Just a “Mouthpiece of Biased Elites?” Populist Party Sympathizers and Trust in Czech Public Service Media

In European countries that have a strong democratic tradition, Public Service Media (PSM) are a crucial part of the national political and cultural public space. As such, they enjoy a significant degree of professionalization and autonomy (Brevini 2014; Sehl et al. 2020), and trust in them is consistently high (Newman et al. 2020: 16). However, for more than two decades, PSM have been under significant pressure from populist politicians, often being accused of bias and being firmly aligned with the ruling elites (Engesser et al. 2017; Sehl et al. 2020), being criticized for the inadequate use of taxpayer funds, ¹ and being threatened with funding cuts (Cushion 2019; Fawzi 2020). Discontent with PSM was also found in audience members with populist attitudes. They express lower trust in PSM, which is significantly less than for private media in many European countries (Schulz et al. 2019; Schulze 2020).

The Czech Republic is one of the countries where this phenomenon occurs (Schulz et al. 2020). Czech PSM—which comprise public service television, public service radio, and public service press agency—have faced constant attacks from populist politicians in recent years. Yet, “despite populists” repeated attempts to gain control over them, Czech PSM maintain a key position within the national media environment and remain independent from direct political control or influence (Stetka 2019, 2020; Stetka et al. 2021). Thus, in key parameters, the situation of the Czech PSM is similar to and illustrative for a number of other European countries, such as Austria, the United Kingdom, and Germany, where PSM hold a strong position within the media landscape (Holtz-Bacha 2021; Sehl et al. 2020).

One of the key questions for the relationship between populism and PSM is how people with populist attitudes deal with these media sources, and from what their relationship with PSM stems. This study will look at the differences in trust in Czech PSM between the populist party sympathizers and the sympathizers of other political parties.

The number of empirical studies that have focused on the relationship between populist attitudes and trust in media is so far limited. However, they all agree that there is a negative relationship between these two variables (Fawzi 2019; Fawzi and Krämer 2021; Hanitzsch et al. 2018; Kalogeropoulos et al. 2019; Schulz 2019; Schulz et al. 2020). Fawzi and Krämer (2021) conclude that people with populist attitudes tend to evaluate media in congruence with populist ideology and perceive them as a mouthpiece for biased elites. Yet, empirical evidence that would provide a convincing explanation for this relationship is missing. The only (and yet empirically untested) theoretical assumption that emerges from the existing studies is that people with populist attitudes expect PSM to produce content that is consistent with the anti-pluralist character of populist ideology (Fawzi 2020). However, PSM are based on pluralism and mostly counter these expectations. This creates a contradiction that may contribute to the lower levels of trust in PSM for an audience with populist attitudes.

Therefore, this study contributes to filling this empirical gap. It explores how populist party sympathizers differ from the sympathizers of other, non-populist political parties in terms of what they expect from the news media they trust, and how this difference affects their trust in PSM. Specifically, the study focuses on the importance of two predictors for audience trust in PSM: the expectation that trusted media will be objective and impartial, and the expectation that trusted media will serve “us” and “our” in-group. As the method of analysis, we use multigroup structural equation modeling.

Populism and Its Vision of (Public Service) Media

Populist ideology is based on the in-group stylization of “ordinary” people as a homogeneous entity and on the fact that this is threatened by out-groups that are comprised of political elites and anyone who does not correspond to the majority (Mudde 2004: 543), including ethnic and religious minorities, such as refugees, immigrants, and Muslims (Fawzi 2019; Jagers and Walgrave 2007; Schulz et al. 2020; Wirth et al. 2016). In other words, populist ideology is inherently based on anti-pluralism (Galston 2018; Müller 2016).

In recent years, Europe and North America have had significant increases in the popularity of political parties that can be described as populist (Mudde 2016; Norris and Inglehart 2019). The economic crisis that began in 2008 and the so-called refugee crisis in 2015 that was triggered by the war in Syria have contributed to this rise in popularity (Ibid.). We witnessed the manifestations of the so-called populist surge in 2016 in both the U.S. presidential election and the British referendum to leave the European Union (Mudde 2016). The Czech Republic is not an exception. Populist parties significantly changed the Czech political map as early as 2013, when the populist movement Action of Dissatisfied Citizens 2011 (in Czech abbreviated as “ANO 2011”) took second place in the general elections (Hanley and Vachudova 2018). The movement was led by Andrej Babiš, one of the wealthiest Czech entrepreneurs and a media owner, who, four years later, became the Czech prime minister.

Studies show that populism's rise in popularity is due to dissatisfaction with traditional political parties, which, in turn, reinforces mistrust in the status quo (Norris and Inglehart 2019) and reduces trust in political and public institutions in general (e.g., Hanitzsch et al. 2018). The anti-pluralistic character of populism contributes to the antagonization of the relationships between the populist party sympathizers (“in-groups”) and all the others (“out-groups”) (Galston 2018; Iyengar and Westwood 2015).

The affinity for populism unsurprisingly affects the citizenry's relationship to the news media. As a source of news, people with populist attitudes tend to choose media that are in sync with their populist attitudes (Fawzi and Krämer 2021; Hameleers et al. 2017). They tend to perceive the mainstream news media (i.e., non-tabloid media owned privately, PSM) as hostile (Fawzi 2019; Schulz et al. 2020) and significantly less trustworthy (Engesser et al. 2017; Fawzi 2020). In the European context, PSM specifically are even less trusted than quality private media by people with populist attitudes (Schulz et al. 2020; Schulze 2020).

Public Service Media as a Principal Contradiction to Populist Ideology

People with populist attitudes perceive PSM as both an extension of the general public discourse and as a biased mouthpiece for elite opinions (Fawzi and Krämer 2021; Schulze 2020). A similar view is expressed by populist politicians themselves, who argue that PSM “represent an inappropriate use of taxpayers' money and they are biased towards the ruling parties” (Sehl et al. 2020: 2).

Several studies reported that the problem in the relationship between trust in media and populist attitudes could be in the essence of populist ideology as such. The studies note that, just as populists antagonize out-groups and expect the media to do the same, they reject those media groups and label them as mistaken for not conforming to this expectation (Fawzi 2019; Fawzi and Krämer 2021; Krämer 2018; Krämer and Langmann 2020; Schulz et al. 2020).

PSM do not meet this expectation by design because they are constituted as institutions that directly support and promote pluralism. From the perspective of Normative Journalism theory, media, in general, should “provide access . . . to a wide range of voices in society, thereby exercising, if not absolute, at least some equality” (McQuail 2013: 81). Specifically, PSM, with their traditional mission to “provide a universal and democratic mass media service” (Goyanes 2021: 21), are typically expected to emphasize the inclusion of minorities in media coverage and social integration (Biltereyst 2004; Debrett 2009; Donders 2012; Goyanes 2021; Just and Büchi 2017; Vyslouzilova 2019), which is often explicitly required by law (Donders 2012; Napoli 2001).

In other words, PSM have a mission that is in direct conflict with the populists' expectation that the media should serve the in-group of “ordinary people” and confront the out-groups. Thus, people with populist attitudes may perceive PSM as dissonant, or they expect something that these media are not entitled to provide. In either case, we assume that it is this value-based alienation from PSM that negatively influences their level of trust.

Populism and the Two-Dimensional Notion of Trust in Media

In our examination of trust in media in relation to populism, we built on a strong integrative base provided by several recently published studies that critically overviewed and evaluated the existing scholarship (Fawzi et al. 2021; Strömbäck et al. 2020). According to Strömbäck et al. (2020) there is a consensus that “news media trust refers to the relationship between citizens (the trustors) and the news media (the trustees) where citizens, however tacit or habitual, in situations of uncertainty expect that interactions with the news media will lead to gains rather than losses.” Fawzi et al. (2021) further define trust in media as “the individual's willingness to be vulnerable to media objects, based on the expectation that they will perform a) satisfactorily for the individual and/or b) according to the dominant norms and values in society (i.e., democratic media functions).” In line with these definitions, it is possible to say that trust in news media stems from and indirectly refers to the audience members' evaluations of their previous experiences with news media (e.g., Kohring and Matthes 2007; overview in Prochazka and Schweiger 2019), and, at the same time, it is future-oriented and encompasses the audience members' expectations for the future actions of new media.

Both Strömbäck et al. (2020) and Fawzi et al. (2021) also note that the referent of trust in media may vary from a particular news outlet or journalist to media types, media institutions, or the whole media system. In our research, we were confronted with the necessity to employ a perspective that would correspond to the long-term process of the fragmentation of the media field and the media audiences (Knobloch-Westerwick et al. 2015; Tewksbury and Rittenberg 2012), and to the revival of populism and its tendency to seek out only certain specific media outlets or clusters of media outlets, and to ignore or reject other media as biased (Strömbäck et al. 2020). Therefore, we rejected trust in media types and trust in the whole news media system as unsuitable for our inquiry (Daniller et al. 2017). Instead, in this study, we tackle and empirically link two modalities of institutional trust in media. One of them represents specific trust in media and the other represents generalized trust in media (c.f. Fawzi et al. 2021). As specific trust, we assess participants' trust in concrete news media outlets (or, namely, in Czech PSM). Generalized trust refers to the news media that are considered by participants as their individually important sources of news and information, because the focus on news media that they actually use and find relevant better fits the current fragmented high-choice media environments (Daniller et al. 2017).

Moreover, the conventional ways for examining trust in news media are usually (though often implicitly) built on the assumption that trust is linked to the expectation that trusted news media will stick to professional norms and be impartial and objective (Westerståhl 1983). However, this disregards the audiences' needs that are associated with the expectation that trusted news media will represent and support in-group identities, worldviews, interests, and values (McQuail 2013). Our assumption is that trust in news media integrates both types of expectations because the in-group-oriented elements are inherent to any social communication, and we suppose that the in-group-related expectations are more vocal for populism, for which they play a constitutive role (Krämer and Langmann 2020). Therefore, in this study, we consider trust in the news media that participants considered to be their individually important sources of news and information as comprising two dimensions that draw from distinct sets of audience members' expectations regarding the future performance of trusted news media.

The first dimension, which is referred to as normative trust in media, is based on the expectation that trusted news media will perform under a professional ethos of objectivity, impartiality, and independence. In other words, normative trust involves expectations articulated by the well-founded normative theory of news media performance (Berkowitz 1997; Knudsen et al. 2021; McQuail 2013; Westerståhl 1983). Therefore, normative trust refers to news media as institutions that form a specific expert system that is specialized in the production and dissemination of news and information (Giddens 1990).

By contrast, the second dimension, which is referred to as cohesive trust in media, includes the expectation that trusted media will be in service of representation and the defense of the in-group (i.e., that “our” media will serve “us”). Therefore, cohesive trust refers to the fact that news media are trusted because they operate as platforms for the construction and reflection of shared symbolic spaces that enable the existence of imagined communities and related collective identities (Anderson 1991; Carpentier 2017; Mouffe 2005).

Regarding trust in media that comprises several subdimensions, there is a consensus among the scholars that in order to fit discriminant validation criteria, trust in media should be measured and treated as a higher-order construct (Prochazka and Schweiger 2019; Strömbäck et al. 2020). However, in this study, we treat the normative and cohesive dimensions of trust in media as distinct because we suppose that it enables us to better observe the connection between trust in media and populism. We assume that populist party sympathizers (relative to the sympathizers of other parties) expect that the media they trust will be in-group-oriented rather than impartial. So, for them, cohesive trust in individually important media will be a stronger predictor of their trust in PSM.

At the same time, and with Prochazka and Schweiger's (2019) validation argument in mind, we acknowledge that these two components form the complex attitude of trust in media only together and through interplay and that trust in media involves both types of expectations. Therefore, normative and cohesive trust are not separate or even mutually exclusive, and the conceptualization does imply that populist party sympathizers tend only to trust cohesively because they also have normative expectations (Krämer and Langmann 2020).

Public Service Media in the Czech Republic

Although trust in the media in the Czech Republic is relatively low and since the 1990s has been steadily decreasing, Czech PSM have maintained a stable position and remain the most trusted media in the country (Stetka 2020). Specifically, in 2022, 56 percent of the population trusted Czech Television and 60 percent trusted Czech Radio (Ibid.). These levels of trust are comparable to the results achieved by the British BBC (55 percent) and the Austrian ORF (62 percent) (Newman 2022; Sparviero and Trappel 2022). By contrast, PSM in other Central European countries, such as Poland (24 percent) and Hungary (31 percent), are significantly less trusted (Bognar 2022; Makarenko 2022). In addition, Czech Television, which has an average weekly reach of 55 percent of the population, is the most-watched Czech broadcaster, and its news channel CT24 has been repeatedly ranked among the most-watched public service news channels in Europe (Czech Television 2021).

Czech public service television and radio are established and regulated by laws under which they have an obligation to preserve media pluralism and to produce content that meets the democratic, cultural, and social needs of society (Act on Czech Television, 1991; Act on Czech Radio, 1991). At the same time, they are to provide verified, objective, and balanced information, and create space for the development of the cultural identity of the citizens, including members of national and ethnic minorities (Ibid.). By joining the European Broadcasting Union, both institutions committed themselves to the values of universality, independence, excellence, diversity, accountability, and innovation (EBU 2012). The independent audits conducted by EBU at Czech Television in 2015 and at Czech Radio in 2019 concluded that both institutions have managed to adequately fulfill these values of public service (EBU 2015, 2019).

Compared to the situation in Hungary and Poland, where PSM have been transformed into government media in recent years, and despite the fact, that the populists were in power in the Czech Republic between 2017 and 2021, Czech PSM still maintain their neutrality and independence. They are closer in their position to PSM in Western Europe (Horowitz et al. 2022; Tóth et al. 2022). However, Czech PSM have recently been under increased political and institutional pressure from populist politicians. For example, Tomio Okamura, leader of the populist Freedom and Direct Democracy party, has declared the intention to nationalize PSM (Bastlová 2017). The intention to encroach on the powers of Czech Television also came from the populist ANO movement, which was the strongest ruling party between 2017 and 2021. In the 2020 elections for the Czech Television Board, only candidates with ties to this party were shortlisted (IPI 2022), which sparked protests from the public and the EBU. These events, when the populist party tried to use its position to influence the PSM in its favor (Sehl et al. 2020), make the Czech situation an interesting and relevant case. They clearly reveal that, for populist parties, the PSM are not problematic for being too close to the state or the government, but rather because these media, unlike in Hungary or Poland, manage to remain independent of the state and the government, while they are considered, by the populists, as ideologically distant or hostile.

Aims and Research Questions

We explore the discrepancy in trust in PSM between people who sympathize with populist parties and people who sympathize with other political parties in order to answer the question:

We seek to answer this question by testing how the normative and cohesive dimensions of trust in the media considered by the participants to be their important sources of news contribute as predictors for the participants' trust in PSM. With the help of multigroup structural equation modeling, we compare participants who declare their agreement for populist parties to people who declare their agreement for parties that do not have elements of populist ideology. Since populist attitudes include the rejection of views other than those that are consistent with populist ideology, we expect that populist party sympathizers, in their relationship with PSM, will accentuate more strongly the cohesive (i.e., in-group oriented) component of trust in media than the normative dimension.

However, in our research, we assessed the cohesive and normative dimensions of trust only for generalized trust in the media considered by the participants as their sources for important news and information. Trust in specific media (and thus in PSM) was assessed by a more conventional one-dimensional measure (for details see below). Therefore, we observe the strengths of the cohesive and normative dimensions through the relationship between specific trust in PSM and generalized trust in individually important media.

Specifically, we hypothesize that:

Methods

Measures and Model

To answer the research questions, we constructed a model to test the extent to which the dimensions of generalized cohesive and normative trust in media (i.e., independent variables) predict trust in PSM (i.e., dependent variable). Socio-demographic characteristics are control variables. We tested the model using multigroup structural equation modeling and looked at the differences between the model for populist party sympathizers and the model for sympathizers of other political parties.

Dependent Variable

The dependent variable, Trust in Public Service Media, was constructed as the sum of the scores of two variables that measured trust in Czech public service radio and Czech public service television (i.e., “Are you familiar with the following media? If so, how trustworthy do you consider them to be?”) on a 5-point Likert scale where “1” was “Absolutely untrustworthy” and “5” was “Absolutely trustworthy” (Spearman-Brown  =  .85). The Public Service Press Agency was not included in our analysis because of its function; its role is to be a distributor of news for standard “retail” media, and, in general, their services are mediated through their subscribers such that they do not have direct access to an audience (Boyd-Barrett and Rantanen 2002: 57).

Independent Variables

The independent variables include the socio-demographic variables of gender, age, and level of education, and two variables for generalized trust in media (i.e., normative and cohesive trust).

Normative trust in the media that was considered by the participants to be their important sources of information was assessed by agreement with the following statements: “These professional media can be trusted because they work with verified information”; “These professional media can be trusted because they focus on important facts and topics”; and “These professional media can be trusted because they are independent in their work” (Cronbach's α  =  .91). The wording of the items is derived from Kohring's and Matthes' (2007) measure, though the items used in our study are thematically nonspecific because the measurement assesses the trust in news media as institutions and not trust in coverage of concrete topics. The items were measured on a 5-point scale that ranged from “totally disagree” (1) to “absolutely agree” (5).

Cohesive trust in the media that was considered by participants to be their important sources of information was assessed by agreement with the following statements: “For people like me, these professional media [which I consider to be important sources of news and information for me] are trustworthy because they show the world as it really is”; “These professional media can be trusted because they stand for political opinions that match what people like me think”; and “These professional media can be trusted because they oppose the interests and opinions that are not close to people like me” (Cronbach's α  =  .85). The items were measured on a 5-point scale that ranged from “totally disagree” (1) to “absolutely agree” (5) (Figure 1).

OPEN IN VIEWER

Figure 1.

Outline of the tested model. Note: The model was further tested for each of the groups, separately. Source: Authors.

Results

Means, standard deviations, and inter-correlations between the variables are shown in Tables A2 and A3 in the Supplementary Information file. Scores that represent cohesive and normative trust in media were computed from the sums of the items. A higher total score means a higher value for the particular variable. In the case of the variable that measures the gender of the respondent, the category of female with respect to male was used. All correlations between the independent variables and the dependent variable were statistically significant. However, the correlation between education and trust in PSM was significant only for the sympathizers of other political parties, and the correlation between trust in PSM and gender was significant only for the populist party sympathizers. Correlations between cohesive trust in media and normative trust in media were relatively high in both subsamples. None of the correlations had a higher value than .80.

In the next step, we separately investigated the difference in mean values for trust in PSM for populist party sympathizers and the sympathizers of other parties. The analysis showed that populist party sympathizers would trust PSM significantly less (M  =  3.19, SD  =  1.03) than the sympathizers of other parties (M  =  3.79, SD  =  0.9, t(2020)  =  −14.36, p  =  .00). The effect size for this analysis (d  =  .62) was found to exceed Cohen's (1988) convention for a medium effect (d  =  .50).

The software Mplus was used. We employed multigroup structural equation modeling for the analysis.

As an initial step, multigroup confirmatory factor analysis was used to assess whether cohesive and normative trust formed two separate dimensions and whether the measurement of both dimensions was invariant between the populist and the non-populist groups. The cohesive and normative dimensions of trust were treated as latent variables that were constructed through individual items. A configural model that assumed cohesive and normative trust to be two separate dimensions (with two residual correlations allowed) had a better fit (χ² ₁₂  =  45.846, p < .001; CFI  =  .996; RMSEA  =  .051, 90% CI  =  [.036, .066]) than the model that assumed that cohesive and normative trust formed one common dimension (χ² ₁₄  =  306.501, p < .001; CFI  =  .968; RMSEA  =  .138, 90% CI  =  [.124, .151]). Thus, cohesive and normative trust were distinguishable from each other, although their inter-correlations were relatively high (i.e., .833 in the populist group and .782 in the non-populist group). This corresponds to the theoretical assumption about the two-dimensional nature of trust in the media, and about the complementarity of these two dimensions, where one dimension is not interchangeable with another. Standardized factor loadings in the two-dimensional model were sufficiently high, ranging from .535 to .902. Next, a two-dimensional model that assumed scalar invariance (i.e., identical factor loadings and intercepts) between the two groups was estimated. This model led to no considerable deterioration in model fit (χ² ₂₀  =  69.274, p < .001; CFI  =  .995; RMSEA  =  .047, 90% CI  =  [.035, .060]) compared to the configural model, which suggested that both dimensions of trust had the same meaning in the two groups and could be directly compared.

Hypotheses 1 and 2 were tested by estimating an unconstrained model in which the cohesive and normative trust and control variables (i.e., gender, education, age) were set to predict trust in PSM freely across both groups (χ² ₆₄  =  196.471, p < .001; CFI  =  .986; RMSEA  =  .043, 90% CI  =  [.036, .050]). Then, we tested whether the effects of cohesive and normative trust on the trust in PSM significantly differed between the populist and the non-populist group. First, we constrained the effect of cohesive trust to be the same in both groups. A chi-square difference test showed a significant change in model fit compared to the unconstrained model (Δχ² ₁  =  6.334, p < .05), which suggested that the effect of cohesive trust differed between the groups. Second, we constrained the effect of normative trust to be identical. This again led to a significant difference from the unconstrained model (Δχ² ₁  =  17.631, p < .001), suggesting that the effect of normative trust also differed between the groups.

Standardized effects based on the unconstrained model are presented in Figure 2 (i.e., a model for populist party sympathizers) and Figure 3 (i.e., a model for other political parties' sympathizers). While the cohesive component of media trust emerged as the strongest predictor of trust in PSM in the model for populist party sympathizers (.31, p < .001), the normative component of media trust was not significant (−.05, p  =  .537). At the same time, both dimensions of media trust were significant predictors for trust in PSM for the sympathizers of other parties, with the normative dimension of trust (.44, p < .001) emerging as the strongest predictor and the effect of the cohesive dimension proving to be weak (.14, p < .05). This means that Hypotheses 1 and 2 were met.

OPEN IN VIEWER

Figure 2.

Model for the populist party sympathizers.

OPEN IN VIEWER

Figure 3.

Model for the sympathizers of other political parties.

Discussion

The main findings of our study are consistent with the theoretical assumption that the anti-pluralism of populist ideology may influence the expectations that the populist party sympathizers would have for PSM (Fawzi 2019; Fawzi and Krämer 2021; Fawzi and Mothes 2020; Krämer 2018; Krämer and Langmann 2020; Schulz et al. 2020) and provide empirical grounds for why these people trust PSM significantly less. Our analysis shows that, for respondents who sympathize with populist parties, the strongest predictor for trust in PSM is the cohesive dimension of generalized trust in media, while the normative dimension is marginal and not significant. Conversely, for respondents who sympathize with other political parties, the strongest predictor is the normative dimension of generalized trust in media and the effect of the cohesive dimension on trust in PSM is markedly weaker. In other words, populist party sympathizers' trust in PSM is linked to how much they trust “their” media because they expect that these media outlets will represent their in-group values and attitudes, while other political sympathizers' trust in PSM is more positively linked to the extent to which they trust “their” media outlets because they expect that they will be impartial and objective. Thus, we think that trust in individually important media enables us to indicate what is important to audience members when it comes to trusting news sources, including PSM. Therefore, the identified differences in the dimensions of generalized trust in media suggest that populist party sympathizers tend to trust PSM more when they believe that PSM will not conflict with their worldview, while non-populist party sympathizers trust PSM more primarily when they believe that PSM will meet their expectations for impartiality and objectivity.

The expectation that PSM should reflect and meet one's values and opinions is not necessarily in direct conflict with the mission or performance of PSM or with the expectations of PSM to be impartial: Cohesive trust does not exclude normative trust and they commonly occur together, and the requirement to represent one's perspective does not preclude an emphasis on a plurality of perspectives. Our study clearly demonstrates that cohesive trust plays a role as a predictor even for people who sympathize with other political parties, so it cannot be said to be a dimension unique only to populist party sympathizers. However, for populist party sympathizers, the cohesive expectations are dominant or principled and, at the same time, these audience members tend toward an anti-pluralist orientation that may sharply contrast with the mission of PSM to produce content for all, to include minorities, and to emphasize different points of view (Biltereyst 2004; Goyanes 2021).

At the same time, we may expect a similar situation (i.e., populist party sympathizers trust PSM only when they believe that PSM will meet their opinions and support their values) in other countries where PSM are under the constant pressure of populist politicians, such as in Austria, United Kingdom, Germany, or Sweden (Holtz-Bacha 2021; Sehl et al. 2020). For example, the German populist party AFD has labeled the German PSM as “Lügenpresse” (lying press) and accused them of suppressing debates on immigration and being biased against their party (Niemeier 2018). This rhetoric is quite similar to that of populist parties in the Czech Republic.

Similar results might be found if this study were replicated in Poland and Hungary, where, in recent years, the ruling parties have transformed PSM into government media (Połońska 2019; Polyák 2019). Their role is no longer based on the support of social inclusiveness and political and cultural pluralism, but rather on the promotion of particular populist or conservative politics. We may only speculate that, for sympathizers of populist parties in these two countries, cohesive trust will play an even more important role as a predictor of trust in PSM, while, for sympathizers of other parties, the effect of normative trust will increase due to the perceived loss of independence of PSM. Thus, the normative dimension of trust will place more emphasis on the perceived loss of the independence of PSM.

The strength and eventual straightforwardness of these conclusions are limited to some degree by the nature of the measures used in the survey data set and the measures employed in the analysis. First, and as noted above, the cohesive and normative dimensions of trust were not measured in relation to particular news media outlets (PSM included). Instead, we measured them in relation to the news media that the respondents considered to be their important sources of news and information. This decision was primarily motivated by the need to decrease the length of the questionnaire and to reduce the respondents' burden.

Second, an inclination to populist attitudes was not assessed with a scale that would enable the measurement of the strength of these attitudes. This was measured through respondents' declarations of the most agreeable political party (i.e., through a binary variable). The questionnaire included two types of this measure: one assessed respondents' electoral preferences, and the other asked for the most agreeable party. We considered electoral preferences to be less valid because, unlike reporting the most agreeable party, respondents' intention to vote for a populist party does not necessarily indicate that they have populist attitudes (Akkerman et al. 2014; Elchardus and Spruyt 2016; Stanley 2011).

And, third (consequent to the second point above), our analysis does not include respondents who did not record their agreement for any of the mentioned parties.

As a result, our model does not aim to provide a more nuanced and direct analysis of the relationship between populist attitudes and trust-related attitudes to PSM. Yet, we believe, the accuracy of the model indicates the general distinctions in trust between respondents who explicitly agree with populist parties and the respondents who explicitly agree with non-populist political parties.

At the same time, the results are open to interpretations that reach beyond our main theoretical hypotheses and focus on the performance of Czech PSM. Some studies suggest that the problem could be linked to the populists' sense that they are under-represented in PSM (Fawzi 2020) or that PSM tend to be more negative toward populists than to other politicians (Esser et al. 2016; Salgado et al. 2021). It is fair to admit that the populist party sympathizers' emphasis on the representation of their worldview may not only be an expression of their anti-pluralism, but it can also stem from shortcomings in the performance of Czech PSM, which results in situations where populist party sympathizers expect something that PSM can, but does not, provide. Thus, the low trust of certain segments of Czech society in PSM may result from PSM's inability to sufficiently or appropriately address the needs of these audience members as legitimate. In such cases, PSM may manage to throw the baby out with the bathwater when they do not sensitively separate these legitimate needs from some of the illegitimate manifestations that accompany their articulation (e.g., hate speech, racism) and reject them, ad hoc. As a result, as the demand for populist politics increases, the populists' trust in PSM decreases. However, the link between PSM performance and the audience members' sense of (under)representation and their trust in PSM requires further and more detailed empirical attention, ideally in the form of a study that would map the expectations that the Czech audience members have for Czech PSM and assess the extent to which these expectations are and could be met.

The data for our study capture the attitudes of respondents before the outbreak of the COVID-19 pandemic and before the start of the war in Ukraine. The influence of these events on the relationship of populist attitudes and trust in PSM can therefore be assessed only speculatively. Confidence in the Czech public service television decreased by 4 percentage points year-on-year in 2022 (Stetka 2022). This could be due to the fact that attitudes toward the COVID-19 pandemic polarized Czech society, while those who were convinced of the non-fatality of the disease tended to have less trust in the media (Ibid.; Mihelj et al. 2022) and at the same time to have stronger populist attitudes (Eberl et al. 2021). With this argument in mind, we can assume that if the data were collected in 2022, the difference in trust in PSM and in the strength of the effect of cohesive trust would be even more pronounced between the sympathizers of populist parties and the sympathizers of other parties than in our study.

Supplemental Material