Framing responsibility of corruption scandals : do different frames lead to different levels of trust?

(1)

Framing responsibility of corruption scandals: do different frames lead to different levels of trust?

Thales Trench de Camargo

Student number: 11896647

Master’s thesis graded as partial examination for the Erasmus Mundus Master’s in Media, Globalisation and Journalism programme at the University of Amsterdam; Faculty of Social and Behavioural Sciences; Graduate School of Communication; Department of Communication Science. Supervised by Lukas P. Otto in Spring 2018.

(2)

Framing responsibility of corruption scandals: do different frames lead to different levels of trust?

Thales Trench de Camargo

Abstract: This study examines the effects of media frames on different political actors against the backdrop of mediated corruption scandals. Drawing on data from an experiment conducted among Brazilians, the author demonstrated that trust in government may have a spillover effect in other political institutions. A positive correlation with trust in Congress and trust in the president was also observed. Yet, exposure to different types of responsibility frames did not lead to significant effects in the level of trust in government, political parties and incumbents. Individuals who paid attention to scandals on television and in newspapers also did not have significant differences in the level of cynicism. These findings are discussed in the light of extant knowledge about framing effects and political scandals.

Keywords: Framing effects, political trust, cynicism, corruption scandals

Introduction

For decades, political scientists have been motivated to understand declining levels of public trust in government (Chanley, Rudolph, & Rahn, 2000). As these changes were

observed for the first time in the United States in the 1960’s, scholars became concerned that the phenomenon could challenge the quality of representative democracy (Van der Meer, 2017). Initial studies have considered policy based satisfaction (Miler, 1974) and incumbents’ performance (Citrin, 1974) as possible explanations for the issue. Adding to the debate, later developments pointed out that economic factors (Citrin & Green, 1986; Hetherington, 1998),

(3)

social-cultural circumstances (Mansbridge, 1997) and political actions (Citrin & Green, 1986; Hetherington, 1998) were also related to the subject. Although scholarly inquiry has already provided several accounts on this matter, it is crucial to shed light on the role of media and further discuss how it can influence political trust.

It is fundamental to look at news media when it comes to understanding what causes alterations in public attitudes (Avery, 2009). A growing body of literature has provided empirical evidence that exposure to news can affect trust and engagement, but there is still some disagreement whether the effect occurs in a positive or negative direction (De Vreese, 2004). Some scholars hypothesized that media can increase the levels of cynicism and disengage citizens (Capella & Jamieson, 1997). Others proposed that it actually stimulates civic engagement and raises the level of trust (Norris, 2000). With the support of existing knowledge on framing effects, one of the concerns of this contribution is to examine the relationship between media and trust.

A considerable amount of literature shows that corruption scandals can influence confidence in public institutions. At the individual level, Seligson (2002) found that those who experience corruption are less likely to believe in the political system and that they demonstrate lower levels of interpersonal trust. Likewise, higher levels of perception of corruption are correlated with decrease in institutional trust (Chang & Chu, 2006; Clausen, Kraay, & Nyiri, 2011; Morris & Klesner, 2010) and with citizens’ trust in local politicians (Chaia & Teixeira, 2001; Solé-Ollé & Sorribas-Navarro, 2014). However, there has been little research on how corruption scandals are framed and on the possible consequences for

political trust. This study aims to assess whether the attribution of responsibility to different political actors can lead to changes in trust among these actors. For this purpose, an

experiment was conducted among Brazilian citizens which were exposed to different responsibility frames about a fictional corruption scandal.

(4)

Having emerged from dictatorship in 1985, Brazil is a young democracy that has been through several scandals in the past decade and is perceived as one of the most corrupt

countries in the world (Transparency International, 2018). The share of Brazilians who mentioned corruption as one of the most serious problems the country has faced ranged between 8% to 14% during the 2006 and 2012 editions of the Americas Barometer Survey (Weitz-Shapiro & Winters, 2016). With this in mind, this paper examines how media framing of corruption scandals affects political trust in Brazil. Consequently, this research will

broaden current knowledge of determinants of political trust, offering an alternative perspective to the majority of studies produced within the context of North America and Europe (Van der Meer, 2017).

This paper is divided into four sections. The literature review is organized in a way that it initially presents what has been found about media effects on political trust, then it examines the question of determinants of trust in government to finally introduce the frame analysed in this study. Section two covers the methodology employed by this research with description of the sample, stimulus material and measures. The next section contains the findings obtained during the data gathering phase and a discussion of the results. Finally, a conclusion section points out the limitations of this study and highlights its achievements.

Media effects on political trust

As discussed, scholarly inquiry has found several determinants of political trust, but before I return to that I will look at the role of media, which plays a big role in mediating information about what is happening in the political arena to citizens. Even though news media are regarded to have a democratic role in society, experts have associated them with decrease in political trust and disengagement from politics (Capella & Jamieson, 1997; De Vreese & Semetko, 2002; Moy & Scheufele, 2000). It is largely argued in the literature that

(5)

negative content towards political actors can evoke political cynicism – citizens’ predisposition to find fault and assume that human action is motivated by self-interest (Dancey, 2011). Thus, content valence was previously hypothesized to affect political trust (De Vreese & Semetko, 2002; Pfau, Moy, Radler, & Bridgeman, 1998).

Another factor that has been associated with changes in the level of trust is the news source. Taking into consideration the characteristics of different media, it is argued that changes in the level of political trust could be due to how news are presented. Few studies have demonstrated that sources tend to report issues in a specific manner. A content analysis comparing media source and tone of coverage towards different institutions in the United States exhibited significant differences among newspapers, magazines, television and radio (Pfau et al., 1998). Print formats were associated with higher levels of political trust whereas broadcasting was correlated with lower levels of trust in Congress and in the Presidency. Similarly, several studies suggest that watching television news may decrease the level of trust in government and that reading newspapers may lead to the opposite effect (see Becker & Whitney, 1980; Miller & Reese, 1982; Pfau et al., 1998).

The possible side effects of exposure to television news gained a term: videomalaise. Robinson's (1976) theory claims that television content is extremely negative towards

political matters, which can lead to public disengagement and negative behaviours. According to Iyengar (1991), the negative relationship between television consumption and lower levels of political trust is partly due to the ‘quality’ of TV content. Robinson (1976) explains that television news may be confusing, affecting the understanding of topics being presented. Some empirical studies tested the videomalaise theory and found evidence linking exposure to TV news with declining levels of voter turnout (Pfau et al., 1998) and diminished civic engagement (Putnam, 2000).

(6)

Notwithstanding, there is still considerable ambiguity that news media has an effect on negative attitudes or malaises (De Vreese, 2005). A great amount of studies did not find supportive evidence that such a phenomenon affects political trust in a negative direction (Hetherington, 1998; Moy & Scheufele, 2000; Norris, 2002). Using data from the American Election Study, Moy and Scheufele (2000) showed that exposure to mass media can affect social trust, but they did not observe the same effect on political trust. On the other hand, Avery (2009) discovered that media exposure can both discourage and promote political trust under some circumstances. Based on data from several panel studies, he underlines that among individuals who previously reported low and high levels of political trust greater exposure to television did not lead to higher levels of trust in government. In contrast, greater newspaper exposure among trusting individuals increased trust.

Some scholars, in fact, claim that media can increase the level of political knowledge of individuals who, in turn, would become more trusting and engaged (Norris, 2000). In a way, the debate is similar to the discussion about the consequences of lower levels of political trust for democracy. There are contradicting findings and explanations about media effects, which makes this topic an interesting field to be further explored. In this study, I follow Cappella and Jamieson’s (1997) argument that media are more likely to instigate cynicism, especially because corruption scandals reveal misconduct of politicians, which can affect how individuals perceive the integrity of political actors. Based on the videomalaise theory (see Robinson, 1976), I hypothesize that individuals who learn mostly about corruption scandals through television will have higher levels of cynicism than people who are informed through newspapers:

H1: Individuals exposed to corruption scandals on television will have higher levels of cynicism than those who learn about it through newspapers.

(7)

Figure 1

Relationship between media exposure and political cynicism

So far, we have seen some general thoughts about how media can influence political trust. To better understand this relationship, instead of just taking into account media

consumption, Kleinnijenhuis, Van Hoof and Oegema (2006) argue that it is necessary to look at the characteristics of news as well. One way of doing so is by looking at how media frame certain topics. According to Entman (1993), framing is a process involving “selecting of some aspects of a reality and making them more salient" (p. 52). Hence, a framing effect is “one in which salient attributes of a message (its organization, selection of content, or thematic structure) render particular thoughts applicable, resulting in their activation and use in

evaluations” (Price, Tewksbury, & Powers, 1997, p. 486). Exposure to different news frames has been related to changes in how individuals perceive and evaluate people, issues and institutions (Iyengar, 1991; Valkenburg, Semetko, & De Vreese, 1999). Through surveys, experiments and content analyses, Cappella and Jamieson (1997) hold the view that frames can affect which type of information individuals will remember of a news story.

Regarding framing effects on political trust, research indicates that the use of strategy frames can increase cynicism. Many scholars have criticized the American press of framing political campaigns in terms of strategic frames, portraying candidates as competitors looking

(8)

for advantages (Capella & Jamieson, 1997; Entman 1989). Cappella and Jamieson (1997) found that strategic frames enhanced political cynicism regarding candidates involved in political campaigns stories. As concern has been raised that the trend could influence electoral outcomes, many other studies built upon Cappella and Jamieson’s work. For example,

Valentino, Beckmann and Buhr (2001) explored the impact of strategy frames on trust in government. They found that strategy frames seem to decrease participation and confidence in government but also highlighted that effects differed among participants. These studies show that certain frames can have effects on both trust in the incumbent and in institutions.

Scholarship on framing recognizes that effects may vary across individuals. The level of political sophistication of a person has been considered to moderate the influence of news media on individuals’ opinions (Price & Zaller, 1993, Zaller, 1992). Avery (2009) argues that since individuals have prior political attitudes and knowledge these characteristics should moderate how they interpret mediated information. For example, the investigation conducted by Valentino et al. (2001) points out that individuals with lower levels of partisan bias and education are more likely to suffer negative effects of strategy frames. De Vreese and Elenbaas (2008) demonstrated a positive interaction among political knowledge, strategic press and publicity news exposure, indicating that the politically sophisticated are more likely to evoke cynical attitudes after being exposed to those frames. Similarly, Winters and Weitz-Shapiro (2016) have found that political sophistication moderate citizens’ responses to corruption scandals. Now that I briefly outlined how media may influence political trust, I will proceed with other factors that have been described by scholars.

Determinants of political trust

Often associated with trust in government, political trust is a concept that relates to citizens’ support for political institutions (Van der Meer, 2017). Scholars have empirically

(9)

referred to political trust as support for principles and political actors even though

conceptually it refers to a specific set of objects (Van der Meer, 2017). To avoid confusion, I will use the term as a reference to regime institutions only. Easton (1975) categorized

citizens’ support into two types: specific and diffuse. Specific support relates to the performance and output of incumbents regarding particular actions. In contrast, diffuse support displays an overall evaluation of an actor on the long term. There seems to be no consensus among researchers about which type of support political trust characterizes. Some studies considered that political trust characterizes an expression of diffuse support, resulting from different experiences over time (Easton, 1975; Miller, 1974). It was also suggested that it refers to specific support (Citrin, 1974; Citrin and Green, 1986). Despite the disagreement among experts, a remarkable achievement of Easton’s conceptualization is that it has been used in many studies, and trust indices tend to contain measures of both types (Hetherington, 1998).

Political trust is considered an essential component to the maintenance of democratic regimes as it is believed to reflect public support towards the decisions of incumbent

authorities (Dalton, 2004). Whereas it has been argued that scepticism stimulates political engagement (Cook & Gronke, 2005), negative changes in the level of political trust have been associated with a threat to the stability of the system (Van der Meer, 2017). Many

investigations have been dedicated to discuss theoretically and empirically causal

explanations for changes in the level of trust in government. Initial debate focused mainly on political factors. Miller (1974) argued that political parties’ inability to comply with citizens’ policy expectations decreased the level of trust in government in the United States, suggesting that policy satisfaction affected political trust. Citrin (1974), on the other hand, attributed lower levels of trust in government to dissatisfaction with incumbent’s performance. As more research in this field was carried out, later studies found evidence that citizens’ evaluations of

(10)

institutions also influenced political trust (Citrin and Green, 1986; Hetherington, 1998). Further discussion about the factors that could lead to changes in the level of political trust came after the report The Crisis of Democracy, published in 1975. The document pointed out that erosion in trust was a trend not only in the United States but also in Japan and in Europe (Crozier, Huntington, & Watanuki, 1975). It basically indicated that the issue was not an isolated case and had global relevance. Van der Meer (2017) states that the content presented in the report has motivated research in the field for decades.

Some of these investigations focused on the relationship between political trust and corruption scandals. Solé-Ollé and Sorribas-Navarro (2014) compared Spanish municipalities affected by corruption scandals with others where scandal did not break out to see whether trust in local politicians changed. The results of their analysis showed that areas affected by a corruption scandal had a lower level of trust in local politicians. In addition, they observed a stronger effect among individuals who obtained information about the scandals from the media. Similar studies were also conducted in other parts of the world with a different focus. Based on data from the East Asia Barometer, Chang and Chu (2006) tested the effect of corruption on institutional trust. They found that the level of citizens' perceptions of corruption decreased their trust toward political institutions. Likewise, Seligson (2002) compared national sample survey data from four countries in Latin America and found supporting evidence that corruption experiences can influence the belief in the legitimacy of the political system.

Other studies addressed not only political factors but also issues regarding the economy and social aspects. Scholars found empirical evidence suggesting that economic performance and citizens’ economic evaluations (Citrin and Green, 1986; Chanley et al., 2000; Hetherington, 1998) as well as concern about criminality (Chanley et al., 2000) also

(11)

influenced political trust. In short, literature indicates that declining trust in government has several potential causes, and the phenomenon proved to be complex (Chanley et al., 2000).

While most studies have treated political trust as a dependent variable, others have challenged its role and considered it an independent variable. Hetherington (1998) analysed whether trust in government could affect other components of government such as

incumbents and institutions. His findings report that trust in government affected citizens’ support towards the president and Congress. Reciprocally, however, citizens’ support for Congress and the president influenced trust in government, pointing out to a possible symmetrical relationship. Hetherington (1998) underlines that trust in government holds a greater causality effect and argues that support for institutions has greater impact on trust in government than attitudes towards incumbents. Chanley et al. (2000) also found similar empirical evidence, showing that lower levels of trust in government could lead to less trust in Congress. Similarly, they found that allegations of misconduct by Congress had greater influence on trust in government than misconduct by the president. These studies highlight that trust in government may have a spillover effect in other components of trust and could have a reciprocal causation. I find it possible to observe spillover effects during the

experiment especially because corruption scandals have been linked with decrease in institutional trust (Chang & Chu, 2006) and may affect how citizens' view the legitimacy of the political system (Seligson, 2002). Hence, based on the work of Hetherington (1998) and Chanley et al. (2000), I expect the following hypotheses:

H2: Individuals with low trust in government will also have low trust in Congress and in the president.

(12)

H3: Trust in Congress will have greater impact on trust in government than trust in the president.

Figure 2

Correlations among trust in government, trust in Congress and trust in the president

Framing corruption scandals in Brazil

Corruption scandals have been found in multiple layers of the Brazilian political arena since the country’s transition to democracy (Balán, 2014). Yet, it cannot be said that electoral punishment is an automatic consequence of exposure to corruption scandals coverage. For example, former president Luis Inácio Lula da Silva (2003-2011) managed to survive allegations of involvement in the Mensalão scandal during his first term and succeeded in winning another election. The case has triggered the interest of Da Silva (2008), who wrote a master thesis about the topic and performed a content analysis of the main Brazilian weekly magazines to assess how each of them framed Mensalão.

An interesting aspect of Da Silva’s research is that it fills a gap in literature about how corruption scandals are reported by the media in Brazil. In all of the publications, valence frames related to three political actors were often mentioned in the articles: Lula, PT

(Workers’ Party) and the government. Contrary to what is known about the effects of strategy frames during electoral campaigns, few studies have investigated how corruption scandals are

(13)

framed, and even less refer to the effects that it may cause on political trust against this backdrop. Panayırcı, İşeri and Şekercioğlu (2016) argue that existing literature on corruption makes use of both generic and issue-specific frames. The first category is characterized as having a universal use because it can be applied to distinct topics both in different cultural contexts and periods whereas issue-specific frames are pertinent only to specific themes (De Vreese, 2005). The main difference between these two approaches, according to De Vreese (2005), is that the later allows for a profound level of details regarding the issue under investigation and the former is more prone to comparability and generalization. For this investigation, it was decided that the best procedure was to use a generic frame rather than an issue-specific frame. Since only few studies have addressed the subject, it is safer to rely on a generic frame to understand possible effects on political trust.

Zamora and Albaladejo (2010) proposed a typology of five issue-specific frames used in the coverage of political corruption scandals. Nonetheless, to the best of my knowledge, they were not tested in other investigations. Additionally, Panayırcı et al. (2016) argue that frames used in earlier studies may be insufficient in some cases due to cultural context. Thus, opting to employ an issue-specific frame proposed by Zamora and Albaladejo (2010) could invalidate the findings of my analysis. A well-documented reference to understand which generic news frames are normally used in political stories is Semetko and Valkenburg's (2000) study. They investigated the predominance of five frames identified by previous work in television news and newspaper articles in the Netherlands. Their results showed that the use of frames depended on the type of issue being covered and the type of outlet as well. Overall, it was observed that responsibility frame was commonly used by television and print news, respectively followed by conflict, economic consequences, human interest, and

(14)

In an effort to establish which frame would be more appropriate for this study, I compared the generic frames documented by Semetko and Valkenburg (2000) with Da Silva’s (2008) content analysis findings. Da Silva (2008) demonstrated that the weekly magazines generally framed the scandal in such a way to either exempt or attribute direct and indirect responsibility to political actors. Hence, out of many generic frames, the responsibility news frame was chosen because it represents the essence of what was described by Da Silva. This generic frame "presents an issue or problem in such a way as to attribute responsibility for its cause or solution to either the government or to an individual or group" (Semetko &

Valkenburg, 2000, p. 96). According to Semetko and Valkenburg (2000), this type of frame was specially employed by outlets regarded as serious and sober.

One of the concerns regarding the use of a responsibility frame refers to changes on citizens’ perceptions of which actor is responsible for an issue (Iyengar, 1991). He argues that when television news focus on a specific event it instigates individuals to find explanations for social issues on the individual level. In contrast, if the frame is used in thematic stories, within a larger historical context, it may encourage people to seek for answers in the macro level, encompassing government and the system. This factor will be considered during the discussion section since I will use the frame to focus on a specific episode rather than on a feature story. This study is exploratory in the sense that it will test whether the responsibility frame can have effects on political trust in the coverage of corruption scandals. Bowler and Karp’s (2004) findings showed partial support that scandals can decrease trust in politicians and have a spillover effect on trust in government and in other institutions. I will re-examine this matter and analyse whether trust among actors involved in the frames varies. I expect that respondents will demonstrate more negative evaluations for the political actor featured in the responsibility frame than for other institutions.

(15)

H4: Exposure to news framed in terms of responsibility decreases trust for political actors mentioned in the frames.

Figure 3

Framing effects on different types of trust

Method design

To investigate responsibility framing effects, I draw on data from an experiment that included three conditions. Each participant was exposed to one condition only. The stimuli were fictitious articles about a corruption scandal involving an incumbent who was

demanding bribes from a mining company to help the organization have easier access to environmental licenses granted by a governmental institution. I produced three versions of the articles, each of which framed the story in terms of responsibility. One version highlighted that the federal government was aware of the bribery mentioned in the article. The second one emphasized that political parties of the ruling base profited with the money, whereas the third article focused only on the incumbent, and it depicted him as a political articulator. Only the headline, kicker and lead were manipulated among the articles to comply with each

(16)

responsibility frame. The rest of the story’s body was left intact to increase comparability among the stimuli. There was no control group for this experiment.

Sample and procedure

A sample was drawn through a nonprobability sampling technique, specifically convenience sampling. Participants were recruited through social media in Facebook groups. Even though this type of sampling has limitations in terms of generalizing findings to the population (Reinard, 2006), the sample consisted mainly of individuals participating in groups for Brazilians living abroad and for students of federal public universities, which are formed by people with different sociodemographic characteristics. A total of 318 Brazilians

participated in the experiment conducted in April 2018 through the software Qualtrics. Entries that did not meet the age criteria (individuals above 18 years old), failed to answer correctly two manipulation check questions and completed less than 50% of the questionnaire were excluded from the final sample of 179 subjects (61.5% female, 88.3% highly educated [M = 4.49, SD = 1.16], aged 18 to 76 years [M = 38.44 years, SD = 12.65 years]).

The subjects gave an informed consent that explained that their participation in the study about the effects of corruption scandals on political trust was voluntary; the study did not offer any type of harm; their data was safely stored and accessed only by the researcher and his supervisor; participants were allowed to withdraw their participation at any time during the survey or by sending an email to the researcher up to seven days after the

completion of the questionnaire. In total, the experiment was composed of three stages. Prior to the stimulus, participants filled a questionnaire containing basic demographics, media consumption habits, political engagement, party and ideological identification, and political interest. After that, they were randomly assigned to one of the three conditions and read a

(17)

manipulated article that attributed responsibility for a scandal either to the government, a former incumbent or political parties.

Before being exposed to the stimulus material, participants were asked to pay attention to the news article that they were about to read. The news article did not contain a byline or any reference to a news outlet to avoid source credibility bias as individuals may distrust news outlets depending on their political view. After exposure to one stimulus

material, participants answered two manipulation check questions and a questionnaire related to political trust, trust in institutions, cynicism and political knowledge. Only after submitting all the answers they were debriefed that the news article was manipulated for the purposes of the research and that despite of the likelihood to a real situation the case did not occur.

Stimulus material

The articles were based on a story published by the Brazilian news website G1, which gathers all the content produced by news outlets maintained by Grupo Globo - one of the largest media conglomerates in the country. The original case was published in March 2018 and mentioned a former deputy indicted on corruption charges following an investigation under the Lava Jato (Car Wash) operation – one of Brazil’s biggest corruption investigations that began in March 2014. The former incumbent was investigated for receiving bribes from foreign companies to help them gain contracts with Petrobras – Brazil’s semi-public

petroleum industry. I selected this story as a starting point because it referred to a specific event inside a large undergoing corruption scandal investigation. Adapting this news article would be representative of a real case and, therefore, harder to discredit by the participants in the experiment because of its similarity to stories that have been published by the media. For my articles, I used the name of another deputy, political parties, and replaced Petrobras for a mining company located in the North of Brazil. Details such as the type of crime the

(18)

incumbent would be accused of and the procedures required for mining activities in protected areas were carefully researched to make the story appealing.

The core parts of the article (the second and third paragraph) were the same among all the versions to assure a high comparability among stimuli. They provided factual and

background information about what led to the investigation of the politician and how the search and seizure warrant at his residence occurred. The manipulation consisted of headline, kicker and first paragraph (lead). Headlines were used to indicate who had responsibility for the case: ‘Former deputy José Geraldo Torres (Worker's Party) is indicted for corruption by the Federal Police’; ‘Guará Operation reveals corruption scheme involving parties of the allied base of Dilma administration’; and ‘Federal government tries to suppress an investigation that reveals a corruption scheme among Dilma’s ruling base parties’.

To make sure that the articles were written in a way that could be representative of a responsibility frame, I followed the criteria employed by Semetko and Valkenburg (2000) on a content analysis study measuring five news frames, including responsibility. They

developed five questions specific for responsibility frames which coders had to answer yes or no. These questions were such as “does the story suggest that some level of the government is responsible for the issue/problem?” and “does the story suggest that an individual (or group of people in society) is responsible for the issue/problem?” This parameter was useful to assess whether the stimuli could be coded as responsibility frame. For the English translation of the articles, see Appendix A.

Measures

Trust measures. Political trust was measured through an index composed by five items that assessed individuals trust in government. They were derived from studies

(19)

in the American National Election Study (NES). Questions referred to individuals’ trust in government in different dimensions: (1) to do what is right, (2) whether it is run by a few big interests, and (3) if it wastes taxpayers' money. It also touched upon individuals' opinion regarding (4) whether government cares about what people think, (5) if citizens have a say about what government does, and (6) if government paid attention to what people think to make decisions over the years. After reverse coding responses to propositions 1, 2 and 6, all the items were added to the political trust index (M = 1.44, SD = .45, alpha = .65).

For trust in incumbents and political parties, only the first item of the political trust index was used. The question was adapted to apply it to each political actor: "How much of the time do you think you can trust ________ to do what is right? The answers were recorded on a 4-point scale: 1 = never; 2 = some of the time; 3 = most of the time; 4 = all the time. The propositions were reverse coded and the scales ranged from 1 (low trust) to 4 (high trust) (trust in incumbents [M = 1.40, SD = .55]; trust in political parties [M = 1.29, SD = .53]). To measure cynicism, I used a proposition previously removed from the trust index to improve its Cronbach's alpha. Even though Hetherington (1998) used it as part of his trust measure, the item tapped on individuals’ perception of politicians’ integrity. The question asked

respondents how many incumbents are crooked. The scale ranged from 1 (low cynicism) to 3 (high cynicism) (M = 1.40, SD = .72) and the item was reverse coded.

Two items were combined into an index to operationalize trust measures for Congress (M = 1.78, SD = .96, alpha = .73) and the president (M = 1.69, SD = 1.06, alpha = .65). Both propositions were adapted using the following text: (1) On a scale from zero to ten, how would you rate ________? 0 = very unreliable; 10 = very reliable. (2) In general, do you approve or disapprove the way ________ has been handling its job? The answers were recorded on a 5-point scale: 1 = strongly disapprove; 2 = disapprove; 3 = neither approve or disapprove or don't know; 4 = approve; 5 = strongly approve. Item 2 was reverse coded.

(20)

Mass media variables. To measure television exposure, two items were combined into an index (M = 3.49, SD = 1.42, alpha = .75). These items tapped upon (1) the amount of days the individual spent watching news on television during the previous week and (2) how much attention they paid to stories on TV about corruption scandals. The same questions were adapted to newspaper exposure, which also was operationalized through a two-item index (M = 2.44, SD = 1.75, alpha = .75). For descriptive statistics of the variables and wording of associated items, see Appendix B.

Results

The initial set of analyses focused on the relationship between media consumption and cynicism. My first hypothesis expected that individuals exposed to corruption scandals on television would have greater levels of cynicism than people who learned about it through newspaper articles. To test this assumption, I ran a linear regression model examining political cynicism as the dependent variable and exposure to television and newspaper as independent variables. The model was not significant, F(2) = .76, p = .468. Therefore, it cannot be used to predict the level of political cynicism. Not surprisingly, the strength of the prediction was also very low: zero percent of the variation in political cynicism could be predicted on the basis of these factors (R2 = .00). Looking at the specifics of the variables, I found that television consumption, b* = .01, SD = .02, t = .45, p = .65, 95% CI [-.04, .7], and newspaper consumption, b* = .03, SD =.03 t = 1.03, p = .303, 95% CI [-.03, .09], had a non-significant, weak association with political cynicism. Based on these findings, it is not possible to infer that differences in the level of political cynicism are due to exposure to different media, hence, my first hypothesis was debunked.

During the second set of analyses, I tested some of the determinants of political trust mentioned in the literature. My second hypothesis assumed that individuals with low trust in

(21)

government will also have low trust in Congress and in the president. To examine the relationship among these variables, I ran a correlations test. There was a significant,

moderately strong positive correlation between individuals' evaluations of Congress and their level of trust in government (r = .42, p < .001). Thus, when citizens do not trust government they also tend to distrust Congress. On a weaker level, a significant, positive correlation between trust in government and trust in the president was also observed (r = .21, p = .004). In sum, it was demonstrated that there is a positive correlation among these variables, which provides support for my second hypothesis.

I also investigated whether trust in Congress has greater impact on political trust than trust in the president, as stated on my third hypothesis. To assess it, I performed a linear regression model examining trust in government as the dependent variable and trust in Congress and trust in the president as independent variables. The model was significant, F(2) = 19.16, p < .001. Therefore, it can be used to predict the level of political trust. Nonetheless, the strength of the prediction was low: 18 percent of the variation in political trust was predicted on the basis of these factors (R2 = .18). Looking at the variables, the regression displayed that trust in Congress, b* = .17, SD = .03, t = 5.34, p < .001, 95% CI [.11, .24] had a significant, weak association with trust in government. In other words, for each additional point on the scale of trust in Congress, which runs from 1 (low trust) to 6 (high trust), the average propensity to trust government increases by .17. In contrast, trust in the president, b* = .02, SD =.03 t = .91, p = .364, 95% CI [-.03, .08], had a non-significant, weak association with trust in government. For these effects, independent variables were assumed to be held constant. This model shows that trust in government is not predicted by trust in the president, whereas trust in Congress has a small effect in the dependent variables. Thus, it can be concluded that these findings provide support for my third hypothesis as well.

(22)

Table 1

Responsibility frames one-way analysis of variance for trust in political institutions Responsibility frames

Trust Incumbent Political parties Government F

Incumbents 1.46 (.58) 1.41 (.56) 1.30 (.50) 1.31 Political parties 1.35 (.61) 1.32 (.54) 1.16 (.37) 2.12 Government 1.46 (.41) 1.51 (.52) 1.35 (.41) 1.70 N 71 58 50

Note. Standard deviations appear in parentheses below means.

Lastly, I examined media effects on different types of trust. My fourth hypothesis stated that exposure to news framed in terms of responsibility decreases trust for political actors mentioned in the frames. For each condition, Table 1 displays the means of each type of trust. I carried out an one-way analysis of variance to assess the influence of frames on trust in incumbents F(2) = 1.31, p = .272, trust in political parties F(2) = 2.12, p = .122, and trust in government F(2) = 1.70, p = .184. No significant effects were found in any of the conditions. Nonetheless, it stands out that on average all the variables had a low mean score of trust. It should be noted that the assumption of equal variances in the population has been violated for trust in politicians [Levene's F (2, 176) = 3.25, p = .041] and trust in incumbents

(23)

[Levene's F (2, 176) = 9.02, p < .001]. Since no significant difference between subjects was established, my fourth hypothesis cannot be confirmed.

Discussion

This study investigated how mediatized corruption scandals can affect political trust and cynicism. Drawing on data from an experiment, it was demonstrated that trust in government has a spillover effect in other political institutions and actors. There was a positive correlation between trust in government with trust in Congress and between trust in government with trust in the president. In addition, the present investigation adds to our understanding of which factors can influence political trust. While lower levels of

congressional approval can affect trust in government, the same cannot be said for trust in the president. These results corroborate the findings of earlier studies that found political trust to be more tied with trust in Congress than with trust in the president (Hetherington, 1998; Chanley et al., 2000). These investigations suggested a symmetrical relationship among these institutions. Hence, I found supporting evidence that the relationship between trust in

government and trust in Congress could be rather symmetrical.

These findings can have important implications for the Brazilian political scenario. If we consider a symmetrical relationship between trust in government and trust in Congress, we can expect that negative evaluations of government are negative to Congress and vice-versa, which in turn provides institutions less leeway to govern effectively and store support (Hetherington, 1998). On average, the mean score of trust in these institutions was low. This implies that it is necessary to change the way governmental institutions are run. Increasing the degree of transparency and effectiveness in these organizations could be a possible solution. If we assume that lower levels of institutional trust are connected to greater perceptions of corruption (Chang & Chu, 2006; Clausen, Kraay, & Nyiri, 2011; Morris & Klesner, 2010), it

(24)

can be argued that the phenomenon is systemic and must be fought in all levels of

government to improve institutional ratings. Notwithstanding, it is not possible to infer that evaluations obtained during the experiment are due to exposure to a corruption story as I did not measure subjects' prior level of trust.

Similar to studies that did not find evidence that media affects political trust in a negative direction (Hetherington, 1998; Moy & Scheufele, 2000; Norris, 2002), this

investigation did not find support for the assumption that individuals who watch more news on television are more cynical than those who read newspapers (Becker & Whitney, 1980; Miller & Reese, 1982; Pfau et al., 1998). It has to be noted that a possible explanation for this outcome could be that nowadays citizens also read and watch news online. In other words, I could have missed those who prefer having a digital subscription of their newspaper or only watch television online. Perhaps a comparison among several media types, including digital media, could enlighten the discussion of how media relates with cynicism.

Moreover, no significant effects were found on the level of trust in government, trust in incumbents and trust in political parties after comparing exposure to different types of responsibility frames. Except for when government was blamed for the scandal, institutions featured in the frames did not elicit lower levels of trust compared to other actors. It is noticeable that political parties had the lowest scores in every condition, which could suggest that they are perceived as even less credible than corrupt incumbents. Although trust in

government had slightly higher scores most of the time, it seems that Iyengar's (1991) concern that frames emphasizing specific episodes may instigate individuals to find explanations on the individual level did not apply to this context. Taking into consideration that none of these actors had a high level of trust and that none of the values are significant, it is not possible to assume that either exposure to corruption scandals was responsible for low evaluations of these actors or that differences in responsibility frames lead to changes in trust among

(25)

institutions. Based on these findings, this research cannot confirm that scandals involving legislators can have negative influence on individuals' attitudes towards institutions as

indicated by Bowler and Karp (2004). However, it cannot be discarded the possibility that the effect of corruption perception in the Brazilian society is so strong that framing effects do not lead to greater differences in different types of trust. This issue needs to be further researched by future studies.

Conclusion

This work has led us to conclude that trust in government may have a symmetrical relationship with trust in Congress. To improve citizens' perceptions of these institutions, it is important that incumbents take actions on all levels of government, since individuals'

evaluations of a single political institution may influence how they perceive others. These findings are in accordance with earlier studies, which suggest that political trust is more connected to trust in Congress than trust in the president (Hetherington, 1998; Chanley et al., 2000). In general, this research has extended our understanding of determinants of political trust and how spillover effects may occur between institutions, but an important empirical question that remains to be answered, however, is whether responsibility frames affect trust in different institutions against the backdrop of corruption scandals. Based on my findings, it is not possible to infer that media framing of corruption scandals affect political trust in Brazil. No significant effects on the level of trust in government, political parties and incumbents were found after exposing individuals to different types of responsibility frames. It is possible that Brazilians are so fed up with corruption that their general perception of any political institution is low. Another explanation could be that these frames do not cause changes in political trust or that my stimuli materials were not representative, although great care was taken to comply it with the criteria proposed by Semetko and Valkenburg (2000). Therefore,

(26)

future studies could further explore whether differences in these frames also contribute to lower evaluations of trust in different institutions. Moreover, this research also did not provide support for studies that hypothesized that citizens who pay attention to news on television are more cynical than those who read newspapers (Becker & Whitney, 1980; Miller & Reese, 1982; Pfau et al., 1998). Hence, it is plausible to agree with De Vreese (2005) that that there is still considerable ambiguity that news media lead to negative attitudes or malaises.

It is important to note that this paper also has some shortcomings. One of the major issues found in my experiment is that the sample was obtained through a convenience

sampling technique, which jeopardizes the generalization of the results (Reinard, 2006, p. 34). This implies that the sample cannot be considered representative of the Brazilian population, as participants were recruited through social media. My findings reflect only a limited

segment of Brazilians. In addition, the experiment did not have a control group, which means that during the analysis it was only possible to compare effects between frames. The inclusion of a control group, where individuals are not exposed to any frame, would have offered a baseline for comparison to capture the impact of each frame (De Vreese, 2010). In this case, the results for the framing analysis were not significant, but supposing that they were, without a control group it would not be possible to infer whether the frames really led to those results or if they were due to other factors. In other words, I would not be able to establish a causal relation between exposure to frames and changes in the level of trust. For the sake of

simplicity, my regression models and analysis of variance did not include other determinants of political trust described in the literature. If this had been done, the models would possibly have improved their representability and adequacy to analyse the dependent variable at stake. Despite of these limitations, this paper has managed to set a basis for future analysis and corroborated to previous studies on determinants of political trust.

(27)

Appendix A

Responsibility frames

Incumbent: Former deputy José Geraldo Torres (Worker's Party) is indicted for corruption by the Federal Police. According to the Federal Public Ministry (MPF) in Pará, Mr. Torres has been investigated since last year under Guará Operation, which investigates fraud in environmental licenses granted to mining company Hydro.

Former federal deputy José Geraldo Torres (PT/PA) was indicted by the Federal Police (PF) for active and passive corruption under Guará Operation, on Monday (April 2nd). Torres has been investigated since 2016 for favouring environmental licenses for the mining company Hydro in conservation units in Pará in exchange for bribes. According to the MPF, he was the main beneficiary of US$ 500,000 paid by the Norwegian multinational responsible for the exploration of aluminium and other minerals in the state. The former deputy was head of Dilma Rousseff's (PT) government from 2014 to 2015 and gained fame as an ecumenical articulator by communicating with opposition parties.

The PF complied with the search warrant and seizure at Torres residence, located in Asa Norte, in Brasilia. After making a search and seizure at the suspect's house, he was taken to the PF headquarters in Brasilia to give testimony. The MPF states that the suspect put pressure on Ibama to allow mining activities in protected areas. Ibama has not made any statements about the case yet.

The investigations began after former director of Hydro Maria Ribeiro da Costa was offered a plea bargain in June 2016 at the PF in the Federal District. Mrs. Da Costa said that the company obtained at least seven environmental licenses between 2010 and 2013 by giving bribes. The Federal Police is due to listen to other suspects who participated in the corruption scheme this week.

(28)

Political parties: Guará Operation reveals corruption scheme involving parties of the allied base of Dilma administration. According to the Federal Public Prosecutor's Office (MPF) in Pará, frauds in environmental licenses granted to mining company Hydro would have benefited at least four political parties.

The Federal Police (PF) indicted on Monday (April 2nd) representatives of PT,

PMDB, PSD and PP for active and passive corruption under Guará Operation. Party members have been investigating since 2016 for favoring environmental licenses to the mining

company Hydro in conservation units in Pará in exchange for bribes. According to the MPF, the amount of US$ 500,000 paid by the Norwegian multinational responsible for the

exploitation of aluminium in the state would have been divided equally among the parties of the governing base. Among the indicted was former MP José Geraldo Torres (PT/PA), political leader that gained fame as an ecumenical articulator by communicating with opposition parties during Dilma's administration

Government: Federal government tries to suppress an investigation that reveals a corruption scheme among Dilma’s ruling base parties. According to the Federal Public

(29)

Prosecutor's Office (MPF) in Pará, the federal government would be aware of frauds in environmental licenses granted to the mining company Hydro.

The Federal Police (PF) indicted on Monday (April 2nd) former federal deputy José Geraldo Torres (PT/PA) for active and passive corruption under Guará Operation, which has been investigating since 2016 favoured contracts offered to Hydro in exchange for bribes. In testimony, Torres claims to have received BRL 30,000 as undue funds to carry out the necessary procedures. He said that he has been threatened by top government officials to remain silent. According to MPF, the mining company paid US$ 500,000 in bribes to members of the PT, PMDB, PSD and PP.

Appendix B

Overview of independent variables

Age: In years (M = 38.44 years, SD = 12.65 years). Gender: 1 = male, 2 = female, 3 = other.

(30)

Education: 1 = less than a high school diploma, 2 = high school, 3 = technical training, 4 = bachelor’s degree or equivalent, 5 = MBA or equivalent, 6 = master’s degree or equivalent, 7 = PhD degree.

Ethnicity: 1 = white, 2 = black, 3 = yellow, 4 = indigenous, 5 = pardo.

Political interest: Two-item index, on a four-point Likert scale, ranging from 1 = hardly at all/not interested at all to 4 = most of the time/very interested. (1) You follow what is going on in government and public affairs; (2) How much interest do you usually have in what is going on in politics? Reverse coded, added, and divided by two to form a scale of political interest (M = 3.49, SD = .64, alpha = .80).

Ideology: Item tapping political-ideological preference on a scale ranging from 1 = extreme left to 10 = extreme right (M = 4.59, SD = 2.45).

Party identification: Which political party do you identify yourself the most? Options included 35 parties and a ‘none of the above’ answer.

Political engagement: Dummy variable tapping voting participation in the general elections held in 2014. 1 = yes, 2 = no.

Television exposure: Two item were combined into an index (M = 3.49, SD = 1.42, alpha = .75). (1) How many days in the past week did you watch news on TV? Eight-point Likert scale ranging from 0 to 7. (2) During the past week, how much attention did you pay to stories on TV about corruption scandals? A great deal, some, not too much, no attention at all?

Newspaper exposure: Two items were combined into an index (M = 2.44, SD = 1.75, alpha = .75). (1) How many days in the past week did you read a daily newspaper? Eight-point Likert scale ranging from 0 to 7. (2) During the past week, how much attention did you pay to newspaper articles about corruption scandals? A great deal, some, not too much, no attention at all?

(31)

Political cynicism: Item tapping perception of the amount of crooked incumbents on a scale ranging from 1 = hardly any to 3 = quite a few (M = 1.40, SD = .72). Item was reverse coded.

Trust in government: Six-item index. (1) How much of the time do you think you can trust the federal government to do what is right? Never, some of the time, most of the time or all the timer? (2) Would you say that the government is pretty much run by a few big interests looking out for themselves or that it is run for the benefit of all people? Few big interests or benefit of all? (3) Out of the money we pay in taxes how much do you think that people in the government waste? A lot, some, not much? (4) Do you agree or disagree with the following statement: I think public officials care much what people like me think. Strongly disagree, disagree, neither agree or disagree, agree, strongly agree? (5) Do you agree or disagree with the following statement: People like me have any say about what the government does. Strongly disagree, disagree, neither agree or disagree, agree, strongly agree? (6) Over the years, how much attention do you feel the government pays to what people think when it decides what to do? Not much, some or a good deal? (M = 1.44, SD = .45, alpha = .65). Items 1, 2 and 6 were reverse coded.

Trust in incumbents: Item tapping trust in incumbents to do what is right on a four-point Likert scale ranging from 1 = never to 4 = all the time (M = 1.40, SD = .55). Item was reverse coded.

Trust in political parties: Item tapping trust in political parties to do what is right on a four-point Likert scale ranging from 1 = never to 4 = all the time (M = 1.29, SD = .53). Item was reverse coded.

Trust in the president: Two items were combined into an index (M = 1.69, SD = 1.06, alpha = .65). (1) On a scale from zero to 10, how would you rate Michel Temer? Zero means very unreliable, and ten means very reliable. (2) In general, do you approve or

(32)

disapprove the way Michel Temer has been handling his job? Strongly disapprove,

disapprove, neither approve or disapprove or don't know, approve, strongly approve? Item 2 was reverse coded.

Trust in Congress: Two items were combined into an index (M = 1.78, SD = .96, alpha = .73). (1) On a scale from zero to 10, how would you rate Congress? Zero means very unreliable, and ten means very reliable. (2) In general, do you approve or disapprove the way Congress Temer has been handling its job? Strongly disapprove, disapprove, neither approve or disapprove or don't know, approve, strongly approve? Item 2 was reverse coded.

National economic evaluations: Item tapping current economic evaluation on a four-point Likert scale ranging from 1 = poor to 4 = excellent (M = 1.50, SD = .60).

Issue satisfaction: Items tapping individuals’ opinion about: (1) poverty (M = 1.31, SD = .31), (2) tax rates (M = 1.61, SD = .68), (3) public security (M = 1.22, SD = .45), (4) unemployment (M = 1.43, SD = .59), (5) healthcare (M = 1.29, SD = .52), (6) education (M = 1.15, SD = .43) and (7) corruption (M = 1.13, SD = .42). (1) Please indicate your opinion about the following issue: __________. Extremely serious problem, serious problem, not too serious problem, not a problem at all. All the items were reverse coded.

Political knowledge: Three-item index, with four multiple-choice questions, tapping factual knowledge about Brazilian national political actors (M = .69, SD = .34, alpha = .67), and binary coded as either 0 = wrong answer or 1 = correct answer: (1) Who is the current president of the House of Representatives? Rodrigo Maia, Eunício Oliveira, Romero Jucá, Fernando Bezerra, I don’t know? (2) Who is the current secretary of state? Eliseu Padilha, Henrique Meireles, Sarney Filho, Gilberto Kassab, I don't know? (3) Who replaced Rodrigo Janot as attorney general? Gilmar Mendes, Raquel Dodge, Alexandre de Morais, Luislinda Valois, I don’t know? The following item was excluded to improve the scale reliability:

(33)

Which president was the first to be part of an impeachment process? Getúlio Vargas, João Goulart, Fernando Collor de Mello, Dilma Rousseff, I don't know?

References

Avery, J. M. (2009). Videomalaise or virtuous circle? The influences of news media on political trust. International Journal of Press/Politics, 14(4), 410–433.

Balán, M. (2014). Surviving corruption in Brazil: Lula’s and Dilma’s success despite corruption allegations and its consequences. Journal of Politics in Latin America, 6(3), 67–93.

Becker, L. B., & Whitney, D. C. (1980). Effects of media dependencies: Audience assessment of government. Communication Research, 7(1), 95–120.

Bowler, S., & Karp, J. A. (2004). Politicians, scandals, and trust in government. Political Behavior, 26(3), 271–287.

Cappella, J. N., & Jamieson, K. H. (1997). Spiral of Cynicism. New York: Oxford University Press.

Chaia, V., & Teixeira, M. A. (2001). Democracia e escândalos políticos. São Paulo em Perspectiva, 15(4), 62–75.

Chang, E., & Chu, Y. (2006). Corruption and trust: exceptionalism in Asian democracies? Journal of Politics, 68(2), 259–271.

Chanley, V. A., Rudolph, T. J., & Rahn, W. M. (2000). The origins and consequences of public trust in government: A time series analysis. Public Opinion Quarterly, 64(3), 239–256.

Citrin, J. (1974). Comment: The political relevance of trust in government. American Political Science Review, 68, 973–988.

(34)

Citrin, J., & Green, D. P. (1986). Presidential leadership and the resurgence of trust in government. British Journal of Political Science, 16(4), 431–453.

Clausen, B., Kraay, A., & Nyiri1, Z. (2011). Corruption and confidence in public

institutions: evidence from a global survey. World Bank Economic Review, 25(2), 212–249.

Cook, T. E., & Gronke, P. (2005). The skeptical American: Revisiting the meanings of trust in government and confidence in institutions. Journal of Politics, 67(3), 784–803. Crozier, M., Huntington, S., & Watanuki, J. (1975). The crisis of democracy. Report on the

governability of democracies to the Trilateral Commission. New York: New York University Press.

Dalton, R. J. (2004). Democratic challenges, democratic choices: The erosion of political support in advanced industrial democracies. Oxford: Oxford University Press. Dancey, L. (2011). The consequences of political cynicism: How cynicism shapes citizens’

reactions to political scandals. Political Behavior, 34(3), 411–423.

Da Silva, V. J. (2008). O escândalo do mensalão em revistas semanais: Uma análise de enquadramento (Master thesis). Retrieved from:

https://repositorio.unesp.br/bitstream/handle/11449/89482/silva_vj_me_bauru.pdf?seq

uence=1

De Vreese, C. H. (2004). The effects of strategic news on political cynicism, issue

evaluations, and policy support: A two-wave experiment. Mass Communication & Society, 7(2), 191–214.

De Vreese, C. H. (2005). The spiral of cynicism reconsidered. European Journal of Communication, 20(3), 283–301.

De Vreese, C. H. (2005). News framing: Theory and typology. Information Design Journal & Document Design, 13(1).

(35)

De Vreese, C. H. (2010). Framing the economy: Effects of journalistic news frames. In D'Angelo, P. & Kuypers, J. A. (Eds). Doing news framing analysis: Empirical and theoretical perspectives (187–214). New York and London: Routledge.

De Vreese, C. H., & Elenbaas, M. (2008). Media in the game of politics: Effects of strategic metacoverage on political cynicism. The International Journal of Press/Politics, 13(3), 285–309.

De Vreese, C. H., & Semetko, H. A. (2002). Cynical and engaged: Strategic campaign coverage, public opinion, and mobilization in a referendum. Communication Research, 29(6), 615–641.

Easton, D. (1975). A re-assessment of the concept of political support. British Journal of Political Science, 5(4), 435–457.

Entman, R. M. (1989). Democracy without citizens: Media and the decay of American politics. Oxford University Press on Demand.

Entman, R. M. (1993). Framing: Toward clarification of a fractured paradigm. Journal of Communication, 43(4), 51–58.

Hetherington, M. J. (1998). The political relevance of political trust. American Political Science Review, 92(4), 791–808.

Iyengar, S. (1991). Is anyone responsible? How television frames political issues. Chicago: University of Chicago Press.

Kleinnijenhuis, J., Van Hoof, A. M., & Oegema, D. (2006). Negative news and the sleeper effect of distrust. Harvard International Journal of Press/Politics, 11(2), 86–104. Mansbridge, J. (1997). Social and cultural causes of dissatisfaction with U.S. government.

In Nye, J. S., Zelikow, P., & King, D. C. (Eds.), Why people don't trust government (133–154). Cambridge, MA: Harvard University Press.

(36)

Miller, A. H. (1974). Political issues and trust in Government: 1964-1970. The American Political Science Review, 68, 951–972.

Miller, M. M., & Reese, S. D. (1982). Media dependency as interaction: Effects of exposure and reliance on political activity and efficacy. Communication Research, 9(2), 227– 248.

Morris, S. D., & Klesner, J. L. (2010) Corruption and trust: theoretical considerations and evidence from Mexico. Comparative Political Studies, 43(10), 1258–1285.

Moy, P., & Scheufele, D. A. (2000). Media effects on political and social trust. Journalism & Mass Communication Quarterly, 77(4), 744–759.

Norris, P. (2000). A Virtuous Circle. Cambridge: Cambridge University Press.

Norris, P. (2002). Democratic phoenix: Reinventing political activism. Cambridge University Press.

Panayırcı, U. C., İşeri, E., & Şekercioğlu, E. (2016). Political agency of news outlets in a polarized media system: Framing the corruption probe in Turkey. European Journal of Communication, 31(5), 551–567.

Pfau, M., Moy, P., Radler, B., & Bridgeman, M. K. (1998). The influence of individual communication media on public confidence in democratic institutions. Southern Journal of Communication, 63(2), 91–112.

Price, V., Tewksbury, D., & Powers, E. (1997). Switching trains of thought: The impact of news frames on readers' cognitive responses. Communication Research, 24, 481–506. Price, V., & Zaller, J. (1993). Who gets the news? Alternative measures of news reception

and their implications for research. Public Opinion Quarterly, 57(2), 133–164. Putnam, R. D. (2000). Bowling alone: America’s declining social capital. In Crothers, L. &

Lockhart C. (Eds). Culture and politics (pp. 223–234). Palgrave Macmillan, New York.

(37)

Reinard, J. C. (2006). Communication research statistics. Thousand Oaks, CA: Sage Publications.

Robinson, M. (1976). Public affairs television and the growth of political malaise. American Political Science Review, 70, 409–432.

Seligson, M. A. (2002). The impact of corruption on regime legitimacy: A comparative study of four Latin American countries. Journal of Politics, 64(2), 408–33.

Semetko, H. A., & Valkenburg, P. M. (2000). Framing European politics: A content analysis of press and television news. Journal of Communication, 50(2), 93–109.

Solé-Ollé, A., & Sorribas-Navarro, P. (2014). Does corruption erode trust in government? Evidence from a recent surge of local scandals in Spain. Retrieved from:

http://www.ieb.ub.edu/phocadownload/documentostrabajo/doc2014-26.pdf

Transparency International (2018, February 21). Corruption perceptions index 2017. Retrieved from:

https://www.transparency.org/news/feature/corruption_perceptions_index_2017

Valentino, N. A., Beckmann, M. N., & Buhr, T. A. (2001). A spiral of cynicism for some: The contingent effects of campaign news frames on participation and confidence in government. Political Communication, 18(4), 347–367.

Valkenburg, P. M., Semetko, H. A., & De Vreese, C. H. (1999). The effects of news frames on readers' thoughts and recall. Communication Research, 26(5), 550–569.

Van der Meer, T. W. G. (2017). Political trust and the “Crisis of Democracy”. Oxford Research Encyclopedia of Politics. Retrieved from:

http://politics.oxfordre.com/view/10.1093/acrefore/9780190228637.001.0001/acrefore

9780190228637-e-77?print=pdf

(38)

Political Sophistication, and the Punishment of Corruption in Brazil. Journal of Politics, 79(1), 60–74.

Winters, M. S., & Weitz-Shapiro, R. (2016). Who’s in charge here? Direct and indirect accusations and voter punishment of corruption. Political Research Quarterly, 69(2), 207–219.

Zaller, J. (1992). The nature and origins of mass opinion. Cambridge university press. Zamora, R., & Albaladejo J. A. M. (2010). La representación simbólica del escándalo

político. Hacia una tipología de los marcos periodísticos (frames) utilizados en la narración del escándalo de corrupción. Razón y Palabra 15(73), 1–27.