The Watchdog Press in the Doghouse: A Comparative Study of Attitudes about Accountability Journalism, Trust in News, and News Avoidance

The watchdog role has been one of the most widely discussed normative functions of the press. In this study, we examine the public’s attitudes toward the news media’s watchdog performance and how they correlate with trust in news and news avoidance, two important phenomena for democracy and the health of the public sphere. We further examine how individual predispositions (e.g. political interest, ideology) and contextual variables (e.g. press freedom) moderate these relationships. Based on data from the 2019 Reuters Institute Digital News Report, and controlling for a range of factors, we find that across 38 countries, watchdog performance evaluations are positively associated with trust in news but that they are also positively associated with higher levels of news avoidance. Last, we find that evaluations of media in other functions like helping citizens understand the most important topics of the day and choosing relevant topics were more strongly associated to trust in news and lower news avoidance levels than watchdog performance evaluations.


Introduction
The watchdog role has been one of the most widely discussed normative functions of the press.This "long-established liberal conception of the news media as the fourth estate," as Norris (2012) describes it, conceives of the press as "an independent guardian located in civil society" that serves as a counterbalance to powerful institutions in the executive, legislative, and judicial branches in government but also those in the private sector as well.Scholars have often differed in their assessments of how well various forms of journalism carry out such an idealized mission (Donohue et al. 1995;Waisbord 2000), do so in reactive rather than proactive ways (Donsbach 1995) or in ways that vary across countries according to differences in journalistic culture (Hanitzsch 2007), but there is little denying its central place in both journalists' professional identities and journalism studies scholarship on the role of news in society.
While there has been a justified emphasis on the normative definitions of watchdog journalism, on how journalists think about the watchdog role, and on differences in the content of watchdog journalism, research looking at how audiences perceive watchdog journalism has been much more limited.While studies have occasionally examined audience expectations about journalism, including how much people say they value watchdog journalism as content or as an essential journalistic norm (Fawzi and Mothes 2020;Loosen et al. 2020), we still know little about differences among audiences in terms of how they view the watchdog role and how much weight they place on it relative to other dimensions of journalism such as its role in providing up-to-date information, directing attention to relevant topics, or analysis about the world.Journalism scholars and practitioners sometimes assume that audiences want the news media to hold powerful leaders and institutions to account, but audiences may not always view the actions of unelected, crusading journalists in such uniformly positive ways.Others may simply see such forms of journalism as secondary compared to other functions they desire from news in their country.
In this study, we examine the public's evaluations of watchdog role performance across more than three dozen media markets.Specifically, we examine what role these evaluations play in relation to trust in news and news avoidance, two variables that serve as indicators of how much the public values journalism both overall and in their everyday life.Academic research has been instrumental in showing the importance of trust in the news for both journalists and democracy more broadly (Hellmueller et al. 2016;Fawzi et al., 2021;Fink, 2019), as well as the positive link between news consumption for individuals and the ability to engage effectively in their political systems (Aalberg and Curran 2012;Carpini and Keeter 1996).However, no prior research has specifically considered what role, if any, attitudes about the watchdog performance of the press plays in relation to these attitudes and behaviors.
The present study tests the nature of these relationships while comparing watchdog performance evaluations alongside audience evaluations of other functions of the press such as providing up-to-date information or shining a light on worthwhile topics.We further test individual-and country-level variables that could affect these relationships, given previous studies that have shown that both the watchdog role and our main dependent variables are context-dependent (Hanitzsch et al. 2018;Márquez-Ramírez et al. 2020;Toff and Kalogeropoulos 2020).We do this by drawing on comparative survey data from the Digital News Report 2019 (Newman et al. 2019) which covers audiences in 38 markets across five continents.We find that (a) watchdog performance evaluations are positively associated with trust in the news but also positively associated with intentional news avoidance (albeit weakly); (b) evaluations of media in other functions like curation (choosing relevant topics for audiences) and offering analysis (helping people understand these topics) were stronger positive and negative correlates of trust in news and news avoidance, respectively, compared to watchdog performance evaluations; and (c) the link between watchdog performance evaluations and trust in news is significantly moderated by political interest and ideological extremity.In other words, the watchdog role of the press appears to be most salient mainly to the segments of the public who are most highly engaged with politics.Furthermore (d), we find no evidence that cross-country differences in the relationship between watchdog attitudes and trust in the news are explained by relative levels of press freedom or differences in the importance that journalists themselves ascribe to the watchdog role in their countries' media system.But we do find (e) citizens in less free countries avoid the news more often the more they think journalists scrutinize powerful people as watchdogs, a finding that could be related to differences in perceptions about the quality of watchdog journalism between countries.Overall, we find that while watchdog performance evaluations are important for what people think about the news media, they are not as strongly or consistently correlated with news avoidance behaviors, particularly when compared with other roles.
In the next sections, we review the scholarly literature on the watchdog functions of the press and known correlators of trust in news and news avoidance, and we specify our hypotheses and research questions on the basis of this existing research.We then proceed to present the data and the rationale for the strategic sample of the countries we focus on as well as the measures used.This is followed by our results and a concluding discussion which returns to the under-explored matter of audience perceptions about the watchdog function of the press.

The Watchdog Function of the Press
The watchdog function is one of the most important and discussed functions of the news media.The watchdog metaphor implies that journalists should act as guards against powerful groups in society.Bennett and Serrin (2005) define watchdog journalism as "1) independent scrutiny by the press of the activities of the government, business and other public institutions, with an aim toward 2) documenting, questioning, and investigating those activities in order to 3) provide publics and officials with timely information on issues of public concerns" (p.169).The news media as the "fourth estate" is one of the most important checks and balances in the division of powers in a democracy, as well as crucial for the protection of civil and political rights (Norris 2000).Academic research on journalistic roles has classified the "critical-monitorial" function of journalism as one of six core functions of the profession (Hanitzsch and Vos 2018).According to the authors, this function encompasses three roles: the monitor role, which is related to being observers of power in a society; the detective role which is related to investigative reporting and active research of powerful conduct (e.g., fact-checking), and lastly the watchdog role which is related to even more active scrutiny of power in a society.Scholars do not necessarily view all watchdog journalism in a positive light, however.Journalism that relentlessly attacks others has also sometimes been criticized as "junkyard dog" journalism (Sabato 1991) or a "burglar alarm that keeps on ringing" (Bennett 2003).
Other than academics, journalists also tend to consider the watchdog role as one of the profession's most valued priorities (Hanitzsch et al. 2011).That said, there is considerable variation in the degree to which journalists subscribe to a watchdog-oriented professional philosophy-especially in non-Western countries, like China, Uganda, or Indonesia where many journalists do not hold such views (Hanitzsch 2011).Márquez-Ramírez et al. (2020) looked at the presence of watchdog articles in prominent newspapers in 18 countries and found more watchdog articles in the press in more democratic countries.However, as the authors point out, the presence of a strong watchdog press is not necessarily purely positive for pluralistic democracy, given that crusading journalism can be used by news media owners for narrow partisan aims or to steer away attention from more relevant news topics (Márquez-Ramírez et al. 2020).
What do audiences think about watchdog journalism?Research findings are mixed.A recent study in Germany suggests that the news media's watchdog function is at the core of citizens' expectations of the news media (Fawzi and Mothes 2020).When asked to evaluate different roles and functions of the press, German citizens ranked revealing political abuses and scandals as the most important ones.This study follows previous similar findings in other Western European democracies like the Netherlands (Van Der Wurff and Schoenbach 2014).Audiences in Germany are also more likely to consider the watchdog role of the news media as important compared to even professional journalists (Loosen et al. 2020).However, there are also indications that citizens in some countries may be repelled by aggressive watchdog coverage and more likely to endorse a "good neighbor" ideal where audiences prefer news that serves the public by responding to everyday societal problems rather than news that fixates on holding powerful figures accountable (Costera Meijer 2010;Heider et al. 2005).
Given the variation in importance that audiences place on the watchdog function and how highly they rank its role conception, this study seeks to assess not only watchdog performance evaluations but what relationship such attitudes may have on two additional variables: the degree to which people trust news and how frequently they say they avoid it.

The Watchdog Role and Trust in News
First, we consider the relationship between watchdog evaluations and trust in news.Trust in news is not only important for journalists and news organizations, but also for democracy.While prior research has shown a close relationship between trust in the news media and how the public tends to view other political institutions-what Hanitzsch et al. (2018) refer to as the "trust nexus"-trust in at least some sources of professional news helps people effectively navigate the complex digital media landscape and political information environment (see e.g., Toff et al. 2020).Ladd (2010) has shown, for example, that those who lack trust in the news are more likely to rely on their partisan predispositions when making evaluations of politicians.As a result, the positive effects of keeping up with news (e.g., learning about politicians' policies and evaluating them) are partly dependent on trust in news.In a qualitative study examining the relationship between trust in news and how people gathered and made sense of information during the COVID-19 pandemic, Ternullo (2022) shows how those with generalized skepticism toward all news struggled to form opinions and therefore engage in effective political action.
Research looking at how watchdog role evaluations impact trust in news is limited.Jebril (2013) found that watchdog performance evaluations had a highly positive effect on satisfaction with political news coverage in Denmark, Britain, and Spain, a concept close to but not necessarily the same as trust in news.A recent study looking at how different news media functions predict trust in news in the United States found that a belief in the adversarial role of the news media towards the government and businesses is strongly and positively correlated with news media trust, while belief in the contextual role of the news media (a role that includes the press investigating government claims) was not found to have a significant relationship with audience trust in the news (Abdenour et al. 2020).
On the one hand, if audiences perceive watchdog journalism to be a core function of the news media, perceptions about the failure of the press to perform along these lines should go hand in hand with audiences saying they lack trust in the news in their country.That leads us to our first hypothesis (H1): news audiences' watchdog performance and positive evaluations will be positively associated with trust in news.
On the other hand, trust in news is shaped by a multitude of factors beyond simply attitudes about the watchdog role of journalism.Research has shown that trust in news is also shaped by variables ranging from individual-level factors around the propensity to trust (Jackob 2012) to characteristics of the media environment (Fletcher and Park 2017;Kalogeropoulos et al. 2019;Tsfati and Ariely 2014) and political contexts in which media organizations are embedded (Hanitzsch et al. 2018).Constructivist approaches to understanding trust in the news have also pointed to the importance of production practices like transparency and authenticity as well as communitycentered initiatives that focus on inclusion and representation within newsrooms so that coverage can be more responsive to the everyday concerns of the audiences media organizations seek to serve (Coleman et al. 2012;Schmidt et al. 2019).This work also underscores that trust is a relational concept, established over time through a range of other normative functions journalism provides individuals and communities.For instance, according to Hanitzsch and Vos (2018), the informative-instructive function of journalism is related to providing accurate information about current affairs to citizens.The analytical-deliberative function includes the analyst role related to providing in-depth explanatory information about current events.Another typically discussed role of journalists is related to curation, or "setting the agenda of public opinion" (Christians et al. 2009), which from the audience's perspective can be interpreted as choosing relevant topics to cover.
These additional journalistic roles, as well as other perceived characteristics of news media, including how prone to the negativity it tends to be (Lengauer et al. 2012;de Bruin et al. 2021), are thought to have an impact on how much audiences engage with news and political information (Valentino et al. 2001) and therefore trust as well.This leads us to our first research question (RQ1): What is the relationship between watchdog performance evaluations and trust in news relative to other attitudes about the performance of the press?
The Watchdog Role and News Avoidance Next, we examine the relationship between watchdog evaluations and news avoidance.News consumption is important for political knowledge, political participation, and for keeping the electorate informed about policies and the affairs of the state (Carpini and Keeter 1996;Curran et al. 2009).A growing number of related studies, however, have begun to examine news avoidance as a potential democratic problem (for a review, see Skovsgaard and Andersen 2020) and one that may be growing, particularly among less advantaged and politically engaged subgroups (Karlsen et al., 2020;Strömbäck et al., 2013).Growing media choice is thought to make it easier for those less interested in news to opt-out of consuming it (see also Prior 2007), but declining levels of engagement with the news may make it harder for news avoiders to make informed political decisions.
Given the emphasis that citizens in Western democracies place on the watchdog role of the news media as demonstrated by prior studies (Fawzi and Mothes 2020;Van Der Wurff and Schoenbach 2014), one would expect that watchdog performance evaluations might be inversely related to levels of news avoidance.Indeed, Toff and Kalogeropoulos (2020) show that news avoidance internationally is strongly correlated to press freedom: people in countries with low levels of press freedom showed higher levels of news avoidance, a finding that could be related to the perceived ability of the press to perform their watchdog role adequately.This leads to our second hypothesis (H2): news audiences' watchdog performance positive evaluations will be negatively associated with news avoidance.
However, Toff and Kalogeropoulos (2020) did not examine the reasons behind news avoidance; the study only demonstrated that a relationship existed at the country level between levels of news avoidance and press freedom.In fact, other factors, including genre preferences, held even greater explanatory power when it came to understanding variation in news avoidance.In other words, even if watchdog performance evaluations are an important factor in explaining some avoidance of news, they may be less important than other factors.Qualitative studies of news avoidance further underscore this possibility.Palmer et al. (2020), for example, find that many habitual news avoiders in the United Kingdom and in Spain see news and politics as intertwined institutions that many remain broadly skeptical of the watchdog norm altogether but also largely dismissive of journalism as holding any relevance to their lives.Previous studies have also highlighted concerns about the tone of news and the constant refrain of "doom and gloom" (Palmer et al. 2020;Schrøder and Ørsten 2016) as another major factor driving avoidance, which is closely connected to journalistic negativity (Lengauer et al. 2012) and its detrimental effects (Valentino et al. 2001).In other words, it is difficult to assess to what degree watchdog performance evaluations might shape news avoidance or whether other attitudes about news might outweigh such concerns.This leads to our second research question: (RQ2): What is the relationship between watchdog performance evaluations and news avoidance relative to other attitudes about the performance of the press?

Individual-and Country-Level Differences
Apart from examining the correlation between watchdog performance evaluations with trust in news and news avoidance, we are interested not only in characterizing the overall relationships but also in how individual-and country-level differences may relate to these dynamics.At the individual level, how people evaluate the performance of journalists in their various roles is linked with how they value these different journalistic roles normatively, which may also vary by country.In addition, people with different individual backgrounds may value the importance of the watchdog norm with different standards in mind and/or subsequently assess their country's news media performance differently.Thus, an important question is, for which individuals watchdog performance evaluations are most closely related to levels of trust and news avoidance?
These individual-level differences are particularly important to consider because the watchdog role of the press is also one of its most politicized functions.For example, Fawzi and Mothes (2020) show that in Germany, those with a strong left ideology and those with high levels of political interest are more likely to expect the news media to perform their watchdog role.Considering the politicized nature of the watchdog role when compared to other journalistic roles, we hypothesize that (H3a) the relationship between watchdog performance evaluations and trust in the news will be stronger among those who are more politically interested, and likewise (H3b) the relationship between watchdog performance evaluations and news avoidance will also be stronger among those who are more politically interested.Looking more specifically at ideology, we further hypothesize that (H4a) the relationship between watchdog performance evaluations and trust in the news will be stronger among those who have strong left-or right-wing ideology and likewise, (H4b) the relationship between watchdog performance evaluations and news avoidance will also be stronger among those who have strong left-or right-wing ideology.
In addition, we ask whether the strength of the association between watchdog performance evaluations and both trust in news and news avoidance might vary across countries in systematic ways.We are interested in two factors.The first is press freedom (as explored in Toff and Kalogeropoulos 2020).Press freedom creates an environment in which journalists are able to hold powerful politicians and businesspeople accountable for their actions.In fact, scholars have pointed toward a causal relationship between a strong free press and lower levels of corruption (Brunetti and Weder 2003).In addition, we are interested in differences in the journalistic culture across countries as operationalized through professional norms, namely how journalists think about their watchdog role conception and whether they value it in their work.Hanitzsch (2019) differentiated between four different categories of journalistic roles: normative (what journalists ought to do), cognitive (what they want to do), practiced (what journalists do in practice) and narrated (what journalists say they do).We know from previous research that there is variation in how journalists conceive and perform these roles in different countries and different media systems (Brüggemann et al. 2014;Hellmueller and Mellado 2016;Mellado et al. 2020).One expectation is that in freer countries and in countries where journalists more heavily value watchdog journalism as important to their overall role conception, audiences may also place greater weight on the importance of watchdog journalism more so than in countries where watchdog journalism is comparatively less emphasized.However, the reverse is also possible.Audiences may value watchdog journalism even more so in countries where it is relatively absent, whether due to a lack of press freedom or because it is not part of the journalistic culture, or both.Given the lack of clear research findings from previous literature and the potentially ambivalent role audiences may have toward watchdog journalism, we pose a third research question: (RQ3): How do these relationships vary in different countries with (a) different levels of press freedom and (b) different journalistic cultures?

Methods
In this section, we describe the data and the methods used to test our hypotheses and gather evidence with respect to our research questions.We employ a comparative survey design, drawing on data from the 2019 Digital News Report survey (Newman et al. 2019), a large-scale study focused on news consumption and attitudes toward news, conducted annually by the Reuters Institute for the Study of Journalism and fielded by YouGov.The 2019 version of the study was conducted in 38 markets with approximately 2,000 respondents each. 1

Dependent Variables
The main dependent variables in this study are (a) trust in news and (b) frequency of self-reported news avoidance.To measure trust, survey respondents were asked to denote whether they agreed or disagreed with the statement, "I think you can trust most news most of the time" (using a 1-5 Likert scale).To measure avoidance, respondents were asked how often they found themselves "actively trying to avoid the news these days" (coded on a 0 to 3 scale corresponding to "Never," "Occasionally," "Sometimes," and "Often").Early in the survey questionnaire, respondents were provided with a definition suggesting that "by news, we mean national, international, regional/local news, and other topical events accessed via any platform (radio, TV, newspaper, or online)."

Independent and Control Variables
The primary independent variable used in this study is an evaluation of watchdog role performance.This question was asked as one of five statements about functions of the news media with which respondents were asked to state their level of agreement.The item for the watchdog role performance evaluations was "The news media monitors and scrutinizes powerful people and businesses."Responses were coded on a 1-5 point Likert scale.Average levels of watchdog role performance evaluations are summarized in Figure 1, which shows substantial variation within countries, in some cases more so than between them.The highest levels of watchdog performance evaluations were found in Brazil while the lowest was in Hungary.
The four additional evaluations about other functions of the news included statements about the curation and analytical roles of the press (as in Hanitzsch and Vos 2018) and attitudes about journalistic negativity (Lengauer et al. 2012), which has been examined in light of its negative effects on audience engagement (Valentino et al. 2001).Respondents were asked whether they agreed or disagreed (measured on the same 1-5 Likert scale) with the statements: "The news media keeps me up to date with what's going on," "The topics chosen by the news media do not feel relevant to me," 2 and "The news media helps me understand the news of the day"-each of which taps into audience perceptions about the performance of journalists in setting the agenda, selecting what it worth paying attention to, and helping to explain complex phenomena.The item "The news media often takes too negative view of events" 2 is related to perceptions that news may be overly pessimistic.Overall, these statements measure the public's evaluations of the news media in a non-exhaustive list of crucial journalistic roles in public life.By comparing the strength of associations between trust in news and news avoidance alongside watchdog role performance evaluations, we can better assess how important these factors may be for some respondents versus others.The bivariate correlations between all performance evaluations can be found in the Supplementary Information file in Table A-4.We also included a number of additional independent variables related to political attitudes.Respondents were categorized as holding a "strong" ideological outlook on the basis of their self-reported placement on a 7-point left/right scale, collapsing responses on the very left and right (1/2 and 6/7 in the strong ideology category) and grouping those who said they were in the three center values or don't know in other category.Political interest was measured by asking "How interested, if at all, would you say you are in politics?"and responses were coded on a 1-5 Likert scale, from "Extremely interested" to "Not at all interested."Last, we included three sociodemographic variables as controls: age (m = 45.8,SD = 16.2),gender (52.2% females), and educational attainment (m = 5.94, SD = 1.84, asked on a 1-10 scale ranging from "I did not complete any formal education" to "Doctoral degree").
To examine country-level variation in the role of watchdog evaluations for correlations between trust in news and news avoidance, similar to Toff and Kalogeropoulos (2020) we included a variable for press freedom, as measured in 2019 by Reporters Without Borders, which uses an expert questionnaire and data on journalistic abuses (Reporters Without Borders 2019) for all markets included in our data.We further included a measure of the importance that journalists in different countries place in the watchdog role in 35 of our 38 countries 3 , taken from the Worlds of Journalism Study (The Worlds of Journalism Study 2019), collected between 2012 and 2016.The measure is the country-level average of responses on a 1-5 scale to two questions asked to journalists: "Please tell me how important each of these things is for your work: Monitor and scrutinize political leaders; Monitor and scrutinize business." 4

Effects of Watchdog Performance Evaluations on Trust in News
To examine the role of watchdog performance evaluations on trust, we estimated OLS models with levels of trust in the news as the dependent variable across the pooled data, both with and without controls for sociodemographics (age, gender, and education), political interest, and ideological strength (the full model output is summarized as Model 1b in the Supplementary Information file). 5In Figure 2, we plot the coefficients for all five media performance evaluations.We find that positive watchdog performance evaluations are significantly associated with higher levels of trust in news (b = .25,p < .001)confirming our first hypothesis.The coefficient was significant also when including other performance evaluations and controls (b = .09,p < .001).
A one-unit increase in the five-point watchdog performance scale was associated with a .09increase on the five-point trust scale.
When compared with attitudes toward other functions of news (RQ1), the coefficient for watchdog role performance evaluations was comparable in magnitude to attitudes about whether the press used an appropriate tone when covering events (b = .07,p < .001)and larger than the coefficient for topic relevance (b = .02,p < .001).However, a perception that the news media help people keep up to date with key events (b = .23,p < .001)and help them understand the news of the day (b = .26,p < .001) held stronger explanatory power and independent variable of trust in news.Both variables explained more than twice as much variation in trust in the news compared to watchdog role performance evaluations, suggesting that meeting audience expectations around more basic explanatory functions of journalism like keeping the public informed and up-to-date were potentially more important on average for trust than views about the press' performance scrutinizing powerful institutions.
The third analysis we performed sought to examine the degree to which political interest (H3a) and political ideology (H4a) moderated the relationship between watchdog performance evaluations and trust in the news.We estimated an additional model with trust in the news as the dependent variable that included one interaction term for political interest and another one for ideological strength.(Model 1c in the Supplementary Information file).We find that both political interest (b = .03,p < .001)and ideological strength (b = .02,p < .001)significantly and positively moderate the effect of watchdog performance evaluations on trust in news, as seen in Figures 3 and 4. In other words, watchdog performance evaluations were important correlators of trust in the news for those most interested in following political affairs and those with strong left-or right-wing ideology.

Effects of Watchdog Performance Evaluations on News Avoidance
Next, we estimated a similar set of models but instead focused on news avoidance instead of trust.In Figure 5, we plot coefficients for the five different performance evaluations, controlling again for the same sociodemographic and political variables as in the previous section (see Model 2b in the Supplementary Information file).These models also include respondents' trust in news in keeping with previous news avoidance studies (Avery 2009;Toff and Kalogeropoulos 2020).In fact, while we do find a significant negative relationship between trust and avoidance, contrary to our second hypothesis (H2) that watchdog role performance evaluations would be inversely related to news avoidance, we instead find evidence of a slightly positive coefficient (b = .01,p < .01)when including all other attitudes.This relationship, however, is negative when control variables are not included (−.05, p < .001,as seen in Model 2a), suggesting only a weak association between watchdog evaluations and news avoidance.In addition, when the relationship is examined without control variables, the model fit is particularly poor (R 2 < .01).In comparison, all four other evaluations about the news media were negatively correlated with higher levels of news avoidance (RQ2).In other words, the higher the evaluations of news media performance in these other aspects, the respondents avoided news less frequently.The strongest correlation of these evaluations with news avoidance was a perception that the news media succeed in keeping respondents up-to-date (b = −.11,p < .001),suggesting that, even more so than trust, news avoidance in the aggregate may be even less correlated with political attitudes when it comes to views about the watchdog performance of the press and more so with meeting the public's basic information needs.
With respect to H3b, we also found a positive interaction between political interest and watchdog performance evaluations.Those with higher levels of political interest were more likely to avoid news when perceiving the media do a good job in their watchdog role (b = .02,p < .001),whereas those with low levels of interest were less likely to avoid it due to high watchdog performance evaluations as illustrated in Figure 6 (full output on Model 2c in the Supplementary Information file).We did not find a significant interactive effect for the relationship between news avoidance and ideological strength (H4b) (b = .00,p > .05).
Our last set of models focused on investigating country-level differences.We estimated four additional models-multilevel regressions with trust in news and news avoidance as the dependent variables (Models 3a-4b in the Supplementary Information file, respectively) to examine our final research question (RQ3).These models include random effects by country as well country-level measures of press freedom and journalist attitudes about the importance of the watchdog role.
Overall, we find that a country's level of press freedom was not significantly associated with levels of trust in the news (as it is in Kalogeropoulos et al. 2019) nor is it a significant moderator for the relationship between audience evaluations of watchdog role performance and trust in the news (Models 3a and 3b).Journalistic evaluations of the importance of the watchdog role in their country were not associated with either trust in news or news avoidance.
Significant but small cross-country differences were found, however, with respect to the relationship between press freedom and news avoidance (Models 4a and 4b).In line with Toff and Kalogeropoulos (2020), we find that higher levels of press freedom are associated with less frequent news avoidance (b = −.01,p < .001).We also find that the relationship of audience watchdog evaluations on trust in news is negatively moderated by press freedom (Figure 7).In other words, in freer countries (like Norway), more positive watchdog performance evaluations are weakly but negatively associated with more frequent news avoidance, whereas this relationship is reversed in places with lower press freedom (such as Turkey) where the perception that news organizations are watchdogs is positively associated with more frequent news avoidance.We find no significant country-level effects associated with the importance that journalists ascribe to the watchdog role in their country.

Discussion
This study examines the relationship of watchdog performance evaluations on trust in news and news avoidance frequency using a comparative sample of 38 global media markets.First, we demonstrate that watchdog performance evaluations are a significant factor in whether people trust news across a wide range of democracies, but we also show that this is particularly the case for those who are politically interested or have a strong left/right ideological outlook.These important individual-level differences underscore that the watchdog role may not be important to all or even most of the public when it comes to whether they trust news in their country but mainly for the segment of the public that pays the most attention to political elites and may be more likely to expect news media to do so, too.
Our findings regarding the relationship between watchdog role performance evaluations and trust in the news are in line with previous research showing that in general audiences do in fact ascribe importance to the watchdog role of the press (Fawzi and Mothes 2020;Van Der Wurff and Schoenbach 2014), not just when asked about it but also when evaluating how much they trust the news.However, for most people, watchdog performance evaluations may not be as influential for trust as other functions of journalism, like the curatorial or analytical roles.This finding extends previous studies that suggested that the watchdog role may be less important for audiences than other roles, for instance, those related to responding to concerns in people's everyday lives (Costera Meijer 2010;Heider et al. 2005).
Second, our findings regarding the relationship between watchdog performance evaluations and news avoidance are more complicated.Overall, we find that in many cases news avoidance may be less correlated with negative watchdog role performance evaluations and instead more closely related to nonpolitical concerning other aspects of news coverage such as the press' performance in curating topics or providing up-to-date information, as explored in previous research (Mummolo 2016;Schrøder 2019).
There are several potential explanations for the finding that watchdog evaluations were less strongly correlated with explaining news avoidance than with trust in news.First, we note that trust in news is itself a significant covariate in these models, so to the extent that negative watchdog role performance evaluations reduce trust, it may also indirectly influence news avoidance but almost entirely through these attitudes.All else equal, evaluations about watchdog performance had little independent effect on news avoidance or even a slightly positive one.Second, it is quite plausible that many citizens may want the news media to perform this role in the abstract but they do not necessarily want journalists scrutinizing powerful figures and institutions in their communities-as Meyer (1988) pointed out in an early study of news credibility.Another explanation may be related to the nature of the survey question, which asked specifically about whether news outlets performed as watchdogs, not about how they thought normatively about such a role.People who think that the media are scrutinizing powerful elites may at the same time conceive some aspects of this role as non-desirable.As pointed out by Márquez-Ramírez et al. ( 2020), the watchdog role can at times be harmful to democracy; for example, corrupt media owners or politicians in charge of state-controlled media may use the watchdog role of the press as a justification for attacking opponents or diverting attention from some topics, as pointed out previously by Sabato (1991).This may also explain the finding that watchdog news evaluations are more positively related to news avoidance for more politically interested citizens who pay more attention to politics and their coverage.Last, survey respondents could more easily evaluate news media on news media characteristics that are more related to their everyday news use.
Third, when looking at macro-level variables like press freedom and the importance that journalists place on watchdog journalism in their countries, we find no significant relationships related to trust in news, but small conditional effects related to news avoidance.Differing levels of freedom to conduct watchdog journalism or different journalistic cultures around the importance of watchdog journalism do not appear to impact audience trust in news beyond that which is explained by audience evaluations about the performance of the press as watchdogs.However, we do find that in contexts with low press freedom (for instance, countries like Mexico, Turkey, or Greece in our sample) people are more likely to avoid news if they perceive the news media to scrutinize the powerful.This finding may again be related to corruption or perceptions that equate such practices with forms of undesirable attack dog journalism.They may also be a reflection of a desire by many in such contexts to simply avoid news about political affairs.
Overall, our findings further indicate that despite the emphasis that academics have placed on the watchdog ideal and its importance for trust in news and democracy more broadly, it may be less important for what many audiences actually want and expect from the news.Accountability journalism may be a crucial democratic function of the press, but it may not be as important to bringing reluctant audiences back into the fold as journalists or journalism studies scholars sometimes think it is.
Our study has at least three limitations.First, as noted above we examine watchdog role performance evaluations but not watchdog role expectations; we are only able to infer the importance respondents place on the role when making assessments about news or choices about their own media behaviors by examining how well such evaluations correlate with these other variables.A more direct measure would be preferable.Second, we measure our key concepts by employing a survey that has examined them using single items.More extensive batteries of questions capturing perceptions about individual news outlets and news media, in general, might provide more reliable and consistent measures subject to less measurement error.As with many comparative studies, we trade depth in our measures for breadth in our focus on multiple global media markets.Third, since we rely on cross-sectional data, we do not examine a potential causal relationship between watchdog performance evaluations and trust in news/news avoidance or the direction of these relationships.Future studies could examine these using panel or experimental data.
Future research could look not only at watchdog role performance evaluations but also at audience expectations around watchdog role conceptions and how they relate to trust in news and news avoidance, particularly across an even wider range of countries and journalistic cultures.In addition, qualitative research that looks at the more complex relationship between watchdog evaluations and news attitudes and behavior can give important insights and shed more light on these relationships.

Figure 1 .
Figure 1.Average levels of watchdog role performance evaluations.Error bars denote standard deviation.

Figure 2 .
Figure 2. Effects of media performance evaluations on trust in news.Coefficients are depicted with 95% confidence intervals.Full output on Supplementary Information file (Model 1b).

Figure 3 .
Figure 3. Marginal effect of the watchdog evaluations on trust in the news for different levels of political interest with 95% confidence intervals.

Figure 4 .
Figure 4. Marginal effect of the watchdog evaluations on trust in the news for different levels of ideological strength with 95% confidence intervals.

Figure 5 .
Figure 5. Effects of media performance evaluations on news avoidance.Coefficients are depicted with 95% confidence intervals.Full output on Supplementary Information file (Model 2b).

Figure 6 .
Figure 6.Marginal effect of the watchdog evaluations on news avoidance for different levels of political interest with 95% confidence intervals.

Figure 7 .
Figure 7. Marginal effect of the watchdog evaluations on news avoidance for different levels of press freedom (left) (full output on model 4b in the supplementary information file).