Introduction

Populism has received much attention in the academic literature since the beginning of the COVID-19 pandemic. Especially in the second year of the pandemic, as lockdowns and other restrictions became common government responses, and policymakers pushed people to get vaccinated, populist actors increasingly appealed to the disaffected and those opposing government policies (i.e., Lasco 2020; Brubaker 2021; Stoeckel et al. 2022; de Lange 2022). Austria was no exception, as the populist radical right Freedom Party of Austria (FPÖ) positioned itself diametrically opposed to all other parliamentary parties' more or less unified stance on anti-COVID-19 policy (Fallend and Miklin 2022). This raises the question of the types of voters the FPÖ may have wanted to appeal (see Eberl et al. 2021). Did people with populist attitudes differ from the outset from other citizens in the way they perceived the pandemic and its impact on them? In addition, were these individuals also more opposed than others to the anti-COVID-19 measures being considered by the government at a time before COVID-19 politics became a major area of partisan competition?

This focus on the political demand side rather than the supply side is important because the link between populism and the COVID-19 pandemic has been studied mainly from the perspective of political parties (Gugushvili et al. 2020; McKee et al. 2020; Katsambekis and Stavrakakis 2020; Wondreys and Mudde 2020; Kavakli 2020; Meyer 2020; de Lange 2022; Lynggaard et al. 2023). Fewer contributions have dealt with the populist mindset of voters, especially early into the pandemic (but see Vieten 2020; Barnieri and Bonini 2020; Eberl et al. 2021; Serani 2023). However, there is evidence that populist parties were initially uncertain about how to respond to the pandemic (de Lange 2022; Lynggaard and Kluth 2023). They may have been who ultimately followed their potential voters. To understand the mindset of these voters, we require survey data on citizens’ attitudes in the earlier part of the pandemic. The question here is whether individuals who strongly distrust the political mainstream and experts, in general, have a different perception of the pandemic than the rest of the population. Yet, even people who trust governments and experts may still regard the pandemic as an imminent threat and experience great stress and insecurity. It needs to be clarified what impact this combination may have. To examine whether the first factor influences the second, this paper asks the following research question: How do populist attitudes affect individuals’ reactions to a crisis like COVID-19? More specifically, we ask two sub-questions: a) what impact do populist attitudes have on the way individuals perceive the COVID-19 pandemic, and b) what impact do such attitudes have on support for policies to combat COVID-19 under conditions of a liberal democracy?

The underlying argument is that populist voters are different because populism is generally associated with a perceived antagonism between the (corrupt) elites and the (common) people (Lasco 2020; Huber 2022). For example, concerning scientific expertise on vaccinations, populist opposition predates COVID-19 (Kennedy 2019; Dryhurst et al. 2020). As such, populists may perceive COVID-19 as a more significant threat as untrustworthy elites and media communicate it. They may also resent the various government-imposed and expert-recommended solutions but may vary in their assessment of these measures, as will be theorized further below. This is because populism hardly exists in a pure form but is connected with illiberal attitudes, especially in Austria (Heinisch and Wegscheider 2020). Thus, individuals with a populist mindset face a potential dilemma: support restrictive measures to protect the community or reject them based on their distrust of experts and government authority.

To investigate our research question, we conducted an original survey of 1,200 Austrians about the impact of the pandemic, fielded relatively early in the pandemic in September 2020. Because the survey was fielded a few months after the first COVID-19 wave and a nationwide lockdown, respondents were familiar with the pandemic and various policies to contain COVID-19. Moreover, at the time of the survey, Austria’s largest populist party, the FPÖ, had not yet fully exploited the government’s Corona response as a key mobilization tool, having advocated highly restrictive measures only a few months earlier (Fallend and Miklin 2022).

First, we ask about the extent to which individuals with populist attitudes feel personally affected by COVID-19. As we are interested in their assessment of affectedness, we intentionally wanted to measure citizens’ subjective understanding, as we believe that this, rather than an objective measure of impact, drives political behavior. We asked respondents to rate their affectedness regarding their physical health and economic well-being. Second, we use a unique experiment to examine the relationship between populism and COVID-19 concerning policy solutions to the pandemic. Here, our focus lies on how populist and non-populist respondents react to different policies that impinge on their freedom or the freedom of certain social groups as well as measures that curtail the rule of law.

We find that populist attitudes shape individuals’ perception of the pandemic, both in how much they feel affected in their health (but not in their economic situation) and in how they react to policies that combat the spread of the COVID-19 pandemic. Populists feel more affected in their health, and they reject, to a higher degree, impingements on their individual freedom, and the idea that effectiveness is more important to them than the constitutionality of a policy. These results show that populist attitudes shape individuals' perception of reality outside the obvious political realm and even under the extreme conditions of a global health crisis. Furthermore, our analysis indicates that the general affinity of populists with anti-liberal attitudes is restricted and even reversed when the anti-liberal ideas are connected with technocratic claims of authority and not oriented toward pluralism.

Our article is organized as follows: first, we discuss the literature and theoretical approaches connecting populism and the COVID-19 crisis, developing our hypotheses about how populism should influence how individuals perceive the effect of the pandemic on their lives and possible policy responses designed to counteract it. We, second, introduce our unique survey and conjoint survey experiment. Subsequently, we present our two analyses and discuss the implications of our findings in the conclusion.

Theoretical discussion-populism and the perception of being affected by COVID-19

Before discussing our theoretical argument about the connection between populism and the COVID-19 pandemic, it is important to clarify the conception of populism we employ. Following the ideational approach, populism is defined by the belief in an antagonistic relationship between the good people and the corrupt elite whose power must be broken (Mudde 2004; Hawkins et al. 2018; Taggart 2004; Rooduijn 2014; Mudde and Rovira Kaltwasser 2018). Populism argues that politics should be an expression of the general will of the people (Mudde 2004). On the part of voters, populism exists, according to the ideational approach, as a set of widespread attitudes among “ordinary citizens that lie dormant until activated by weak democratic governance and policy failure” (Hawkins et al. 2018: preface; on the general point see also Akkerman et al. 2014; Castanho Silva et al. 2018; Hawkins and Rovira Kaltwasser 2018). Since “the people” are always right in the populist imagination, common sense forms the basis for all decisions (Mudde 2004).

The subsequent argument will make the case that populists feel especially impacted by the COVID-19 pandemic due to what they perceive as representation failures of the elites (Castanho Silva and Wratil 2021) and that, in turn, this has consequences for how individuals with populist orientations view policy measures mandated by governments intended to counteract the pandemic. Thus, our study draws on arguments Hawkins et al. (2018) put forth about failures of democratic representation and policy discontent (also, Huber et al. 2023).

Populism and feeling affected by COVID-19: the divergence hypothesis

At first glance, it seems plausible to assume that populists may think COVID-19 is a hoax because experts and mainstream media tell them otherwise. If one follows this logic, populists would be less affected by something they do not believe to be real. However, following the literature, a better case can be made for the divergence hypothesis, meaning that populists diverge in their perception from other voters in feeling actually more affected. As populists are, by definition, skeptical about the explanations and recommendations offered by the medical establishment while feeling unrepresented and alienated from those in power, they are likely to view government policies as meaningless or indefensible (Caramani 2017). At the same time, people with populist orientations find themselves in a crisis caused by the pandemic in terms of illnesses and deaths and society’s reaction to it. In these circumstances, even populists can feel very vulnerable. However, because of their distrust of expert advice and policymakers, they cannot rely on the political elites to have the best intentions and that the solutions adopted will produce the promised results. As a consequence, individuals may, therefore, feel uniquely distressed by COVID-19. Similarly, populists might feel affected, even if they thought the pandemic was harmless, by the fact that they were nonetheless forced to comply with variously restrictive measures whose meaning they disputed.

When public policy is connected to such imminent individualized threats, self-interest and experience become dominant factors in individuals’ decision-making (Chong et al. 2001). In either case, populists would likely diverge in their reaction from non-populists. This feeling may be exacerbated by the perceived lack of a comprehensive scientific consensus on the proper way to deal with the coronavirus outbreak (Katsambekis and Stavrakakis 2020). Moreover, as access to alternative information, especially via social media, is straightforward, experts’ prescriptions can be constantly questioned and seemingly easily refuted (Brubaker 2021; Mede and Schäfer 2020).Footnote 1

To capture this individualized perception of threat, we focus in our study on two types of COVID-19-related impacts that appear to matter most to people directly: health and the economy. It should be emphasized that we leave the interpretation of being affected to the individual because, ultimately, any causally linked behavior does not result from an objectively given norm but from a person’s inner impulse to act, which results from the individual interpretation of a situation.

However, we introduced a distinction as to whether being affected refers to the person concerned with him- or herself or to other persons close to them, as will be explained in more detail in the following section on the methodology used. It stands to reason that in a pandemic, the affectedness in health is paramount (Every-Palmer et al. 2020; Huber and Langen 2020). Few things are more valuable to individuals than their health, which is ultimately tied to physical survival. However, the pandemic can impact in a variety of ways. First, there is the immediate consequence of the coronavirus, which can lead to severe illness, and second, there is the experience of fear that particularly vulnerable groups have, even if they do not fall ill. In addition, there is a negative impact on the health of people who suffer from other diseases whose treatment was postponed due to limited hospital capacity or who suffer from psychological problems because their movements have been restricted. Therefore, we emphasize the importance of the subjective feeling of being affected in one’s health; we must leave it to the respondents to determine their affectedness, however ‘objectively’ ill a person may have been. It is the mental state that determines attitudes and behavior.

The other important and widespread impact is economic and, by implication, social, as many individuals lost their jobs, confined themselves to their home offices, or lost their businesses, leading to existential fears for many. This was also a widely reported effect in the media and was, thus, highly salient. Again, this has a strong subjective component, as some people are more resilient than others in the same situation and would judge their affectedness differently. Thus, our concept of affectedness is not limited to the disease itself or the loss of a job or business but remains on a subjective level, which is thus open to political influence and public debate.

Hypothesis formation

While there is no research yet that directly addresses the issue of perceptions of health and economic impacts on people with populist attitudes, at least not for the early stages of the pandemic, we can draw on the extensive literature on the relationship between ontological uncertainty and populist attitudes to make our case (Jost et al. 2007; Kinnvall and Svensson 2022; Heinisch and Jansesberger 2022). The literature shows that populist citizens are more likely to react emotionally to a public health crisis (Filsinger 2023). It is in this context of uncertainties that these individuals are likely feeling particularly affected, especially since they are by definition skeptical of those in power, have less trust in established politics and experts than other citizens, and may even link their experience of COVID-19 to conspiracy theories (Eberl et al. 2021). Thus, the experience of insecurity combined with the feeling of being disconnected and of being distrustful vis-à-vis those who manage the health crisis would arguably increase an individual’s sense of vulnerability. While this sense may have changed as populist politicians began reframing the illness and its impact, our argument pertains to the likely state of mind in the early part of the pandemic when the survey was taken.

Additionally, the expert-driven approach to politics contradicts the preferred model of democracy of populists, as they typically assume politicians pretend to act in the interests of the people but do not necessarily consider their concerns (Caramani 2017). This divergence between the way politicians act during the Corona crisis and the idea of what political action should look like leads people with populist tendencies to feel more negatively affected by the pandemic. As such, the divergence argument guides us in our hypothesis formation.

H1

Individuals with populist attitudes are more likely to feel affected by the Corona pandemic than individuals without populist attitudes.

Conceptualizing populism and anti-COVID-19 policy preferences

As populist attitudes are likely to affect individuals’ perception of the crisis severity, they are also likely to affect the perception of potential policy solutions. First, we note that much of the literature on this question focuses on party and government reactions (Gugushvili et al. 2020; McKee et al. 2020; Kavakli 2020; Meyer 2020) and the effectiveness of lockdown measures (Barbieri and Bonini 2020). As COVID-19 is an unprecedented challenge to policy-making and poses a threat to the entirety of society, the dominant theoretical approach is that public opinion is most likely to follow the leadership of experts and politicians, as is often the case in new and highly complex issues (Lenz 2012; Zaller 1992).

Nonetheless, we proceed from the idea that at the beginning, the management of the pandemic was a policy area with an extremely high degree of uncertainty regarding the effectiveness of the proposed measures since the specific COVID virus was unknown, and the last similar situation occurred more than 100 years earlier. Therefore, in the early stages of the pandemic, we would expect people to follow the advice of experts and adopt expert-driven policies (Gadarian et al. 2020, p. 3). However, these policies were highly intrusive and included no-contact orders, nighttime curfews, and mandatory mouth-to-nose protection in public areas of life. Not surprisingly, studies have shown that individuals have different preferences regarding these policies (e.g., Arceneaux et al. 2020). Thus, as the initial rally around-the-flag effect (Kritzinger 2021) and general respect for authority in a time of crisis were beginning to wear off (Johansson 2021) and consequences of the first lockdown measures were absorbed (for Austria, see Łaszewska 2021), we expect people to begin diverging in their assessments of the pandemic and public policy.

Responses to COVID-19 increasingly pitted fundamental political interests against each other, particularly health concerns against economic development, but also disagreements about the appropriate level of government interference in individual decisions and lives. As populists are ideologically predisposed to distrust experts and may feel differently about affectedness than the public in general, there are three reasons for their likely opposition to the government’s COVID-19 policy: activation, blame attribution, and populist ideology. First, the strong sense of being affected by the pandemic is likely to act as a trigger activating and strengthening populist orientations such as those vis-à-vis decision makers. We know such mechanisms from experiences in other crises (Brubaker 2021; Thele 2022; Verbalyte et al. 2022). Second, populists are likely to blame both the pandemic and their negative experiences on those in power. As a result, populists would be more willing to reject their policies. Third, this attitude is reinforced by a more deeply rooted ideological principle, according to which populists describe themselves as the true champions of popular freedom, especially in the face of restrictions imposed by what they see as illegitimate elite decision-makers (Heinisch and Wegscheider 2020; Caramani 2017). This suggests resistance to restrictive measures imposed by the government (Brubaker 2021; Katsambekis and Stavrakakis 2020).

Further, drawing on the growing literature on this subject, we focus on the relationship between efficacy and the liberal principle (Katsambekis and Stavrakakis 2020; Wondreys and Mudde 2020). Thus, two central questions lie at the core of understanding how populists might react: one concerns the tradeoff between efficacy and the liberal principle, and the other concerns the relative importance of illiberalism versus anti-pluralism (Bauman 2013; Basile and Mazzoleni 2020; Kallis 2018). Liberal principles such as constitutionality and minority rights are typically rejected if they get in the way of restoring stability and security to the community (Bauman 2013). Therefore, we may consider this matter also from the question of which of the two impulses of populism, anti-liberalism or anti-pluralism, is the more important factor in responding to COVID-19. Some scholars, notably Pappas (2019), see populism as predominantly illiberal, which suggests that populists are less likely to reject illiberal measures than other individuals do. Differently stated, constitutional safeguards protecting minority positions should not stand in the way of effective protections for the community when dealing with COVID-19.Footnote 2 Due to the contradicting rationals of these arguments, we formulate the following hypotheses:

H2

Populist attitudes negatively affect the agreement toward policy measures mitigating the spread of COVID-19.

H2a

Individuals with populist attitudes have a higher likelihood to disagree with any policy measures mitigating the spread of COVID-19.

H2b

The more a policy is targeted at the individual, the higher the likelihood of individuals with populist attitudes to disagree with the policy mitigating the spread of COVID-19.

H2c

The more a policy targets democratic principles, the higher the likelihood of individuals with populist attitudes to disagree with the policy mitigating the spread of COVID-19.

Case selection, data, and methods

Case synopsis and the context of the survey

To test our theoretical assumptions, we chose Austria as a case study. It is one of the most prosperous and stable democracies in Western Europe, with a good public health system and an efficient public administration that should be able to cope comparatively well with the pandemic economically and administratively. However, Austria is also a country with one of the longest-standing and strongest populist parties, regularly supported by more than 20% of the electorate. The FPÖ claims to defend a national-liberal heritage (hence its name) while espousing authoritarian and radical right-wing ideas. This relates directly to our theoretical argument. The way the pandemic developed and the government’s response, described below, also make Austria an excellent and relevant case for this research question. During the pandemic, the government comprised a coalition of the mainstream right Austrian People's Party (ÖVP) and the Greens. Under the leadership of Chancellor Sebastian Kurz, the ÖVP had moved to the right, while the Greens were considered a center-left party with a focus on the environment. However, the Greens also controlled the Ministry of Health, which became the main target of government critics as the pandemic progressed. The second largest party, the Social Democratic Party (SPÖ), the FPÖ, and the small liberal NEOS party were in opposition. While all parties except the FPÖ more or less maintained a consensus on Corona, at least in the early stages of the pandemic, the FPÖ began accusing the government of overreach and incompetence as early as the summer of 2020. However, at that point, the FPÖ was still dealing with the fallout from an earlier scandal and a split leadership between Norbert Hofer, the party leader, and Herbert Kickl, the parliamentary group leader, both of whom held different positions on dealing with COVID-19 and its severity (see Fallend and Miklin 2022 on the Austrian government’s handling of Corona and its political implications).

To assess the state of mind of Austrian voters, including those with populist leanings, regarding the pandemic and government responses, it is important to consider polling data collected relatively early in the pandemic period, in September 2020. At this point, citizens’ attitudes are still little influenced by the extensive policy debates, and thus political framing and partisan cueing, that occurred later. At the same time, we can safely assume that respondents already had a good enough understanding of the pandemic, possible countermeasures, and their consequences at the time of the survey.

As stated above, our survey was fielded in September 2020, following summer months with relatively low rates of COVID-19 incidents but amidst rising concerns that the pandemic was returning. This situation was preceded by relatively resolute and successful government efforts in coping with the first wave (Fallend and Miklin 2022) while being under the impression of the increasingly dire situation in neighboring Italy. However, when COVID-19 numbers started rising again in late summer of 2020, the government seemed unsure how to respond. At the time, the Austrian Constitutional Court had ruled several of the previously adopted measures to curb the spread of COVID-19 unconstitutional. There was also growing disagreement among the political parties about how best to respond, though consolidated new positions had yet to emerge.

The survey

To test our hypotheses empirically, we fielded a nationally representative survey (see Table 3 in the Appendix) in Austria, tapping into various aspects of the COVID-19 pandemic. The stand-alone survey went into the field at the beginning of September 2020 and had a sample size of 1,200 people. Importantly, this survey captured public opinion before the more specific discussions about vaccine procurement and vaccination skepticism arose that have since dominated public and scientific discourse. It, therefore, provides a relatively unvarnished, rare insight into the views of the Austrian public on the relationship between perceptions of the pandemic and government action on the one hand and those individually affected on the other. Nevertheless, we may assume that the orientations surveyed here form the basis for the debates and sentiments that have emerged since then.

Table 1 Example of conjoint, COVID-19 policy packages showing all potential policy options
Full size table

The survey was fielded online through the survey company Market Institute, which randomly sampled respondents from their respondent pool of 25,000–30,000 Austrians. The survey contained two questions about the respondents’ feeling of being affected by the COVID-19 crisis, one targeting their own health and the other asking about their own financial situation. Here, we focus exclusively on people's subjective perception so that the health impact is not limited to having contracted COVID-19, and the financial impact is not limited to having lost one's job. It is left to the respondents and their subjective assessment whether and to what extent they felt affected.Footnote 3

On a scale of 0 to 10, how much did you feel negatively affected [in your health well-being OR economically] by the Corona crisis?

While populist attitudes have been measured in a range of ways, and a battery of up to 18 questions has been identified as the most reliable way of measurement (Akkerman et al. 2014; Castanho Silva et al. 2018), the same research shows that all of these questions correlate very highly. Thus, given the constraints of a survey tapping into various aspects of the pandemic, our approach was to measure populism using one item proxy that combines the three central characteristics of populism: the common people, the bad elites and a Manichean conflict between them. Our item asks for respondents’ agreement to the following statement, with the response scale ranging from 0 (do not agree at all) to 10 (agree very much):

In our country, the powerful listen far too little to the common people.

We define as 'populists' those respondents who agree with the above statement with a value of 7 to 10 on the scale. The average and median value of responses is 7, making this a logical cutoff point.Footnote 4 We acknowledge that this measure is a proxy and does not have the same level of internal validity as one of the full batteries of populism survey questions that measure each populism dimension in several questions. As mentioned, practical constraints around survey timing and cost led us to use this proxy. However, we are confident that this proxy measures populism for several reasons: first, it captures the main conceptual ingredients of a populist sentiment: people-centrism and anti-elitism are explicitly included, and the question wording constructs a conflict between people and elites. Second, 82% of FPÖ voters among our respondents are populists, while only 62% of all other respondents fall into this category. This lends face validity since the FPÖ is Austria’s dominant populist party. We also show in the Appendix that the regression results are similar when replacing populism with FPÖ vote, though the results in health show that the two measures are distinct. Moreover, we cross-validated our findings with the results of the Austrian National Election Study 2017 (Wagner et al. 2017), which employs an expanded populism battery based on Silva et al. (2018). It finds that about 72% of Austrians have at least moderate populist attitudes,Footnote 5 which corresponds to the results in our survey and shows further validity.

As a third central element, the survey contained a conjoint experiment that simulates a decision-making situation by giving respondents a binary choice between two policy packages designed to combat the spread of COVID-19. Of course, a survey experiment cannot perfectly simulate a real policy-making decision with all its personal or societal consequences. However, Hainmueller et al. (2015) have shown that forced-choice conjoint experiments mimic real-world behavior very closely, using field experiments as the benchmark. Thus, while our experiment is subject to the usual limitations, voters’ responses to conjoint experiments are a functional substitute for real-world behavior.

The conjoint experiment asked respondents to indicate their preferences between two policy packages, and Table 1 shows an example experimental decision. To set up this decision, we provided the respondents with situational context and the instruction to imagine a hypothetical situation. The text shown (in German) is as follows:

Please imagine the following situation:

After months of opening up the economy, with shops and restaurants finally recovering and tourists also flocking to the country again, there is an explosive resurgence of COVID-19 disease that can no longer be controlled by conventional measures.

In turn, the government again severely restricts freedom of movement in the country and announces additional measures as essential.

In September 2020, at the time of the survey, stable and low COVID-19 infections and hospitalizations marked Austria’s situation. At the same time, there were widespread discussions about the disease. Thus, the above hypothetical situation represented the not-yet-existent but realistic threat of another wave of infections and the resulting necessary policy measures.

We introduced the policy packages with the following statement:

Below, we show you different packages of measures that the government can take. Please select the package of measures you would personally prefer.

The first three policies are versions of proposals that were discussed in Austria or elsewhere, while we included the fourth more general policy-making principle to hone in on the democratic nature of decision-making during a crisis. Each policy relates to a different aspect of liberal democracy. A technical device to trace and reconstruct personal contacts impinges on personal freedom and privacy, while indefinitely isolating a specific social group impinges on their freedom of movement and conduct as well as their protection against discrimination. Harsher law and order measures, especially in conjunction with the idea of calling on the population to report suspects, is closely connected to the type of social control implemented by non-democratic regimes and impedes the rule of law, especially the freedom from unlawful accusation and persecution. All three measures are targeted at individuals, though particularly measures 1 and 3 affect the integrity and security of every individual person. The fourth measure is a general policy-making principle that contrasts effectiveness and constitutionality, which shows no perceivable impact on individuals but targets the decision-making process in general.

Analysis: populism and COVID-19

We first present descriptive statistics to give an overview of the prevalence of the feeling of being affected by the Corona pandemic and certain political attitudes. The first result is surprising in two respects: first, the overall magnitude of respondents feeling affected in their health is far higher than actual case numbers would have suggested. By the time of the survey, Austria had had a total of about 29,000 confirmed cases out of a population of nearly 9 Million (WHO/global 2019). However, 23% of our respondents rated their health affectedness as higher than the mid-point 5, and 31% of respondents felt affected in their financial situation to this degree. Of course, in a subjective assessment, individuals may interpret health effects very differently. However, regardless of the form in which the pandemic was experienced, respondents still registered it as a health problem. What matters for our study is what effect people subjectively attribute to COVID-19 and how they react.

Second, given the low number of reported COVID-19 cases, but the nationwide lockdown, soaring unemployment, workplace closures, and supply chain disruptions, as well as the massive impact of travel restrictions in a country where tourism is of enormous importance, it was reasonable to expect that the perceived economic impact would be far greater than the health impact. Yet, the distribution of feeling affected is remarkably similar for both dimensions. Figure 1 shows that the majority of Austrians felt not or only slightly affected by COVID-19 and that there is little difference between the two measures. Only at the extreme end of feeling strongly affected do we find slightly higher percentages for the financial situation compared to health.

Fig. 1
figure 1

How affected Austrians feel in their health and financial situation

Full size image

In agreement with previous studies (Plasser and Ulram 2003; Huber et al. 2020), populist attitudes are widespread among Austrians. In our survey, 63.79 percent of our respondents are categorized as populist.

We conduct the following investigation using OLS regression models, which allow us to test not only populist attitudes but also the possible influence of other factors on subjective concern about the Corona pandemic. As policies are strongly connected to the government at the time, we interact populist attitudes with respondents’ status as government or opposition voters to control the robustness of the populism effect. Among the control variables, we include sociodemographic factors such as gender, age, and education level. Further, we include the subjective assessment of the respondents’ income because it will likely affect individuals’ assessment of their economic situation during the COVID-19 pandemic and the political attitude of authoritarianism, as it will likely affect individuals' general response to the proposed policy solutions.Footnote 6 We do not use partisanship in the full model because in Austria, being populist is too closely related to being a voter of the FPÖ. In our models, we treat health and economic concerns of respondents separately.

Populism and feeling affected by COVID-19

We begin our analysis by testing hypothesis 1 about the relationship between populism individuals’ levels of feeling affected by the COVID-19 pandemic. We first turn to individuals’ levels of feeling affected by COVID-19 in their health. Figure 2 shows the coefficient plots of three OLS regressions, the first only including populist attitudes, the second adding whether the respondent was a government or opposition voter and interacted this factor with populism as a control for the robustness of the populism effect, and the third model adds the control variables. All models include weighting for party voters. For individuals’ health, we can clearly confirm our H1 that populist attitudes increase feeling affected in one's health. This effect is robust against controlling for government support as a closely related concept.

Fig. 2
figure 2

Coefficient plot for regression explaining feeling affected in own health, three models

Full size image

Turning to how much respondents feel affected in their economic situation, Fig. 3 replicates the same analyses we undertook for health. Generally, we find similar patterns but with slightly weaker effects and significance. Figure 3 shows that populist attitudes increase respondents’ feeling of being affected, albeit the effect loses its significance when introducing further variables. This lends less support for H1.

Fig. 3
figure 3

Coefficient plot for regression explaining feeling affected in own finances, three models

Full size image

In summary, this part of our analysis shows that our hypothesis explaining feeling affected with populism is confirmed for respondents’ health situation. This highly individualized perception is thus clearly guided by respondents’ relationship to the established system. However, we find less support for our hypothesis in our analysis of feeling affected in one’s economic situation. It seems that the assessment of one’s socioeconomic status is based on factors like relative deprivation instead of populism. Nonetheless, our results significantly challenge the conventional wisdom by showing that people with populist orientations, those strongly convinced that selfish elites disregard the interests of ordinary people, are themselves not dismissive about the effects of COVID-19. In fact, as we suspected with respect to health, these individuals feel more affected than others because of their alienation from those who make the critical decisions. Thus, the relationship between populist sentiment and affectedness is more complicated, and disagreeing with decision-makers and having a non-mainstream perception of the disease does not automatically mean that one does not feel affected. Tapping into this particular sense of being affected can provide the impetus for political activation and eventual action.

Populism and the policies to counteract the COVID-19 pandemic

Turning to our second research question, we analyze the response patterns of populist and non-populist respondents to our experiment. Figure 4 shows each policy’s average marginal component effects (AMCE) in the experimental policy packages against COVID-19. The AMCE is one standard procedure to analyze conjoint experiments. It is interpreted as the marginal effect of changing a given characteristic on the population’s probability of choosing a policy package, which is averaged over all possible values of the other characteristics. In other words, it shows whether including the illiberal or undemocratic version of a policy affects (positively or negatively) the likelihood that the respondents choose a policy package. Both panels in Fig. 4 show that illiberal policies like a mandatory tracing App, forced isolation of vulnerable social groups and the encouraged reporting of non-compliant people to the police for stronger punishment each have negative effects on package selection. Thus, when given the choice, respondents prefer to avoid these. We find no effect for the more abstract measure of valuing effectiveness over constitutionality, meaning that respondents care little about its inclusion.

Fig. 4
figure 4

Average marginal component effects for policies depending on respondents with or without populist attitudes

Full size image

To test hypothesis 2, that we find differences between populist and non-populist respondents and its sub-hypotheses, we need to compare the two panels in Fig. 4.Footnote 7 The Figure shows (and Table 2 showing the numerical values confirms) that the disagreements with forced isolation and reporting non-compliance are the same in both panels. Thus, we can reject H2a that populists oppose all policies. However, populist respondents show a stronger adverse reaction to the tracing App, where the coefficients are clearly outside each other’s confidence intervals, and to the effectiveness over constitutionality argument, where the coefficients are just at the border of the 95% confidence intervals. Thus, we cannot conclude that populists are more likely to disagree with any COVID-19 measures but, in fact, do so in certain conditions and not others, which hypotheses 2b and 2c explore.

Table 2 AMCEs for populists and non-populists
Full size table

We find evidence for our hypothesis 2b that populists react more negatively to policies that are directed at the individual. Arguably, the tracing App is the only policy directly targeting individuals (while constitutionality is more abstract) and affects all individuals equally. The forced isolation is restricted to specific social groups and reporting to those showing specific behaviors. Thus, survey respondents can reasonably deny belonging to a group targeted by these two latter policies. Yet, including the all-affecting tracing App decreases the likelihood of package selection by 20% among populist respondents and only about 10% among non-populists. This finding confirms H2b. The tracking App is the only measure targeting individuals that does not provide for a selective opt-out but automatically includes everyone. The difference between populists and non-populists also becomes clear when we examine our hypothesis 2c, that populist individuals reject policies that target democratic principles. The general valuation of efficiency in fighting COVID-19 over a policy’s constitutionality attacks democratic principles, as it fundamentally rejects the rule of law. Surprisingly, non-populist voters are indifferent to this measure; the main effect is nearly on the zero line, and the confidence interval reaches into the areas of positive and negative effects. Populist respondents, on the other hand, have a greater distaste for this measure, with the main effect being a reduction of likelihood by about 5%. While the confidence interval straddles the zero line, the effect is clearly negative. Although this measure does not target individuals, it empowers ‘elite decision makers’ to impose comprehensive restrictions, like the tracking App, that can cover everyone.

Thus, finally, we can generally confirm our hypothesis 2 that populist attitudes influence how respondents react to illiberal and anti-democratic measures to combat COVID-19. The finding that populists are more likely than other citizens to defend the liberal position when choosing between it and effectiveness challenges conventional wisdom. It also suggests that, under certain circumstances, the dual impulses of illiberalism and anti-pluralism inherent in populism unravel so that a preference for a liberal position emerges, weakening the anti-pluralist one.

Conclusions

The COVID-19 pandemic has radically disrupted the lives of people worldwide, their economies, and the political decision-making in their countries. Since one might assume that such a global crisis threatening everyone’s livelihood would overwrite the political divides of ‘normal’ times, this presents a unique condition to test how deeply political attitudes like populism affect individuals’ perception of reality. Do populist individuals perceive the pandemic and its proposed solutions differently than non-populists? To answer this question, we developed a special survey with a conjoint experiment that measured Austrian respondents’ populist attitudes, the extent to which they felt affected by COVID-19 in their health and economic well-being, and their reaction to illiberal policies designed to address the dangers of the pandemic.

Before delving into the specific conclusions, it is important to acknowledge not only that the number of people who report feeling their health as impaired is significantly higher than the official COVID-19 caseload suggests, but more importantly that what might be considered primarily a biomedical experience is so profoundly shaped by political attitudes.

In examining which respondents feel more or less affected by COVID-19, our analysis shows that populism is clearly connected to feeling more affected in terms of individual health. This highly individualized perception of personal health is thus clearly guided by the individual’s relationship to the established system and the government. On the other hand, we find little support for the effect of populism on feeling affected in one’s economic situation. Instead, whether one perceives one’s income as sufficient has a strong influence.

Turning to the question of how respondents react to various illiberal and undemocratic policies to combat the COVID-19 pandemic, we have two main findings. First, the general patterns of our survey experiment show that the Austrian respondents tend to reject policies that are illiberal and more aligned with authoritarian values. This contrasts with US and UK findings, where respondents are more likely to choose authoritarian policies (Arceneaux et al. 2020). Thus, while a call for stricter rules and harsher enforcement seems logically consistent with an unprecedented global health crisis, we find evidence against a general trend. Whether this is due to country differences, differences in the timing during the pandemic, or the degree of authoritarianism within the policy proposals can only be investigated in a meta-analysis of all the individual country- and small-N comparative studies being published.

Second, and directly speaking to our overall research question, we find clear evidence that populist attitudes matter for individuals’ reactions to illiberal and undemocratic policies. Although it is often claimed that populist attitudes are directed against liberal principles (cf., Galston 2018; Pappas 2019; Plattner 2010; Taggart 2004; Urbinati 1998; Rovira Kaltwasser 2014), we find the exact opposite: that populists are more likely to reject illiberal policies than non-populists. We theorize that populists’ rejection of the political establishment weighs heavier in their perception of government policies than their illiberalism, which lies more in a rejection of pluralism. Given that COVID-19 policies were mostly based on expert assessments of the situation rather than on popular demand (Brubaker 2021; Katsambekis and Stavrakakis 2020), this finding also confirms the general tension between populism and technocracy (Caramani 2017). While both are anti-liberal in that they restrict the range of acceptable sources of policy development, the COVID-19 pandemic shows the inherent contrast between these two viewpoints. As the pandemic approaches the technocratic ideal of elected politicians abdicating decision-making responsibilities to medical experts, it also highlights the competing claims to authority.

Our research challenges the conventional wisdom in two ways: first, that resentment toward elites implies a lower degree of being affected by a disease that those elites are trying to manage, and second, that populists would be less likely to defend liberal positions against unconstitutional measures than non-populists. Our findings call for further analysis of how these views evolved as the pandemic progressed. It also sheds light on the initial difficulties of radical populist parties trying to appeal to voters who fear effective but seemingly authoritarian measures. Finally, we see in these findings, which emerged when the pandemic appeared to be largely under control, an early indication of what manifested in widespread skepticism about expert-led government vaccination campaigns in the coming year. Our main theoretical conclusion for populism research, however, is that these results suggest that populists are not indiscriminately anti-liberal.