title: Economic insecurity and compliance to the COVID‐19 restrictions
date: 2021-11-29
journal: Eur J Soc Psychol
DOI: 10.1002/ejsp.2828
sha: 63ddd4c51891186943dddc0e81c7f2f2d3dc5292
doc_id: 833515
cord_uid: mqcq4kwn

The present research investigates economic insecurity as one potential determinant of citizens’ compliance to restrictive policies implemented to combat the spread of the COVID‐19 virus. Two pre‐registered studies (N (Study 1) = 305; N (Study 2) = 175) were conducted in France during the second and the third wave of the pandemic to test correlational (Study 1) and causal (Study 2) links between economic insecurity, perceived constraints and transgressions (self‐reported, Study 1; intended, Study 2). We hypothesized that the effect of economic insecurity is particularly strong for restrictions involving social affiliations (e.g., not meeting with friends and families). Results indicated that economic insecurity indeed increases perceived constraints and the tendency to transgress but for all types of restrictions (involving social affiliation or not). We propose that economic insecurity poses a threat to individuals’ self‐agency, which triggers psychological reactance to any form of restrictions to individual freedom. This article is protected by copyright. All rights reserved

To fight against the pandemic, most countries implemented restrictive policies to break the spread of the virus. Those policies evolved over time and across countries from "soft" (e.g., washing hands, wearing masks) to "hard" measures (e.g., closing borders, curfews and lockdowns). Although following recommendations was strongly encouraged and disobedience may even be fined, the success of these restrictive policies ultimately depended on citizens' compliance. Unfortunately, the exceptional pandemic went together with an exceptional economic recession, thereby aggravating the climate of uncertainty and threat. The present research aims to examine whether economic insecurity undermines citizens' compliance with the restrictive policies aiming to contain the virus transmission.

A growing amount of research in psychology has attempted to identify the psychological mechanisms that underlie individuals' adoption of prescribed and recommended behaviors to curb the risk of getting infected and transmitting the SARS-CoV-2 virus. Among them, the belief in conspiracy theory seems to hold an important place. Numerous correlational studies suggest that stronger endorsement of conspiracy beliefs regarding is negatively related to compliance with restrictive policies (Bierwiaczonek et al., 2020; Freeman et al., 2020; Imhoff & Lamberty, 2020) . This negative association seems to be increased by a lower level of anxiety about the risks related to the pandemic (Erceg et al., 2020) , right-wing political attitudes (Farias & Pilati, 2021) and vertical individualism (i.e., viewing oneself as an autonomous individual who accepts inequality) (Biddlestone et al., 2020) . Other research reported that risk tolerance (Müller & Rau, 2021) and lower levels of political trust (Bargain & Aminjonov, 2020) negatively affect compliance to social distancing and lockdown prescriptions.

Despite acknowledging that the socio-economic climate played an undeniable role in the aforementioned psychological factors, so far, little research has been conducted to specifically examine how economic stressors may affect compliance to restrictive policies. To this matter, Probst, Lee and Bazzoli (2020) recently collected responses from US workers across 43 states and found that economic insecurity (i.e., job and financial insecurity) was negatively correlated to selfreports of compliance with prevention behaviors. The more participants experienced economic insecurity, the lower their compliance to the recommended prevention behaviors was (e.g., physical distancing, staying at home, disinfecting items and surfaces, washing hand frequently, avoiding touching one's face, and coughing inside one's elbow or in a tissue). Drawing on scarcity theory (Mani et al., 2013; Mullainathan & Shafir, 2014) , authors interpreted this relationship as a result of a cognitive impairment caused by economic vulnerability. Scarcity theory posits that dealing with a lack of valued resources (e.g., time, food, money) induces a temporary mindset that consumes cognitive resources. With less cognitive resources available when engaging actions, individuals with a scarcity mindset would be less able to make appropriate decisions and to exert self-control while focusing on the most urgent and pressing needs. With regard to the pandemic, dealing with economic insecurity and having to anticipate the consequences of potential job or income losses could have translated into less cognitive resources to comply effectively with the prescribed and recommended health-related behaviors. Importantly, Probst et al. (2020) also found that the negative relationship between economic insecurity and self-reports of compliance with prevention behaviors was attenuated in US states offering more generous unemployment benefits. In other words, knowing that they have a safety net in case of job loss could contribute to relieving US employees from economic insecurity worries, thereby preserving cognitive resources.

However, in their study, Probst et al. (2020) focused on six behaviors that involved only personal and respiratory hygiene (e.g., hand washing, disinfection of surfaces, coughing inside one's elbow). Those behavioral recommendations all belong to the category of "soft" policies compared to harder policies that were implemented in many countries such as curfews and lockdowns. These policies are considered "hard" since they curtail individuals' freedom of action regarding their personal and, more importantly, social life. Not only they are more constraining but they are also distinct in nature from the "soft" policies. In fact, not being allowed to assist social gathering or to visit family members takes on the social dimension of individual's life. And with this difference in nature may come differences in the psychological processes that underlie their compliance. While it seems plausible that forgetting to wash one's hands systematically after contact with potentially contaminated surfaces may result from a temporary lack of cognitive resources (due to a scarcity mindset), the same rationale hardly extends to getting involved in social gatherings (but see Xie et al., 2020 for a different contention). Not complying with "hard" restrictive policies could rather emanate from making a deliberate decision than from cognitive lapses. For this matter, whether economic insecurity plays a role in compliance with restrictive social policies remains an open question.

The phenomenon of social affiliation in situations of stress is well-known in psychology (Gump & Kulik, 1997; Schachter, 1959) and despite competing theories on its determinants, there is a large consensus about its existence (Baumeister & Leary, 2017) . Indeed, the experience (or anticipation) of a stressful and ambiguous situation increases the need to affiliate with people who share the same experience in order to exchange information and attitudes that helps mitigating the anxiety associated with the unpredictability of the situation. Economic insecurity arises from the experience of uncertainty regarding one's employment and financial resources (Probst, 2005) , which makes this situation particularly likely to increase the need for social affiliation. Moreover, the pandemic and its economic collateral damages altogether concur to a very stressful context, as attested to by the impact on individuals' well-being and mental health (Holmes et al., 2020; Martinelli et al., 2021) . For those who do not benefit from stable employment status or whose professional activity directly suffers from the consequences of the pandemic, uncertainty and stress may be at their highest peak. One basic and instinctive way to cope with stress is to seek for social contacts, and preferably with similar or close others (Taylor, 2011) . And that is precisely what was strongly recommended not to do during periods of lockdown. Coping with stressing situations would therefore directly enter in conflict with the restrictive social policies that are necessary to contain the virus propagation.

The objective of the present research is to examine the role of economic insecurity in individuals' compliance to restrictive policies. The two pre-registered studies (Study 1: https://osf.io/cg3ks/?view_only=16b9069c47d7465e8bc45fd43a19250a; Study 2: https://osf.io/g3hdz/?view_only=5167370d48a54e439f8a6412269a555d) were conducted in France during the second and the third wave of the pandemic in 2020. We tested the hypothesis that experiencing economic insecurity is associated with the perception of restrictive policies as more constraining and with a greater tendency to transgress them. In addition, in both studies, the nature (social vs. non-social) of the restrictions is tested as a within-participants variable: the association between economic insecurity and perceived constraint and intention to transgress was expected to be particularly true for "social" restrictions, namely, restrictions that deprive people from interacting face-to-face with others (e.g., not meeting with friends and families). The full material, including data and scripts are available on https://osf.io/s5qr8/?view_only=16786ba0b51f4e2ca2e0d52652ea8ee8 . In both studies, we report how we determined our sample size, all data exclusions , all manipulations, and all measures (Simmons et al., 2012) .

Study 1 was run during the second lockdown in France, starting from October 30, 2020, which involved several measures and policies restraining citizens' mobility. Residents were allowed to go outdoors only for essential needs (e.g., doing groceries, going to a medical appointment). Except for imperative family reasons (e.g., picking up one's child from school), social contacts with friends and family members were not an exempting condition. On November 24, the French government announced that the lockdown would be eased in three progressive steps, starting on November 28. Study 1 was then run from November 30 to December 8 and involved questions regarding the past four weeks in November that took place under lockdown.

We conducted an a priori power analysis on the basis of an effect size of d=.40 (f²=.04), which constitutes our smallest effect size of interest (SESOI; Lakens et al., 2018) . Our SESOI was established by averaging the effect sizes of the correlations between compliance with CDC guidelines and two different predictors of economic insecurity (job and financial insecurity) reported by Probst et al. (2020) . The power analysis using G*power (Faul et al., 2007) indicated that 265 participants are required to have 90% chances to detect an interaction effect of a f²=.04 size. We planned to oversample and recruit around 300 participants. Three hundred five French participants (207 women, 96 men and 2 undefined) were recruited online through social media and large email diffusion (M age = 38.83; SD age = 13.109) and took part on a voluntary basis.

Economic insecurity. Economic insecurity was assessed with the Affective Financial Stress measure (Petitta et al., 2020) composed of 4 Likert scale items translated into French by a bilingual researcher. Participants indicated on a 5-point frequency scale ranging from 1 "Never" to "5 "Always" how often they experienced worries and concerns related to their current financial situation (α = .913; M = 2.565; SD = 1.188).

Based on the measures that were applicable during the November 2020 lockdown in France, a list of 11 items related to restrictive policies was created. Participants indicated to what extent each item was inconvenient to them on a 100-point scale ranging from 1 "Not inconvenient for me at all" to 100 "Extremely inconvenient for me" (α = .832; M = 53.488; SD = 20.882). Five of the items were a priori categorized as social, as they clearly involve social restrictions (e.g., "Not being able to visit family, friends and acquaintances"; α = .784; M = 61.313; SD = 24.375) while the other six items were categorized as non-social, as they did not primarily involve social restrictions (e.g., "Not being able to buy anything I want"; α = .707; M = 46.966; SD = 22.262).

Based on the measures that were applicable during the November 2020 lockdown in France, a list of 13 non-recommended or prescribed behaviors, a priori defined as social or non-social, was created. Participants indicated how often they exerted each behavior during the 4-week of lockdown using a 7-point Likert scale (i.e., 1 "Never", 2 "Only once the past 4 weeks", 3 "two to three times the past 4 weeks", 4 "Between 4 and 7 times the past 4 weeks (approximatively once a week)", 5 "Between 8 and 15 times a week the past 4 weeks (2 to 3 times a week)", 6 "16 to 20 times for the past 4 weeks (approximatively 4 times a week)",7 "More than 20 times over the past 4 weeks (almost every day or more)" ; α = .784; M = 2.059; SD = 0.755). Nine behaviors involved social contacts (e.g., "Invite family or friends to come over at my home"; α = .805; M = 2.068; SD = 0.924) and the other four behaviors did not primarily involve social contact (e.g., "Going outdoor for a walk for more than 1-hour"; α = .423; M = 2.097; SD = 0.907) 2 .

After giving consent, participants first responded to the economic insecurity measure, followed by the perceived constraints and the self-reported transgressions measures. The design of the study is correlational with factors of interest being economic scarcity, perceived constraints (social vs. nonsocial) and self-reported transgressions (social vs. non-social). We predicted that economic insecurity positively correlates with perceived constraints and self-reported transgressions, but more specifically with the ones involving social contacts.

Correlation matrix and preliminary analyses. ---Insert Table 1 here --- Preliminary analyses (cf. Table 1) indicated that gender was significantly related to most of the dependent variables (see Supplementary material) and was therefore included as a covariate in further analyses. Data were analyzed using mixed-ANCOVA (General Linear Model) 3 . The model tested whether economic insecurity predicted perceived constraints and transgressions of various restrictions. The social vs. non-social nature of the constraints and transgressions were tested as within-participants variables.

Participants significantly perceived restrictive social policies as more constraining than non-social policies, F(1, 300) = 83.639, p < .001, η 2 = .227, 90% CI [0.153; 0.282]. Moreover, the main effect of economic insecurity indicated that the more economically insecure participants felt, the more they perceived all kinds of restrictive policies as constraining, F(1, 300) = 8.038, p = .005, η 2 = .026, 90% CI [0.005; 0.062]. However, economic insecurity did not interact with the nature (social vs. non-social) of restrictive policies (F < 1). Contrary to our prediction, economic insecurity was not associated with the perception of restrictive social policies as more constraining than non-social ones.

Participants did not report transgressing more social than non-social restrictions, F(1, 300) = 2.169, p = .142. The main effect of economic insecurity was significant, F(1, 300) = 5.853, p = .016, η 2 = .019, 90% CI [0.002; 0.052], meaning that the more economically insecure participants felt, the more they reported transgressing all types of restrictive policies. However, economic insecurity did not interact with the nature (social vs. non-social) of restrictive policies (F < 1). Contrary to our prediction, economic insecurity was not associated with a greater tendency to transgress restrictive social policies than non-social ones.

We hypothesized that the economic insecurity engendered by the consequences of the pandemic would enhance individuals' need for affiliation (Gump & Kulik, 1997; Schachter, 1959) , which would motivate individuals to seek social contact (Martinelli et al., 2021) despite the restrictions imposed to curb the virus transmission. Accordingly, our main prediction was that experiencing economic insecurity would be associated with the perception of restrictive social policies as more constraining and with a greater self-reported tendency to transgress them than non-social ones. Our results support the hypothesis that economic insecurity is related to poor compliance to the restrictions. However, they did not support a moderation of this association by the nature of the restrictions. Indeed, economic insecurity was related to greater perceptions of constraints and greater reports of transgression for both social and non-social restrictions.

Thus, results from Study 1 confirm that the more participants felt economically insecure, the more they perceived the (social and non-social) restrictions as constraining, and the more likely they were to transgress restrictive policies. However, Study 1's material suffers from some limitations. In line with the November 2020 lockdown in France, two lists of social and non-social restrictive policies were created. A posteriori, the categorization of some of the non-social items seemed problematic. For example, "being restricted in one's mobility" could easily fall into the "social constraints" category if individuals usually travel to visit family or friends. The same logic applies to "not being able to do indoor sport" that can represent both social and non-social contexts. Since most of the items could eventually be social in nature, findings from Study 1 leave open the possibility that the social nature of the restrictions increases the association between economic insecurity, perceived constraints and transgressions of these policies.

In Study 2, the distinction between social and non-social measures was clearer. In addition, while most studies on the determinants of individuals' compliance to restrictive policies are correlational, Study 2's design was experimental and tested the causal role of economic insecurity. More precisely, we designed a manipulation procedure that aimed to put participants either in an economic security or insecurity mindset before reporting how constraining they perceived the restrictions to be and how often they planned to transgress them.

We hypothesized that participants who are led to feel economically insecure should perceive restrictive policies as more constraining and should be more willing to transgress them than participants led to feel economically more secure. In Study 1, this main effect was not moderated by the social (vs. non-social) nature of the restrictions. Thus, in Study 2, two alternative pre-registered hypotheses were tested. On the one hand, with the dichotomy between social vs. non-social restrictions being clearer, and in line with our initial hypothesis, we expected the above effect to be stronger for social restrictions than for non-social restrictions. On the other hand, if results from Study 1 were not due to the ambiguity of the items used, a main effect of the economic insecurity condition should be observed on both perceived constraints and self-reported transgressions, irrespectively of their social nature.

After a series of local lockdowns, the French government announced a third national lockdown starting from April 3 2021. Study 2 was launched on April 8, soon after this announcement. Since the lockdown had just started, Study 2 did not measure self-reported past transgressions but rather intentions to transgress the restrictions that were in force at that time. On April 22, the French government announced that the lockdown would be eased in the following days. As a consequence, data collection was stopped on April 23.

We conducted a power analysis using the smallest effect size detected in study 1 (i.e., f²=.01), which constituted our smallest effect size of interest to detect the predicted interaction effects. The a prior power analysis using G*power (Faul et al., 2007) indicated that 262 participants were required to achieve 90% chances to detect an effect of this size in a design involving within-and betweenparticipant comparisons as well as their interaction. Since we had access to a large pool of participants through course credit (i.e., students from two French universities), we initially planned to oversample and recruit 300 participants. Unfortunately, data collection had to be stopped before we reached the target sample size due to the lift of the lockdown, making our dependent variables obsolete. As a consequence, we were able to collect responses from 180 participants. One of them was exposed to both conditions due to a program failure and three of them failed the attention check measure and were excluded from analyses. The final sample is N = 175 participants (161 women, 13 men, 1 who did not wish to answer the question; M age = 20; SD age = 2.591). The sensitivity power analysis assuming 90% power indicates that we were only able to detect an interaction effect of a size greater than f² = 0.015 with our final sample size.

Perceived constraints. The items list from Study 1 was adjusted and updated as a function of the official measures that were applicable during the April 2021 lockdown in France (see Supplementary material). Participants' mean ratings of inconvenience was calculated for policies a priori categorized as social policies (M = 73.059; SD = 20.809; α = .790) and non-social policies (M = 54.909; SD = 21.807; α = .767).

Intentions to transgress. The item list from Study 1 was adjusted and updated as a function of the official measures that were applicable during the April 2021 lockdown in France. For each item, participants indicated how frequently they intended to exert the behavior on a 7-point Likert scale ranging from 1 "Never" to 7 "Several times in a week" (see Supplementary material). For one item ("moving outside of your county for non-essential trips and involving only you or your family"), participants indicated how likely they were to exert the behavior on a 7-point Likert scale ranging from 1 "No, I do not plan on doing it all" to 7 "Yes I will". Participants' mean intention to transgress was calculated for items a priori categorized as social policies (M = 3.582; SD = 1.436; α = .862) and non-social policies (M = 2.802; SD = 1.010; α = .532) 4 .

At the end of the study, participants were asked to indicate to what extent the affirmation "I am currently worried and stressed by my economic and professional future situation in reason of the COVID-19 crisis" was true for them on a 7-point Likert scale ranging from 1 "Absolutely not true for me" to 7 "Absolutely true for me".

After giving consent, participants were told that the study aimed to collect students' experience of the pandemic. They were randomly assigned to one of the two experimental conditions of 2 (Economic situation induction: insecurity vs. security) x 2 (Perceived constraints: social vs. non-social measures) x 2 (Intentions to transgress: social vs. non-social measures) mixed design, with the economic situation induction as a between-participants variable and the two others as withinparticipants variables. Participants in the "economic insecurity" condition (N = 94) were first reminded that the crisis was going to result, in the years to come, in a major economic crisis with negative consequences on the job market. Participants in the "economic security" condition (N = 81) were first reminded that the COVID-19 crisis had some negative social and psychological repercussions and that, for these reasons, more jobs will be available in the psychology job market. Next, all participants were given as much time as needed to express their thoughts by writing anything that worry (insecurity condition) or reassure (security condition) them as a student, on both the financial and professional dimensions related to the COVID-19 crisis. Then, they completed both measures of perceived constraints and intentions to transgress. They also completed socio-economic and biographical measures as well as the attention and the manipulation check items. Finally, participants were thanked and debriefed.

Mixed-ANOVA (General Linear Model) was used to analyze the data. As in Study 1, the social vs. nonsocial nature of the constraints and transgressions were tested as within-participants variables.

Participants in the economic insecurity condition reported higher levels of economic and professional related worries and stress due to the COVID-19 crisis (M = 4.74; SD = 1.772) than participants in the economic security condition (M = 3.96; SD = 2.015), F(1, 173) = 7.458, p = .007, η 2 = .041, 90% CI [0.006; 0.099]. Our manipulation seems to have been effective.

Participants perceived restrictive social policies as more constraining than non-social policies 5 , F(1, 173) = 129.346, p < .001, η 2 = .428, 90% CI [0.337; 0.502]. However, neither the main effect of the economic situation induction nor the interaction with the nature of perceived constraints reached significance (all Fs < 1).

Participants' indicated greater intention to transgress social policies than non-social policies, F(1, 173) = 59.743, p < .001, η 2 = .257, 90% CI [0.168; 0.340]. The main effect of the economic situation induction was also significant, F(1, 173) = 4.472, p = .036, η 2 = .025, 90% CI [0.001; 0.075]. Replicating results of Study 1 with an experimental design, participants in the economic insecurity condition indicated greater intentions to transgress both types of restrictive policies (M = 3.295; SD = 1.035) than participants in the economic security condition (M = 2.936; SD = 0.938). Contrary to our initial hypothesis, the interaction of economic insecurity condition with the social nature of restrictions failed to reach significance [F(1, 173) = 2.882, p = .095]. The effect of the economic (in)security condition on intentions to transgress did not depend on the type of restrictive policies involved (social vs. non-social).

Study 2 aimed to test the same hypotheses as in Study 1 by manipulating, instead of measuring, economic insecurity. In line with Study 1's results, we found that participants who were led to reflect on the negative economic outcomes the COVID-19 crisis might have on their future economic situation intended to transgress restrictive policies (both social and non-social) to a greater extent than participants who were led to reflect on the potential positive economic outcomes of the crisis for their economic future. We also attempted to test the predicted interaction, which failed to reach significance in Study 1, that participants exposed to a highly insecure economic future would perceive as more constraining and would intend to transgress more restrictive social than non-social policies. Once again, analyses did not reveal such an interaction effect. Indeed, although social restrictions are perceived as more constraining and are more likely to be transgressed than non-social policies, these effects were not moderated by the economic insecurity condition. The early stop in data collection resulted in a lower statistical power than expected, and could explain the non-significance of the interaction. However, this interaction was not observed in Study 1 either. In addition, the fact that the main effect of insecurity was significant, consistently with Study 1, sustain that all things equal, higher economic insecurity is associated to lower compliance to any prescriptions, both social and non-social ones.

The present research aimed to examine how the experience of economic insecurity in the crisis influenced individuals' compliance with the mitigation measures that were recommended and prescribed in many countries. Two studies were conducted to test the correlational (Study 1) and the causal (Study 2) links between economic insecurity, perceived constraints and self-reported past transgressions (Study 1) as well as intentions to transgress (Study 2).

In line with Probst et al. (2020) , the results of the two studies show that economic insecurity may explain part of citizens' non-compliance with COVID-19 restrictive policies. Indeed, the more economically insecure participants felt, the more likely they were to report past (Study 1) or future (Study 2) transgressions. A major strength of our work lies in the causal nature of the effect of economic insecurity (Study 2), which qualifies previous correlational findings and support that, beyond other factors, the mere temporary feeling of economic insecurity was sufficient to increase transgressions of the COVID-19 restriction policies.

Among those restriction policies, some involved a social component (e.g., "Invite family or friends to come over at my home") while others did not (e.g., "Not washing hands before going outdoors"). Our rationale was that the experience of economic insecurity constitutes a highly stressful and uncertain situation which enhance individuals' need for social affiliation. Our main hypothesis was that economic insecurity would lead individuals to seek social contacts, which would translate into 1) perceiving social restrictions as more constraining than non-social restrictions, 2) a greater tendency or intent to transgress social than non-social restrictions. Results of both studies supported that social restrictions were perceived as more constraining and were more often transgressed than nonsocial restrictions, sustaining the view of these restrictions as "hard" (vs. "soft") restrictions. However, results did not support the moderation hypothesis. Indeed, economic insecurity was related to a perception of both types of restrictions as more constraining (Study 1) and to a greater tendency or intent to transgress both types of restrictions (both studies). Worthwhile to note that the absence of the predicted interaction in Study 2 may also be partly caused by a lack of statistical power to detect a small-sized effect, given that data collection had to be stopped right after the announcement of the lockdown lift. Its absence is nevertheless congruent with the results from the well-powered Study 1.

Incidentally, Study 1's additional analyses indicated gender differences in perceived magnitude of constraints depending on their nature: Women (but not men) perceived restrictive social policies as more constraining than non-social ones and men reported more transgressions than women, particularly non-social policies. Study 2's sample characteristics (i.e., psychology students among which men are very much underrepresented) did not allow further investigation of gender effects in compliance to social restrictions. Gender did not moderate the link between economic insecurity and perceived constraints or self-reported transgressions evidenced in Study 1. However, beyond economic insecurity, the question of whether gender may affect perceptions of social and non-social restrictive policies as well as compliance to these restrictions would be worth investigating in future research.

At first glance, the consistent and positive main effect of economic insecurity on (self-reported past and intended future) transgressions could seem surprising given that insecure individuals are those who may suffer the most from the crisis, and thus, would be expected to comply to the restrictions in order to limit the spread of the virus. However, and contrary to that assertion, results support that they are those who (actually or intend to) transgress the restrictions the most. The reasons why economic insecurity diminishes individuals' compliance to all types of restrictive policies would deserve attention. Probst et al. (2020) have suggested that economic insecurity triggers a scarcity mindset, which consumes part of the cognitive resources necessary to deal with the restrictive policies (e.g., washing hands frequently, disinfecting surfaces and objects, etc.). Our initial argument was that a scarcity mindset might not be the best candidate to account for diminished compliance with social restrictions (e.g., getting involved in social gatherings, inviting friends to come over one's home). Instead, we suggested that, in an attempt to cope with the stress and uncertainty related to economic insecurity, individuals would feel an increased need for social affiliation (Schachter, 1959) and would therefore seek more social contacts despite the restrictive official recommendations. In the end, our findings suggest that economic insecurity decreases compliance to all forms of restrictive policies, and not only to the ones that prevent social contact, which cannot be accounted for by an increased need for social affiliation. Unless one assumes that a scarcity mindset can equally result in failures to remind oneself to wash hands and in failures to avoid meeting up with friends or family, the present findings leave alternative explanations open.

As an aggravating factor, economic insecurity may have emphasized the threat that individuals experienced during the pandemic situation . Individuals not only feared to get contaminated but also feared that their future professional and economic situation would be compromised. Because official restrictions and recommendations reduce individuals' freedom of action and thus their sense of agency, economic insecurity may encourage people to try to regain it by using defensive strategies such as psychological reactance (Brehm & Brehm, 2013; Reiss et al., 2020) . In support of this explanation, other recent research showed that individuals with high reactance expressed a decreased preference for security-related behavior and were more inclined to break curfews set in several countries as a mitigation measure against the COVID-19 pandemic (Soveri et al., 2020; Zhu et al., 2020) . Therefore, it seems plausible that economic insecurity poses a threat to individuals' self-agency, and individuals in turn adopt defensive strategies, such as reactance to restrictions to their individual freedom.

Economic insecurity could also be experienced as a consequence of unfair or poor management decisions made by the government that negatively impacted the job market. Indeed, procedural justice models predict that when individuals feel that they have been treated unfairly by an authority, they tend to comply less to the rules or norms established by the said entity, which is perceived as illegitimate (e.g., Murphy, 2017; Murphy & Tyler, 2008) . Individuals experiencing an increase in economic insecurity, could hold the government responsible and considered them even as illegitimate. This could explain the association observed between economic insecurity and compliance to official restrictions, especially in Study 1, the manipulation used in Study 2 being less likely to change justice perceptions (as it is based on the positive vs. negative impact of the crisis on the psychology-related job sector, rather than on the action of the government).

The present research further contributes to understanding the conditions under which citizens may be reluctant to comply with restrictions that are imposed (or highly-recommended) in several countries across the world in a pandemic situation. While a great deal of research has focused on the negative role of belief in conspiracy theories, our findings, along with those from Probst et al. (2020) , point to the aggravating role of economic insecurity in compliance to mitigation measures. Interestingly, economic insecurity is also a strong predictor of belief in conspiracy theories (Douglas et al., 2019; Mao et al., 2020) . Further research is needed to better understand the central role that economic insecurity may play by leading people to want to regain control over their lives, which could further translate into greater adherence to conspiracy theories and defense strategies such as reactance to restrictive policies. Given that many experts from all horizons predict that humanity will have to face major global challenges in the future (e.g., restrictions policies to limit global warming), investigating the impact of economic insecurity on citizens' compliance seems crucial.

Notably, the two reported studies were conducted in France, which tops social welfare spending ranking among OECD countries in terms of percentage of global domestic product (OECD, 2021). While Probst et al. (2020) found that the negative correlation between economic insecurity and selfreports of compliance with prevention behaviors was attenuated in US states offering more generous unemployment benefits, our findings indicate that economic insecurity still has a negative effect on compliance to official restrictions even in a country where the social welfare system is considered as one of the most generous in the world. 

Trust and compliance to public health policies in times of COVID-19

The need to belong: Desire for interpersonal attachments as a fundamental human motivation

Cultural orientation, power, belief in conspiracy theories, and intentions to reduce the spread of COVID-19

Belief in COVID-19 Conspiracy Theories Reduces Social Distancing over Time

Psychological reactance: A theory of freedom and control

Understanding Conspiracy Theories

Misbehaving in the Corona crisis: The role of anxiety and unfounded beliefs

COVID-19 as an undesirable political issue: Conspiracy beliefs and intolerance of uncertainty predict adhesion to prevention measures

G* Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences

Coronavirus conspiracy beliefs, mistrust, and compliance with government guidelines in England

Stress, affiliation, and emotional contagion

Multidisciplinary research priorities for the COVID-19 pandemic: A call for action for mental health science

A Bioweapon or a Hoax? The Link Between Distinct Conspiracy Beliefs About the Coronavirus Disease (COVID-19) Outbreak and Pandemic Behavior

Between Conspiracy Beliefs, Ingroup Bias, and System Justification: How People Use Defense Strategies to Cope With the Threat of COVID-19

Equivalence testing for psychological research: A tutorial

Poverty impedes cognitive function

Are individuals from lower social classes more susceptible to conspiracy theories? An explanation from the compensatory control theory

The Impact of the COVID-19 Pandemic on Vulnerable People Suffering from Depression: Two Studies on Adults in France

Scarcity: The true cost of not having enough

Economic preferences and compliance in the social stress test of the COVID-19 crisis

Challenging the 'invariance'thesis: Procedural justice policing and the moderating influence of trust on citizens' obligation to obey police

Procedural justice and compliance behaviour: The mediating role of emotions

Economic stress, emotional contagion and safety outcomes: A cross-country study

Economic stressors. Handbook of Work Stress

Economic stressors and the enactment of CDCrecommended COVID-19 prevention behaviors: The impact of state-level context

From anxiety to action-Experience of threat, emotional states, reactance, and action preferences in the early days of COVID-19 selfisolation in Germany and Austria

The psychology of affiliation: Experimental studies of the sources of gregariousness

A 21 Word Solution (SSRN Scholarly Paper ID 2160588)

Unwillingness to engage in behaviors that protect against COVID-19: Conspiracy, trust, reactance, and endorsement of complementary and alternative medicine

Affiliation and stress

Working memory capacity predicts individual differences in social-distancing compliance during the COVID-19 pandemic in the United States

Debate: Facing uncertainty with(out) a sense of controlcultural influence on adolescents' response to the COVID-19 pandemic

Values in square brackets indicate the 95% confidence interval for each correlation. * indicates p < .05. ** indicates p < .01 and *** indicates p < or =