Media and Corruption: The Other Way Round—Exploring Macro Determinants of Journalists’ Perceptions of the Accountability Instruments and Governmental Pressures

Media and Corruption: The Other Way Round—Exploring Macro Determinants of Journalists’... Abstract This article reverses the direction of the causal link between corruption and media, in particular, journalism, as it has been commonly investigated in communication studies. Few existing studies have explored the effects of corruption on journalists’ practices and self-conceptions. Based on a comparative survey of 1,764 journalists in 12 European countries and two Arab states, conducted in 2011 and 2012, this article investigates how perceived corruption—considered as an extra-media macro determinant—affects journalists’ perceptions of Media Accountability Instruments (MAI) impact and of governmental pressures on journalism quality. The empirical results obtained show that the more widespread corruption is in a country’s public sector, the lower the journalists’ perceptions of the MAI impact and the higher their perceptions of government pressures. News media and corruption have been often associated. The predominant view within political science and communication studies has been that the good functioning of the media is an obstacle to the spread of corruption (Brunetti & Weder, 2003; Camaj, 2013; Chowdury, 2004; Freille, Haque, & Kneller, 2007). This article reverses the causal relationship by investigating whether the perceived spread of corruption in the public sector affects journalists’ practices and professional self-conceptions. Contrary to the existing literature, the aim of this study is not to assess whether a free press can reduce corruption, but rather whether a context of widespread corruption may prevent journalists from being well-equipped to provide transparent and autonomous news content and to inform public opinion properly. More precisely, the article investigates the extent to which perceived corruption influences two different journalistic domains: (1) journalists’ perceptions of the impact of the media accountability instruments (MAI) as any non-State means of making media responsible toward the public; (2) journalists’ evaluations of governmental pressures. This article rests on two main causal assumptions: (1) journalists’ perceptions of their autonomy affect the way they do their work and, as a consequence, the content that they produce; (2) media coverage of corruption shapes public perceptions of corruption. Given the clandestine nature of corrupt interactions, it is particularly hard to provide objective measures of this phenomenon, especially in a cross-sectional empirical design. This is why corruption is often measured through the perceptions of experts as well as common citizens. While it is acknowledged that perceptions may not necessarily reflect the actual extent of corruption in a country, empirical studies have demonstrated that citizens’ opinions on corruption are not disconnected from those of experts (Pellegata & Memoli, 2016) and more in general from the reality (Charron, 2016). More importantly, citizens’ perceptions of corruption are per se particularly important because they strongly affect their attitudes toward parties and institutions and their electoral behavior (Clausen, Kraay & Nyiri, 2011). Citizens draw on various sources to build their corruption perceptions and, among them, the news media play a crucial role. It is therefore of great importance to shift scholarly attention to journalists’ attitudes, given their role as news producers. If corruption regarded as an extra-media context affects how journalists perceive the efficacy of those instruments developed specifically to increase transparency in their profession and how they perceive the level of government pressures, this most likely also has consequences on how they cover the issues of corruption and political influence in the media. We therefore postulate that, in countries plagued by higher levels of corruption, the latter pervades not only important sectors of the public system but also the society in general, fostering a “culture of distrust” in certain institutions and weakening the rule of law (Blais, Gidengil, & Kilibarda, 2015; Melgar, Rossi & Smith, 2010). This overall perception of clientelism, malpractices, and uncertain rules negatively affects citizens’ attitudes toward the elites and the society. The journalistic profession—which performs a crucial role in the good functioning of democracy and is also important for providing representations of corruption—is also affected by corruption. The results of an empirical analysis conducted on a sample of more than 1,700 journalists resident in 12 European countries and two Arab states confirm our hypotheses. Even when controlling the impact of alternative explanatory factors at the professional, organizational, and societal levels, corruption is detrimental to media accountability and the journalists’ perceived professional autonomy from governmental pressures. The article is structured as follows. The first section discusses media accountability instruments, explaining the role that they play for journalism, and how journalists’ perceptions of governmental pressures influence their profession. The second section sets out our theoretical argument, which relates corruption to journalists’ perceptions of the impact of MAI and governmental pressures, and states the research hypotheses. The third section describes the data and the variables used in the empirical analysis, whose results are discussed in the fourth section. A final section presents the main conclusions of the study. Journalists’ Perceptions of MAI and Governmental Pressures Within journalism studies, there are many theoretical analyses (Hallin & Mancini 2004; Shoemaker & Reese, 1996, 2014) and an increasing number of empirical inquiries that explore the factors and contexts that influence journalists’ practices and self-conceptions (Fengler, Eberwein, Mazzoleni, Porlezza, & Russ-Mohl, 2014; Fengler et al., 2015; Hanitzsch & Berganza, 2012; Hanitzsch & Mellado, 2011; Reich & Hanitzsch, 2013). Nonetheless, few attempts have been made to evaluate the role played by perceived corruption in affecting journalists’ practices and perceptions. MAI are considered journalistic tools intended to enhance transparency, to weaken corruption and malpractices in journalism, and to provide remedies for unethical conduct by journalists (Bertrand, 2000; Fengler et al., 2014; McQuail, 2003). They include well-established and discrete instruments (e.g., codes of ethics, press councils, ombudsman, media criticism in the mass media, correction boxes, letters to the editor) as well as innovative tools emerging online (e.g., editorial weblogs, webcasts of internal critique sessions, media-critical activities on Twitter and Facebook; see Bertrand, 2000). Such practices should attempt to balance freedom of the press with social responsibility at the levels of both media structure and media performance. McQuail (2005) refers to media accountability as “voluntary and involuntary processes by which the media answer directly or indirectly to their society for the quality and/or consequences of publication” (p. 2007). Overall, MAI should be understood as instruments intended to increase journalism’s capacity to enhance an informed public opinion (see Bertrand, 2000; McQuail, 1992). Several scholars have already investigated the concept of “media accountability cultures” (see Eberwein, Fengler, Lauk, & Leppik-Bork, 2011), studied how journalistic cultures shape journalists’ attitudes toward media accountability (Fengler et al., 2014, 2015), or measured the impact of MAI in terms of general cultures (Mazzoleni & Splendore, 2014). However, these studies fail to pursue a comprehensive and systematic organization of the three groups of determinants of journalists’ practices and self-conceptions—at the individual, societal, and media system levels—which are generally considered as separate from one another. More importantly, this literature does not empirically test the influence of factors external to media systems, such as corruption. Political influence is usually regarded as a decisive obstacle to journalists’ autonomy (Deuze, 2005; Hallin & Mancini, 2004). The recent literature comprises an increasing number of studies that comparatively investigate the determinants of journalists’ autonomy, focusing on journalists’ perceptions rather than on aspects that lie beyond their consciousness (Hanitzsch & Mellado, 2011; Reich & Hanitzsch, 2013). According to these studies, perceived influences concern journalists’ perceptions of the forces that shape the news production process. To operationalize political influence, this literature mainly relies on survey questions intended to assess journalists’ opinions on government officials, politicians, censorship, and business representatives. Several cross-sectional empirical analyses of the determinants of political autonomy at the individual, organizational, and societal level have been conducted. These studies demonstrate that journalists perceive a higher impact of political influence in contexts characterized by strong traditions of political parallelism and low levels of democratic performance and press freedom (see Hanitzsch & Mellado, 2011; Reich & Hanitzsch, 2013). Only few recent attempts have focused on the negative consequences of corruption on journalists’ perceptions of their role. However, these studies mainly focus of nondemocratic systems or newly established democracies characterized by developing economies. Some of them are qualitative analyses conducted in Asian and Middle Eastern political systems (Pintak & Ginges, 2009; Pintak & Nazir, 2013). Örnebring (2012) investigates the interconnection between clientelism and corruption in the media system of Central and Eastern Europe. Onyebadi & Alajmi (2014) instead consider a direct form of corruption (gift solicitation) among Kuwaiti journalists. Only Relly, Zanger, & Fahmi (2015), drawing on Shoemaker & Reese (1996, 2014), have considered corruption as pertaining to the extramedia dimension of influences on journalism. They accordingly use a quantitative approach to study the determinants of professional role perceptions among Iraqi Kurdish journalists. Hanitzsch & Berganza (2012), applying a similar approach, specifically demonstrate that more widespread corruption in a country’s public sector is associated with a lower level of journalists’ trust in public institutions. The two dependent variables considered here have a certain degree of association. It is plausible that, at least in some media systems, the more journalists perceive the impact of MAI, the more they are professionally autonomous. Nonetheless, we have decided to keep the two variables separate for three reasons. First, we have a sample of different countries with different media systems and journalistic cultures. MAI are particularly well established in liberal and corporatist models (Hallin & Mancini, 2004). Their impact and perceived importance can vary among journalistic cultures (i.e., their influence can be also perceived as political control instead of a source of autonomy). Second, while the influence of a “political” phenomenon like corruption on journalists’ perceptions of governmental pressures appears more plausible, its impact on MAI is much less clear and it has yet to be tested. The third reason is a statistical one: The degree of correlation between the two indicators used to operationalize the dependent variables (see Methods section) is low and not significant according to common statistical standards. Corruption as a Determinant of MAI Impact and Governmental Pressures Corruption is a complex phenomenon that is difficult to capture in a single definition (Andersson & Heywood, 2009; Lancaster & Montinola, 1997). There are many different activities that can be defined as corrupt, and they vary across countries and regions. A narrow definition, by now standard in comparative political studies, essentially considers corruption to be the misuse of public office for private financial gain (see Andersson & Heywood, 2009; Rose-Ackerman, 1999). This definition refers to a “political” concept of corruption because it indicates that each type of corrupt activity typically involves elected officials and/or appointed bureaucrats who abuse the power and authority vested in them for private gain. Such activities take place at the expense of the collective community, thereby violating the norms that regulate the public office. There is a large body of literature that analyzes the detrimental consequences of corruption on diverse aspects of political and economic performance, such as economic growth (Mauro, 1995), economic competition (Emerson, 2006), and quality of governance (Kaufmann, Kraay, & Mastruzzi, 2006). Particularly important are those studies that investigate the impact of corruption on citizens’ attitudes toward democratic institutions. They demonstrate that in countries with higher levels of corruption, citizens express lower confidence in national institutions such as the parliament and the government (Clausen et al., 2011; Mishler & Rose, 2001), and a lower degree of satisfaction with how the democratic process works (Anderson & Tverdova, 2003). In this article, we follow the above-outlined stream of literature but shift our attention to journalists responsible for depicting corruption itself. We postulate that corruption negatively affects journalists’ perceptions of journalistic practices—by their evaluations of the impact of MAI—and professional self-conceptions, in particular in regards to pressures from the government. When corruption is perceived as endemic, it pervades diverse important sectors of a society, from the highest levels of government to the desk-level bureaucrats and street-level public officials (Rose-Ackerman, 1999). Furthermore, corruption spreads also into the society, generating a “culture of distrust” in political authorities and reducing the quality of governance, and in particular the rule of law (Melgar et al., 2010; Morlino, 2011). Given the pivotal role of journalists in connecting politics and public opinion, they cannot be considered immune from the negative effects of corruption. From a field perspective (Benson & Neveu, 2005; Bourdieu, 1998), journalism is considered to be a semiautonomous field with its own logics and practices, but, at the same time, closely intertwined with the political and economic fields (Willig, 2012). One of the key analytical concepts of the field perspective is journalistic doxa. This concept refers to a set of implicit, tacit presuppositions in the journalistic field, not least the practical schemes that editors and reporters take for granted (Bourdieu, 1998). It represents the site where professional self-conceptions take shape. A context of widespread corruption affects the journalistic doxa. Like members of other organizations, journalists may adopt rationalization tactics that let them “accept” corruption and, in some sense, “take it for granted” and deny its negative impact (Anand, Ashforth, & Joshi, 2004). As a consequence, corruption should have detrimental consequences on how journalists perceive the effectiveness and the impact of MAI instruments, that is, precisely those instruments intended to improve the quality of their work by increasing transparency. This does not mean that journalists deliberately reject MAI, but rather that they consider them less effective. This expectation is in line with a more general argument, which states that where corruption is more widespread, anticorruption policies are less effective (Mungiu-Pippidi, 2006; Persson, Rothstein, & Teorell, 2013). We therefore expect that journalists who live and work in a political and social environment plagued by corruption have less confidence in the impact of MAI. Our first research hypothesis is as follows: H1: The higher the level of corruption in a country, the lower the journalists’ perception of MAI impact. When corruption perception is high and the rule of law is weak, politicians, in particular those with executive powers, tend to feel themselves allowed to act above the rules. Therefore, a corrupt political environment is a fertile ground for politicians to interfere in those spheres of the society in which they are not allowed by the law to impose control (Rose-Ackerman, 1999). As stressed by a plethora of reports on the threats to media freedom, this problem should be particularly evident in journalism, given its important role in reporting corruption and scandals affecting government authorities. Corruption is usually flanked by, and strictly connected to, clientelism and political parallelism. In his detailed study on the relationship among clientelism, corruption, and media in Central and Eastern Europe, Örnebring (2012) argues that media are also used as “elite-to-elite” communication tools. Informal clientelistic networks involve politics, business, as well as the media in many different exchanges of favors (Örnebring, 2012). Echoing Roudakova (2008), Örnebring (2012) assumes that clientelism and corruption contribute to producing a stable system where loyalties and relationships between media and politics are seen as predictable. Whether politicians effectively exert pressures on journalists, widespread corruption induces journalists themselves to perceive higher levels of governmental pressures that reduce the quality of their work. Therefore, our second research hypothesis is the following: H2: The higher the level of corruption in a country, the higher the journalists’ perception of government pressures. Although our causal argumentation mainly focuses on extramedia determinants of journalistic quality, in the analyses now described, we considered other possible dimensions of influence at both the individual and organizational level (see Hanitzsch & Berganza, 2012; Reich & Hanitzsch, 2013; Shoemaker & Reese, 1996, 2014). Methods To conduct our empirical analysis, we used data gathered from the MediaAct data set (Fengler et al., 2014). This is a comparative survey of 1,764 journalists in 12 European countries and two Arab states, conducted between May 2011 and March 2012 by the European Union (EU)-funded research project “Media Accountability and Transparency in Europe” (MediaAct). Countries included in the survey are Austria, Estonia, Finland, France, Germany, Italy, Jordan, the Netherlands, Poland, Romania, Spain, Switzerland, Tunisia, and the United Kingdom. As the components of the MediaAct consortium stated, a “most different system” design was followed and these countries were chosen to include different journalism cultures and media systems (Eberwein, Fengler, Philipp, & Ille, 2014, p. 70). It is for this reason that three Eastern European countries (Estonia, Poland, and Romania) and two Arab countries (Jordan and Tunisia) were included in the sample. The aim of the survey is to investigate journalists’ and media professionals’ opinions on media accountability. Moreover, the data set also includes information on sociodemographic and professional aspects of the respondents. The MediaAct study was as a quasi-experimental design, that could not aim to create entirely representative samples (Eberwein et al., 2014, p. 70). Nonetheless, using a theory-based sampling scheme, the research consortium was able to recruit comparable populations, despite the lack of reliable statistical data on journalists in some of the countries. To identify journalistic populations, the project considered as journalists all professional actors occupied full time or almost full time with the collection, description, and publication of relevant information in journalistic media. Each team involved in the project collected statistical data on the basic population of journalists in their country. The national sample size was determined from the response behavior in the pretests. Supplementary Appendix Table A1 shows the sizes of the populations and of the country samples. Journalists were selected through a two-step sampling scheme, which was adapted to cope with the idiosyncrasies of the different countries to reach a meaningful sample in each national survey (Eberwein et al., 2014, p. 72). The first step differentiated journalists along nine different types of media, while in the second step the quota criteria of journalistic hierarchy was applied.1 To fulfill the quotas and avoid biased results, the journalists in each category were randomly selected. The survey was conducted using an Internet-based method Computer Assisted Web Interviewing (CAWI). To test our research hypotheses, we developed two dependent variables measuring (1) the journalists’ perception of the impact of MAI and (2) the journalists’ perception of governmental pressures. Given the nature of the survey, our first dependent variable (MAI impact) does not measure the objective effects of the established media accountability instruments, but rather the journalists’ perception of their impact in terms of accountability. More precisely, the question administered to respondents was, “Which of the following have most impact on journalists’ behavior in [country]? Please rate on a scale from 1 (No impact at all) to 5 (Very high impact).” The 19 MAI instruments that respondents rated are the following: company editorial guidelines, ombudsman/reader’s editor, in-house media blog, legal department/media lawyer, user comments, laws regulating the media, professional codes of ethics, press council, regulatory authority, journalism trade journals, media criticism in the news media, online media criticism by journalist bloggers, print or broadcast satire/comedy about the media, blogs about the media, viewers’ associations, criticism on social media, journalism education, academic analysis of journalism, and NGOs/foundations. MAI impact is an additive index developed by aggregating the scores that respondents attributed to these instruments. The Cronbach’s alpha of MAI impact is equal to .86, indicating a rather high level of internal consistency of the elements composing the index. Considering the cultural and political differences among the countries composing our sample, we conducted a multigroup confirmatory factor analysis to test the invariance of factor loadings across countries. Results confirm the equivalence in the measurement of the perceived impact of MAI across countries. Considering how the question is phrased, the higher the value of MAI impact, the higher the journalists’ perception of the impact of these instruments on media accountability. The second dependent variable measures the journalists’ perceived governmental pressures (Government pressures). This variable was developed from the following question: “What do you regard to be a problem about journalism in [country]? Please rate on a scale from 1 (This is not a problem) to 5 (This is a major problem): Governmental pressure damages journalistic quality.” Thus, the higher the level of government pressures (meaning that pressure from government is considered a major problem), the higher the journalists’ perceptions of the presence of interferences by the government. For the sake of simplicity and to increase their comparability, the two dependent variables were normalized to vary on the same 0–1 scale. Supplementary Appendix Figure A1 provides a descriptive overview of country differences in the distribution of our dependent variables. The graph in the upper panel displays the mean and the standard deviation of MAI impact in each of the 14 countries included in the analysis, while the graph in the lower panel shows the mean level and the standard deviation of government pressures. Regarding the MAI impact (upper panel), we cannot detect a general pattern in the distribution of the data. Finnish journalists are those who perceive the highest impact of MAI, followed by journalists in Jordan and Spain. At the other extreme are Tunisia, Italy, and the Netherlands. Countries in Eastern Europe (Poland and Romania) and the Arab world (Jordan and Tunisia), where MAI are less consolidated, display higher standard deviations. Turning to government pressures (lower panel), a general pattern emerges whereby Western European countries (with the exception of Estonia)—usually associated with more transparent public sectors—record average levels of pressures lower than those of Southern and Eastern European countries and Jordan and Tunisia. Our main independent variable is the level of corruption of a country’s public sector. It was measured at the aggregate level. To operationalize the level of corruption, we relied on one of the indexes most commonly used in comparative research: the 2011 Corruption Perceptions Index (CPI) developed by the international network Transparency International. CPI is an annual index, released since 1995, that measures experts’ perceptions of corruption in the public sector. To make the interpretation of this index more intuitive, we inverted its original scale, making CPI range between 0 (lower corruption) and 10 (higher corruption). CPI is particularly useful for the purpose of our study because it provides a measure of the perceived level of corruption present in a country’s public sector, directly following the narrow definition of corruption that we have adopted, as explained in the previous section (Lambsdorff, 2006). We included in the analysis a set of country-level and individual-level control variables represented by potential explanatory factors of media accountability and perceptions of governmental pressures. At the individual level, we identified and tested objective individual factors referring to journalists’ profiles and careers. Among them, two variables related to sociodemographic aspects of respondents and two others referred to their profession. Gender is a dummy variable that assumes value of 1 for females and 0 for males. Education is an ordinal variable that measures the journalists’ educational level on a 6-point scale. Journalism training measures the formal training in journalism of respondents. This is an ordinal variable with three different modalities: “no formal training,” “informal training” (journalists who have received in-house training in a news outlet or possess a vocational diploma or a certificate in journalism), and “formal training” (journalists who attended a journalism school or had a university degree in journalism or a journalism-related postgraduate degree). Professional experience is an ordinal variable that defines the level of respondents’ experience in journalism in six different categories. The lowest category identifies journalists with <1 year of experience, while the highest category groups those with >20 years of experience. It is plausible that education, journalism training, and professional experience are positively associated with MAI impact. More educated and better trained journalists, as well as professionals with more experience, generally have greater competence with and better knowledge of the dynamics present in newsrooms. Higher experience and competence should induce respondents to perceive a greater impact of self-regulation instruments. On the contrary, although the existence of a causal relationship between these control variables and government pressure is plausible, we are agnostic about its sign because of the presence of opposite explanations. Another variable measures respondents’ opinions regarding the impact of newsroom organization pressure to produce quality journalism. Management role estimates the media professionals’ perceptions of the role played by the newsroom. This is gathered from the answers given to a question that specifically asked: How does your organization seek to maintain high journalistic standards? (1) The managers encourage newsroom debate about quality issues; (2) The management encourages and/or reacts to audience complaints; (3) The company Web site highlights my organization’s commitment to high standards; (4) The management contributes financial support to the press council/media regulator; (5) I would be called in by my editor if the integrity of my work were challenged by members of the public; (6) My supervisors acknowledge if members of the newsroom maintain high standards even under difficult circumstances. Please rate the following statements on a scale from 1 (I totally disagree) to 5 (I fully agree). The variable is built by summing the scores that journalists attributed to each of the six items proposed and normalizing the value to range on a 0–1 scale. The Cronbach’s alpha of management role is .74. The higher the level of management role, the higher the journalists’ perceptions of the impact of the newsroom organization. Following Fengler et al.’s (2015, p. 260) reasoning that newsrooms are key factors in encouraging or discouraging media accountability, we expect higher values of Management Role to be associated with higher values of MAI impact and lower values of government pressure. Journalists that operate in better-organized newsrooms that support their work and induce them to produce high-quality products should be more conscious of the effects of MAI and feel themselves more protected from governmental pressures. At the societal level, the effects of corruption are controlled for the role played by an important explanatory factor of different aspects of the quality of journalism. Freedom press is an index developed by Freedom House, which since 1980 has annually measured, through experts’ surveys, media freedom in all countries and territories around the world. It varies from 0 (the most free) to 100 (the least free). This index is widely used in the field and we expect it to be negatively associated with MAI impact and positively with government pressure. Although freedom press and CPI are highly and significantly correlated, they do not overlap conceptually. We decided to include freedom press in the analysis because it is an important and widely used indicator in this field of studies and, consequently, a serious potential source of spuriousness in our results. Variance Inflation Factor test excludes that multicollinearity biases the results of the empirical models including both CPI and freedom press. Nevertheless, we also ran our regressions substituting the continuous (0–100) freedom press score with a dummy variable, which assumes value 1 for countries rated as “Free” and 0 for countries rated as “Partly Free” or “Not Free.” (log)GDP per capita (taken from The World Bank’s World Development Indicators) measures the 2011 per capita GDP of each country included in the sample and indicates its level of economic performance. Supplementary Appendix Table A2 provides the descriptive statistics for all the variables included in the analysis. To appropriately address the hierarchical nature of the data set, which combined individual observations nested with information at the level of different countries, we applied a multilevel modeling procedure. We realize that with only 14 countries studied the number of highest level units is relatively small, but the literature suggests that a low number (even lower than 14) of highest level units does not necessarily affect regression coefficients (Maas & Hox, 2005). Moreover, the value of intraclass correlation coefficients indicates that the proportion of journalists’ perceived autonomy from governmental pressure that varies at the country level is around 24%, suggesting the use of multilevel modeling. Thus, considering the continuous nature of our dependent variables, we ran Generalized Least Squares (GLS) regressions. However, as the small number of the highest-level units might produce biased standard errors, we checked the robustness of multilevel regression models on replicated samples of data through the bootstrapping approach. Furthermore, we also decided to run our regressions also using pooled Ordinary Least Squares (OLS) with the errors clustered at the country level. By applying both these checks, our empirical results do not change.2 Results Table 1 presents the empirical results of six multilevel regressions. The first three regressions (1, 2, 3) have MAI impact as the dependent variable, while regressions 4, 5, and 6 has government pressure as the dependent variable. The first specification (1 and 4) regresses the two dependent variables, respectively, on CPI and all the controls except freedom press. The second specification (2 and 5) includes freedom press but excludes CPI. Finally, the last specification (3 and 6) regresses the dependent variables on both CPI and freedom press and all the other control variables. This empirical strategy enabled us to disentangle the net effect of CPI while keeping all the other alternative explanatory factors, in particular freedom press, under control. Table 1 Multivariate Regressions of media accountability instruments (MAI) Impact and Government Pressures on Corruption (Corruption Perceptions Index, CPI) and Controls Predictors MAI impact Government pressure Null (1) (2) (3) Null (4) (5) (6) Individual level     Gender (female) .033*** (.007) .033*** (.007) .034*** (.007) .036** (.015) .037** (.015) .037** (.015)     Education .001 (.003) .001 (.003) .002 (.003) −.008 (.007) −.007 (.007) −.007 (.007)     Journalism training:         Informal .025** (.012) .022* (.012) .024** (.012) −.020 (.024) −.019 (.024) −.0212 (.024)         Formal .030** (.012) .027** (.012) .030** (.012) −.023 (.023) −.022 (.023) −.024 (.023)     Prof. experience .001 (.002) .001 (.002) .001 (.002) .003 (.005) .003 (.005) .003 (.005)     Organization role .211*** (.015) .212*** (.015) .211*** (.015) −.021 (.030) −.022 (.030) −.0211 (.030) Country level     CPI −.014*** (.005) −.020*** (.005) .095*** (.021) .070*** (.021)     Freedom press .001 (.001) .002** (.001) .014*** (.004) .008** (.003)      (log)GDP per capita −.018* (.011) .09 (.017) −.001 (.012) .065 (.044) .100 (.062) .137*** (.047)     _const. .436*** (.126) .113 (.192) .233* (.139) −.452 (.498) −.840 (.703) −1.292** (.543)     Observations 1,640 1,640 1,640 1,585 1,585 1,585     Groups 14 14 14 14 14 14     Sigma_u .034*** (.007) .019*** (.006) .025*** (.006) .014*** (.005) .161*** (.031) .089*** (.019) .102*** (.021) .074*** (.016)     Sigma_e .155*** (.002) .141*** (.002) .141*** (.002) .141*** (.002) .279*** (.004) .273*** (.005) .273*** (.005) .273*** (.005)     Prop. of explained variance     Individual level .202 .198 .210 .228 .239 .247     Country level .758 .713 .804 .899 .899 .899     AIC −1535.9 −1732.1 −1726.6 −1734.5 552.6 444.9 448.1 441.8     BIC −1519.4 −1672.7 −1667.1 −1669.7 568.9 503.9 507.2 506.3 Predictors MAI impact Government pressure Null (1) (2) (3) Null (4) (5) (6) Individual level     Gender (female) .033*** (.007) .033*** (.007) .034*** (.007) .036** (.015) .037** (.015) .037** (.015)     Education .001 (.003) .001 (.003) .002 (.003) −.008 (.007) −.007 (.007) −.007 (.007)     Journalism training:         Informal .025** (.012) .022* (.012) .024** (.012) −.020 (.024) −.019 (.024) −.0212 (.024)         Formal .030** (.012) .027** (.012) .030** (.012) −.023 (.023) −.022 (.023) −.024 (.023)     Prof. experience .001 (.002) .001 (.002) .001 (.002) .003 (.005) .003 (.005) .003 (.005)     Organization role .211*** (.015) .212*** (.015) .211*** (.015) −.021 (.030) −.022 (.030) −.0211 (.030) Country level     CPI −.014*** (.005) −.020*** (.005) .095*** (.021) .070*** (.021)     Freedom press .001 (.001) .002** (.001) .014*** (.004) .008** (.003)      (log)GDP per capita −.018* (.011) .09 (.017) −.001 (.012) .065 (.044) .100 (.062) .137*** (.047)     _const. .436*** (.126) .113 (.192) .233* (.139) −.452 (.498) −.840 (.703) −1.292** (.543)     Observations 1,640 1,640 1,640 1,585 1,585 1,585     Groups 14 14 14 14 14 14     Sigma_u .034*** (.007) .019*** (.006) .025*** (.006) .014*** (.005) .161*** (.031) .089*** (.019) .102*** (.021) .074*** (.016)     Sigma_e .155*** (.002) .141*** (.002) .141*** (.002) .141*** (.002) .279*** (.004) .273*** (.005) .273*** (.005) .273*** (.005)     Prop. of explained variance     Individual level .202 .198 .210 .228 .239 .247     Country level .758 .713 .804 .899 .899 .899     AIC −1535.9 −1732.1 −1726.6 −1734.5 552.6 444.9 448.1 441.8     BIC −1519.4 −1672.7 −1667.1 −1669.7 568.9 503.9 507.2 506.3 Note. Robust standard errors in parentheses. *** p < .01; **p < .05; *p < .1. Table 1 Multivariate Regressions of media accountability instruments (MAI) Impact and Government Pressures on Corruption (Corruption Perceptions Index, CPI) and Controls Predictors MAI impact Government pressure Null (1) (2) (3) Null (4) (5) (6) Individual level     Gender (female) .033*** (.007) .033*** (.007) .034*** (.007) .036** (.015) .037** (.015) .037** (.015)     Education .001 (.003) .001 (.003) .002 (.003) −.008 (.007) −.007 (.007) −.007 (.007)     Journalism training:         Informal .025** (.012) .022* (.012) .024** (.012) −.020 (.024) −.019 (.024) −.0212 (.024)         Formal .030** (.012) .027** (.012) .030** (.012) −.023 (.023) −.022 (.023) −.024 (.023)     Prof. experience .001 (.002) .001 (.002) .001 (.002) .003 (.005) .003 (.005) .003 (.005)     Organization role .211*** (.015) .212*** (.015) .211*** (.015) −.021 (.030) −.022 (.030) −.0211 (.030) Country level     CPI −.014*** (.005) −.020*** (.005) .095*** (.021) .070*** (.021)     Freedom press .001 (.001) .002** (.001) .014*** (.004) .008** (.003)      (log)GDP per capita −.018* (.011) .09 (.017) −.001 (.012) .065 (.044) .100 (.062) .137*** (.047)     _const. .436*** (.126) .113 (.192) .233* (.139) −.452 (.498) −.840 (.703) −1.292** (.543)     Observations 1,640 1,640 1,640 1,585 1,585 1,585     Groups 14 14 14 14 14 14     Sigma_u .034*** (.007) .019*** (.006) .025*** (.006) .014*** (.005) .161*** (.031) .089*** (.019) .102*** (.021) .074*** (.016)     Sigma_e .155*** (.002) .141*** (.002) .141*** (.002) .141*** (.002) .279*** (.004) .273*** (.005) .273*** (.005) .273*** (.005)     Prop. of explained variance     Individual level .202 .198 .210 .228 .239 .247     Country level .758 .713 .804 .899 .899 .899     AIC −1535.9 −1732.1 −1726.6 −1734.5 552.6 444.9 448.1 441.8     BIC −1519.4 −1672.7 −1667.1 −1669.7 568.9 503.9 507.2 506.3 Predictors MAI impact Government pressure Null (1) (2) (3) Null (4) (5) (6) Individual level     Gender (female) .033*** (.007) .033*** (.007) .034*** (.007) .036** (.015) .037** (.015) .037** (.015)     Education .001 (.003) .001 (.003) .002 (.003) −.008 (.007) −.007 (.007) −.007 (.007)     Journalism training:         Informal .025** (.012) .022* (.012) .024** (.012) −.020 (.024) −.019 (.024) −.0212 (.024)         Formal .030** (.012) .027** (.012) .030** (.012) −.023 (.023) −.022 (.023) −.024 (.023)     Prof. experience .001 (.002) .001 (.002) .001 (.002) .003 (.005) .003 (.005) .003 (.005)     Organization role .211*** (.015) .212*** (.015) .211*** (.015) −.021 (.030) −.022 (.030) −.0211 (.030) Country level     CPI −.014*** (.005) −.020*** (.005) .095*** (.021) .070*** (.021)     Freedom press .001 (.001) .002** (.001) .014*** (.004) .008** (.003)      (log)GDP per capita −.018* (.011) .09 (.017) −.001 (.012) .065 (.044) .100 (.062) .137*** (.047)     _const. .436*** (.126) .113 (.192) .233* (.139) −.452 (.498) −.840 (.703) −1.292** (.543)     Observations 1,640 1,640 1,640 1,585 1,585 1,585     Groups 14 14 14 14 14 14     Sigma_u .034*** (.007) .019*** (.006) .025*** (.006) .014*** (.005) .161*** (.031) .089*** (.019) .102*** (.021) .074*** (.016)     Sigma_e .155*** (.002) .141*** (.002) .141*** (.002) .141*** (.002) .279*** (.004) .273*** (.005) .273*** (.005) .273*** (.005)     Prop. of explained variance     Individual level .202 .198 .210 .228 .239 .247     Country level .758 .713 .804 .899 .899 .899     AIC −1535.9 −1732.1 −1726.6 −1734.5 552.6 444.9 448.1 441.8     BIC −1519.4 −1672.7 −1667.1 −1669.7 568.9 503.9 507.2 506.3 Note. Robust standard errors in parentheses. *** p < .01; **p < .05; *p < .1. Regressions 1, 2, and 3 confirm the validity of H1. As expected, CPI is always negatively and significantly related to MAI impact, meaning that lower perceived corruption is associated with higher journalists’ perceptions of the impact of MAI. To be stressed is that the significance of the CPI regression coefficient is not affected by the inclusion of freedom press in the analysis. Even when controlling for the impact of media freedom, corruption remains a significant explanatory factor of journalists’ perceived impact of MAI. Moreover, as the Akaike Information Criterion (AIC) and Bayesian information criterion (BIC) values presented at the bottom of Table 1 show, the specifications including CPI among the regressors are those of a higher quality given the data distribution. In particular, a likelihood ratio test confirms that the full specification (Model 3) is the one that best fits our data. Besides this main result, other control variables display a significant effect on MAI impact. Management role positively and significantly impacts on journalists’ perceptions of the impact of MAI. The more positive the journalists’ opinions of a well-organized newsroom, the higher their perceptions of MAI impact. Furthermore, females, as well as journalists with formal training, perceive a stronger impact of MAI than do males and journalists with no professional training or who are only informally trained. Freedom press is significantly associated with MAI impact only in regression 3. Surprisingly, the regression coefficient of freedom press has a positive sign, which is in contrast with our expectations. However, given the high level of correlation between freedom press and CPI, this result could be an “effect” of multicollinearity. On substituting the continuous score of press freedom with the dummy described in the previous section, freedom press turns out to be nonsignificant. The proportion of variance explained by this model specification (Model 3) is 21% at the individual and 80% at the country level.3 Regressions 4, 5, and 6 highlight a positive and significant relationship between CPI and government pressure which confirms H2. Also, freedom press is significantly (and positively, as expected) related to government pressure. However, even on introducing this variable as a control in Regression 6, the coefficient of CPI maintains the same level of significance. Moreover, also in these regressions, if we substitute the continuous score of press freedom with the dummy indicating countries rated as “Free,” the effect of freedom press turns out to be no significant. Not surprisingly, the proportion of variance in journalists’ perceptions of government pressure explained at the country level (79%) is much higher than that explained at the individual level (25%). Finally, the values of AIC and BIC show that regressions that include CPI among the explanatory factors (4 and 6) are better specified according to our data. A likelihood ratio test confirms the best goodness of fit of Model 6, which includes both CPI and freedom press. At the individual level, only gender is significantly related to government pressure. Its regression coefficient displays a positive sign, indicating that females tend to perceive more governmental pressures. By contrast, management role, a variable strongly associated with the other dependent variable, does not display a significant effect. At the country level, though only in Regression 6, (log)GDP per capita displays a positive and significant regression coefficient, which is in contrast with our expectation. However, on plotting the distribution of countries according to the average values of these two variables, it seems that this effect is mostly driven by countries such as Austria, Italy, and Spain, which present good economic performances but, at the same time, high levels of perceived governmental pressures (see Mazzoleni & Splendore, 2014). This conundrum should be the subject of future, more in-depth analyses. To increase the robustness of our results, we carried out a set of statistical checks. First, we built an alternative index measuring journalists’ perceptions of MAI impact, choosing only the most traditional and institutionalized instruments and not considering the online ones. Even when substituting MAI impact with the alternative indicator in the regressions presented above, empirical analyses confirmed the validity of the hypotheses advanced. Second, we tested the effects of alternative indicators of corruption: The World Bank’s Control of Corruption Index (CC) and the Transparency International’s Global Corruption Barometer (GCB). Like CPI also CC is developed by aggregating a large number of surveys on experts, and the two indicators are almost perfectly correlated (r = .98, p < .001). The GCB instead is a large mass survey that, among other aspects, monitors citizens’ experience of corruption. Thus, GCB measures the proportion of respondents who in 2011 reported having paid a bribe in the last 12 months. Whereas CPI and CC are considered subjective indicators because they are developed on the basis of experts’ perceptions, various scholars refer to GCB as an objective indicator of the real spread of corruption in a country because it is developed on the basis of experiences reported by common citizens (Clausen et al., 2011; Treisman, 2007). Both CC and GCB display the same effects on the two dependent variables tested in the empirical analysis, confirming the hypotheses postulated above. This check is particularly important because it demonstrates that our findings are not affected by the method that we used to measure the level of corruption, an issue that is much debated in comparative analysis (Treisman, 2007). Third, even on including a dummy variable that assumed value 1 for nondemocratic regimes, such as Jordan and Tunisia, or on excluding these two countries from the analyses, our main empirical results did not substantially change. However, it is worth noting that while CPI regression coefficients remain statistically significant, freedom press is never significantly associated with either MAI impact or government pressure. Finally, considering that our dependent variables were both left- and right-censored, because in their normalized form they cannot assume values lower than 0 and higher than 1, we reran our models adopting Tobit regressions developed to explicitly address censored dependent variables (Takeshi, 1984). The results obtained from the Tobit regressions were the same as the main results obtained running GLS regressions. Discussion In this study, we have focused on the role played by corruption in affecting journalists’ perceptions of accountability and autonomy in their profession. Our empirical findings have confirmed our research hypotheses, demonstrating that journalists who live and work in countries with a more widespread corruption perceive a lower impact of MAI and more pressures from the government. Journalism is not immune from the detrimental consequences generated by corruption. Corruption, in fact, seems to affect the journalistic doxa, the set of taken-for-granted understandings that guide journalistic practice. The presence of endemic corruption in a system affects journalists’ performances and, consequently, weakens their resolve to produce transparent and accountable journalism. Although the proportion of variance explained by our regression models at the individual level is only around 20%, it is between 70% and 80% at the country level. These findings show how much important are external factors in affecting journalists’ perceptions of accountability and transparency in their profession, in particular, extramedia factors, such as corruption, that do not pertain to journalistic culture, but influence anyway journalistic practices. Nonetheless, there are significant differences in the relationships between corruption and the two dependent variables used in our study. The graph in Figure 1 plots the 14 sample countries according to their CPI score and the country respondents’ average value of MAI impact. The solid line indicates the best linear fit between the two variables according to the data distribution, and shows the negative relationship between corruption and journalists’ perception of MAI impact in producing good-quality journalism. At the same time, the graph highlights the presence of some outliers: in particular Jordan, which displays a particularly high average value of MAI impact. In that country, data were gathered during the so-called Arab Spring, and Jordanian journalists may have reflected that experience, particularly in regard to online MAI (see Fengler et al., 2014). It is not by chance that if we focus on the alternative measure of MAI impact, which takes only the most traditional and institutionalized instruments into account, the association with corruption becomes stronger, with Jordan (and also Spain) displaying lower perceived impact of MAI, in line with countries recording similar corruption scores. Figure 1 View largeDownload slide Linear relationship between perceived corruption (Corruption Perceptions Index, CPI) and journalists’ perception of media accountability instruments (MAI) impact (country average values). Note: Perceived corruption (CPI) on the X-axis. The original 0–10 scale of CPI was inverted, so that higher values of CPI represent higher levels of perceived corruption. Journalists’ perception of MAI impact on the Y-axis. Values of MAI impact are standardized on a 0–1 scale. Higher values of MAI impact refer to journalists’ perception of a stronger impact of MAI on the profession. Figure 1 View largeDownload slide Linear relationship between perceived corruption (Corruption Perceptions Index, CPI) and journalists’ perception of media accountability instruments (MAI) impact (country average values). Note: Perceived corruption (CPI) on the X-axis. The original 0–10 scale of CPI was inverted, so that higher values of CPI represent higher levels of perceived corruption. Journalists’ perception of MAI impact on the Y-axis. Values of MAI impact are standardized on a 0–1 scale. Higher values of MAI impact refer to journalists’ perception of a stronger impact of MAI on the profession. As stressed by the Figure 2, associating the country average CPI scores and Government pressures levels, the relationship between these two variables is now positive and stronger than that between corruption and MAI impact. Journalists who live and work in countries plagued by high levels of corruption tend to perceive higher pressures from governmental authorities. More precisely, though with some distinctions, the graph marks a clear difference between two groups of countries. On the one hand, there are Northern and Western European countries (plus Estonia) characterized by low levels of both corruption and perceived government pressures; on the other, Southern and Eastern European countries and non-European ones with high levels of both corruption and government pressures. Figure 2 View largeDownload slide Linear relationship between perceived corruption (CPI) and journalists’ perception of government pressures (country average values). Note: Perceived corruption (CPI) on the X-axis. The original 0–10 scale of CPI was inverted, so that higher values of CPI represent higher levels of perceived corruption. Journalists’ perception of government pressure (Gov. pressure) on the Y-axis. Values of Gov. pressure are standardized on a 0–1 scale. Higher values of Gov. pressure refer to journalists’ perception of a higher pressure by the government on their job. Figure 2 View largeDownload slide Linear relationship between perceived corruption (CPI) and journalists’ perception of government pressures (country average values). Note: Perceived corruption (CPI) on the X-axis. The original 0–10 scale of CPI was inverted, so that higher values of CPI represent higher levels of perceived corruption. Journalists’ perception of government pressure (Gov. pressure) on the Y-axis. Values of Gov. pressure are standardized on a 0–1 scale. Higher values of Gov. pressure refer to journalists’ perception of a higher pressure by the government on their job. In his seminal work on media accountability, Bertrand (2000, p. 151) affirms that “ethics are not enough,” and that not all journalists are “endowed with a moral sense” (p. 41). His suggestion is therefore that media need accountability instruments to enforce the norms adopted in a given journalistic culture. Our empirical findings demonstrate that corruption jeopardizes journalists’ perceptions of the impact of any instruments that the journalistic field may implement to make journalists accountable. What this article firmly suggests is that considering journalists’ perceptions of their role and their place within the society as a whole (McQuail, 2013) depends also on the society itself, its functioning, and its overall level of corruption. If we consider journalists as a particular category of professionals who provide information about corruption, but who—regardless of the role they cover—are themselves influenced by corruption, this analysis sheds new light on how citizens’ opinions on corruption are linked to those of experts (Pellegata & Memoli, 2016). In environments plagued by corruption, journalists appear to be structurally unable to perform their normative role as watchdogs. Like common citizens, journalists also take corruption for granted, thus reinforcing public opinion perceptions in a “spiral of cynicism.” There are several caveats concerning this study that should be stressed and that should be subject to future research. Although the analysis involved 12 European countries and two Arab states, and although several robustness checks seem to indicate that the empirical results are indifferent to the inclusion or exclusion of the latter, our findings can hardly be generalized to many non-European countries. Media accountability and journalistic autonomy from governmental pressures are mainly conceived in Western and liberal terms. MAI in particular may not have been understood and adopted in the same way across our sample countries. To generalize our conclusions to areas different from Europe, future analyses should cover more non-European countries. Another limitation of this study concerns the potential problem of reciprocal causality between corruption and our dependent variables. Considering that our measure of corruption is developed on the basis of experts’ perceptions, and given the crucial role played by journalists in representing the issue of corruption, reversing our causal argumentation could also be plausible. It may be that experts’ (and even more, citizens’) perceptions of corruption in the public sector are affected by the fact that journalists consider the instruments developed to increase transparency in their work to be ineffective and feel pressures by the government authorities. Whether corruption causes a lower perceived impact of MAI and higher perceived pressures from the government or vice versa is almost impossible to determine with survey data. Further research using more “objective” indicators rather than subjective perceptions is certainly necessary to investigate this relation. Nonetheless, in conclusion we believe that our study offers important contributions to the literature on journalistic accountability and on corruption. First, the article highlights the importance of considering the extramedia dimension when journalists’ perceptions and practices are studied. It is now important to investigate how corruption and diverse journalistic cultures intertwine and are related to each other. Furthermore, our study shows that a category of professionals who play a crucial role in representing corruption is affected by the spread of corruption in the country where they operate. This finding should induce scholars of corruption to reflect on how the relationship between corruption and journalists’ perceptions and representations of it may affect citizens’ opinions of the spread of corruption itself in the public sector. Accordingly, in light of that relation between corruption and journalism also experts’ evaluations, on which the most common measures of perceived corruption are build, should be reassessed. Supplementary Data Supplementary Data are available at IJPOR online. Alessandro Pellegata is a research fellow at the Department of Social and Political Sciences of the University of Milano (Italy). His main research interests are in the field of comparative politics, in particular electoral competition, political representation, public opinion, and corruption. He has published articles in several academic journals, such as Democratization, International Political Science Review, Journal of Information Technology & Politics, and Social Indicators Research. Sergio Splendore is an assistant professor at the Department of Social and Political Sciences of the University of Milano (Italy) His research centers on media sociology with a deep interest in journalism studies and media production. He has authored or coauthored papers in these areas for Journalism, Media, Culture and Society, Journalism Practice, and Oxford Bibliographies in Communication. 1The nine media types considered are daily newspapers, weekly newspapers, magazines, public service radio, private commercial radio, public service television, private commercial television, online news media, and news agencies. Regarding journalistic hierarchy, MediaAct distinguishes between a management level and an operational level. 2Results of all the robustness checks discussed here and at the end of the next section are available on request from the authors. 3The proportion of variance explained at different levels was computed using the formula provided in Snijder and Bosker (1999: p. 101–105). Acknowledgement The authors thank Prof. Susanne Fengler for inspiration and the MediaAct consortium that allowed us to use its data set. We also thank the three referees of the International Journal of Public Opinion Research for their most instructive comments.  This study is based on survey data developed by the project entitled “Media Accountability and Transparency in Europe (MediaAct),” funded by the European Union Seventh Framework Programme (2007-2013. Grant Agreement number 244147. Project Coordinator: Susanne Fengler). References Anand V. , Ashforth B. E., , Joshi M. ( 2004 ). Business as usual: The acceptance and perpetuation of corruption in organizations . Academy of Management Executive ,ssss 18 , sssss – 53 . doi: 10.5465/AME.2004.13837437 Google Scholar CrossRef Search ADS Andersson S., , Heywood P. M. ( 2009 ). The politics of perception: Use and abuse of transparency international's approach to measuring corruption . Political Studies , 57 , 746 – 767 . Doi: 10.1111/j.1467-9248.2008.00758.x. Google Scholar CrossRef Search ADS Anderson C. J., , Tverdova Y. V. ( 2003 ). Corruption, political allegiances and attitudes toward government in contemporary democracies . American Journal of Political Science , 47 ( 1 ), 91 – 109 . doi: 10.1111/1540-5907.00007 Google Scholar CrossRef Search ADS Bertrand C. J. ( 2000 ). Media ethics & accountability systems . New Brunswick, London : Transaction Publishers . Benson R., , Neveu E. ( 2005 ). Introduction: Field theory as a work in progress. In Benson R. , Neveu E. (Eds.), Bourdieu and the journalistic field (pp. 1 – 29 ). Cambridge : Polity Press . Blais A. , Gidengil E., , Kilibarda A. ( 2015 ). Partisanship, information, and perceptions of government corruption . International Journal of Public Opinion Research , 29 , 95 – 110 . doi:10.1093/ijpor/edv037 Bourdieu P. ( 1998 ). Practical reason. On the theory of action . Cambridge : Polity Press . Brunetti A., , Weder B. ( 2003 ). A free press is a bad news for corruption . Journal of Public Economics , 87 , 1081 – 1824 . Google Scholar CrossRef Search ADS Camaj L. ( 2013 ). The media’s role in fighting corruption: Media effects on government accountability . The International Journal of Press/Politics , 18 , 21 – 42 . doi: 10.1177/1940161212462741 Google Scholar CrossRef Search ADS Charron N. ( 2016 ). Do corruption measures have a perception problem? Assessing the relationship between experiences and perceptions of corruption among citizens and experts . European Political Science Review , 8 ( 1 ), 147 – 171 . doi: 10.1017/S1755773914000447 Google Scholar CrossRef Search ADS Clausen B. , Kraay A., , Nyiri Z. ( 2011 ). Corruption and confidence in public institution: Evidence from a global survey . The World Bank Economic Review , 25 , 212 – 249 . doi: 10.1093/wber/lhr018 Google Scholar CrossRef Search ADS Deuze M. ( 2005 ). What is journalism? Professional identity and ideology of journalists reconsidered . Journalism , 6 , 442 – 464 . doi: 10.1177/1464884905056815 Google Scholar CrossRef Search ADS Emerson P. M. ( 2006 ). Corruption, competition and democracy . Journal of Development Economics , 81 ( 1 ), 193 – 212 . Google Scholar CrossRef Search ADS Eberwein T. , Fengler S. , Lauk E. , Leppik-Bork T. (Eds.). ( 2011 ). Mapping media accountability—In Europe and beyond . Berlin : Herbert von Halem Verlag . Eberwein T. , Fengler S. , Philipp S. , Ille M. ( 2014 ). Counting media accountability—The concept and methodology of the mediaact survey. In Fengler S. , Eberwein T. , Mazzoleni G. , Porlezza C. , Russ-Mohl S. (Eds.), Journalists and media accountability. An international study of news people in the digital age (pp. 65 – 79 ). New York, NY : Peter Lang . Fengler S. ( 2012 ). From media self-regulation to ‘crowd-criticism’: Media accountability in the digital age . Central-European journal of communication , 2 , 175 – 188 . Fengler S. , Eberwein T. , Alsius S. , Baisnée O. , Bichler K. et al. ( 2015 ). How effective is media self-regulation? Results from a comparative survey of european journalists . European journal of communication , 30 , 249 – 266 . doi: 10.1177/0267323114561009 Google Scholar CrossRef Search ADS Freille S. , Haque E. M., , Kneller R. ( 2007 ). A contribution to the empirics of press freedom and corruption . European Journal of Political Economy , 23 , 838 – 862 . doi: 10.1016/j.ejpoleco.2007.03.002 Google Scholar CrossRef Search ADS Hallin D. C. , Mancini P. ( 2004 ). Comparing media systems. Three models of media and politics . Cambridge : Cambridge University Press . Hanitzsch T., , Berganza R. ( 2012 ). Explaining journalists’ trust in public institutions across 20 countries: Media freedom, corruption, and ownership matter most . Journal of Communication , 62 , 794 – 814 . doi: 10.1111/j.1460-2466.2012.01663.x Google Scholar CrossRef Search ADS Hanitzsch T., , Mellado C. ( 2011 ). What shapes the news around the world? How journalists in eighteen countries perceive influences on their work . The International Journal of Press/Politics , 16 , 404 – 426 . doi: 10.1177/1940161211407334 Google Scholar CrossRef Search ADS Kaufmann D. , Kraay A., , Mastruzzi M. ( 2006 ). Measuring governance using cross-country perceptions data. In Rose-Ackerman S. (Ed.) International handbook on the economics of corruption (pp. 52 – 85 ). Northampton, MA : Edward Elgan Publishing . Lambsdorff J. G. ( 2006 ). Measuring corruption: The validity and precision of subjective indicators. In Sampford C. , Shacklock A. , Connors C. , Galtung F. (Eds.), Measuring corruption (pp. 81 – 100 ). Burlington, VT : Ashgate . Maas C. J., , Hox J. J. ( 2005 ). Sufficient sample sizes for multilevel modeling . Methodology , 1 , 85 – 91 . doi: 10.1027/1614-1881.1.3.86 Google Scholar CrossRef Search ADS Mauro P. ( 1995 ). Corruption and growth . The quarterly journal of economics , 110 , 681 – 712 . doi: 10.2307/2946696 Google Scholar CrossRef Search ADS Mazzoleni G. , Splendore S. ( 2014 ). Cultures of accountability'™ – or global trends in media accountability? The Hallin and Mancini model revisited. In: S. Fengler , C. Porlezza , T. Erbenwein , G. Mazzoleni , S. Russ-Mohl (eds.), Journalists and Media Accountability. An International Study of News People in the Digital Age ( 145 – 159 ), London : Peter Lang . McQuail D. ( 1992 ). Media performance . London : Sage . McQuail D. ( 2003 ). Media accountability and freedom of publication . Oxford : Oxford University . McQuail D. ( 2005 ). McQuail’s mass communication theory ( 5th ed.). London : Sage . McQuail D. ( 2013 ). Journalism and society . London : Sage . Melgar N. , Rossi M., , Smith T. S. ( 2010 ). The perception of corruption . International Journal of Public Opinion Research , 22 ( 1 ), 120 – 131 . doi:10.1093/ijpor/edp058 Google Scholar CrossRef Search ADS Mishler W., , Rose R. ( 2001 ). What are the origins of political trust? Testing institutional and cultural theories in post-communist societies . Comparative Political Studies , 34 ( 1 ), 30 – 62 . doi: 10.1177/0010414001034001002 Google Scholar CrossRef Search ADS Morlino L. ( 2011 ). Changes for democracy. Actors, structures, processes . Oxford : Oxford University Press . Mungiu-Pippidi A. ( 2006 ). Corruption: Diagnosis and treatment . Journal of Democracy , 17 , 86 – 99 . doi: 10.1353/jod.2006.0050 Google Scholar CrossRef Search ADS Onyebadi U., , Alajmi F. ( 2014 ). Gift solicitation and acceptance in journalism practice: An assessment of kuwaiti journalists perspective . Journalism , 17 , 348 – 365 . doi: 10.1177/1464884914557924 Google Scholar CrossRef Search ADS Örnebring H. ( 2012 ). Clientelism, elites and the media in central and eastern Europe . The International Journal of Press/Politics , 17 , 497 – 515 . doi: 10.1177/1940161212454329 Google Scholar CrossRef Search ADS Pellegata A., , Memoli V. ( 2016 ). Can corruption erode confidence in political institutions among european countries? Comparing the effects of different measures of perceived corruption . Social Indicators Research , 128 ( 1 ), 391 – 412 . doi:10.1007/s11205-015-1036-0 Google Scholar CrossRef Search ADS Persson A. , Rothstein B., , Teorell J. ( 2013 ). Why anticorruption reforms fail—Systemic corruption as a collective action problem . Governance , 26 , 449 – 471 . doi: 10.1111/j.1468-0491.2012.01604.x Google Scholar CrossRef Search ADS Pintak L., , Ginges J. ( 2009 ). Inside the Arab newsroom . Journalism Studies , 10 , 157 – 177 . doi: 10.1080/14616700802337800 Google Scholar CrossRef Search ADS Pintak L., , Nazir S. J. ( 2013 ). Pakistani journalism: At the crossroads of Muslim identity, national priorities and journalistic culture . Media, Culture and Society , 35 , 640 – 665 . doi: 10.1177/016344371348365 Google Scholar CrossRef Search ADS Reich Z., , Hanitzsch T. ( 2013 ). Determinants of journalists' professional autonomy: Individual and national level factors matter more than organizational ones . Mass Communication and Society , 16 ( 1 ), 133 – 156 . doi: 10.1080/15205436.2012.669002 Google Scholar CrossRef Search ADS Relly J. , Zanger M., , Fahmy S. ( 2015 ). Professional role perception among Iraqui Kurdish journalists from a ‘State within a State’ . Journalism , 16 , 1085 – 1106 . doi: 10.1177/1464884914550973 Google Scholar CrossRef Search ADS Rose-Ackerman S. ( 1999 ). Corruption and government. Causes, consequences and reforms . Cambridge : Cambridge University Press . Roudakova N. ( 2008 ). Media-political clientelism: Lessons from anthropology . Media, Culture and Society , 30 ( 1 ), 41 – 59 . doi: 10.1177/0163443707084349 Google Scholar CrossRef Search ADS Shoemaker P. J., , Reese S. D. ( 1996 ). Mediating the message: Theories of influence on mass media content . New York, NY : Longman . Shoemaker P. J., , Reese S. D. ( 2014 ). Mediating the message in the 21st century. A media sociology perspective . Routledge : London . Snijders T. A., , Bosker R. J. ( 1999 ). Multilevel analysis. An introduction to basic and advanced multilevel modeling . London : Sage . Takeshi A. ( 1984 ). Tobit Models: A Survey . Journal of Econometrics , 24 , 3 – 61 . Google Scholar CrossRef Search ADS Treisman D. ( 2007 ). What we have learned about the causes of corruption from ten years of cross-national empirical research? Annual Review of Political Science , 10 , 211 – 244 . doi: 10.1146/annurev.polisci.10.081205.095418 Google Scholar CrossRef Search ADS Willig I. ( 2012 ). News ethnography in a field perspective . Journalism , 14 , 372 – 387 . doi: 10.1177/1464884912442638. Google Scholar CrossRef Search ADS © The Author 2017. Published by Oxford University Press on behalf of The World Association for Public Opinion Research. All rights reserved. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png International Journal of Public Opinion Research Oxford University Press

Media and Corruption: The Other Way Round—Exploring Macro Determinants of Journalists’ Perceptions of the Accountability Instruments and Governmental Pressures

Loading next page...
 
/lp/ou_press/media-and-corruption-the-other-way-round-exploring-macro-determinants-divEmjWLDV
Publisher
Oxford University Press
Copyright
© The Author 2017. Published by Oxford University Press on behalf of The World Association for Public Opinion Research. All rights reserved.
ISSN
0954-2892
eISSN
1471-6909
D.O.I.
10.1093/ijpor/edx008
Publisher site
See Article on Publisher Site

Abstract

Abstract This article reverses the direction of the causal link between corruption and media, in particular, journalism, as it has been commonly investigated in communication studies. Few existing studies have explored the effects of corruption on journalists’ practices and self-conceptions. Based on a comparative survey of 1,764 journalists in 12 European countries and two Arab states, conducted in 2011 and 2012, this article investigates how perceived corruption—considered as an extra-media macro determinant—affects journalists’ perceptions of Media Accountability Instruments (MAI) impact and of governmental pressures on journalism quality. The empirical results obtained show that the more widespread corruption is in a country’s public sector, the lower the journalists’ perceptions of the MAI impact and the higher their perceptions of government pressures. News media and corruption have been often associated. The predominant view within political science and communication studies has been that the good functioning of the media is an obstacle to the spread of corruption (Brunetti & Weder, 2003; Camaj, 2013; Chowdury, 2004; Freille, Haque, & Kneller, 2007). This article reverses the causal relationship by investigating whether the perceived spread of corruption in the public sector affects journalists’ practices and professional self-conceptions. Contrary to the existing literature, the aim of this study is not to assess whether a free press can reduce corruption, but rather whether a context of widespread corruption may prevent journalists from being well-equipped to provide transparent and autonomous news content and to inform public opinion properly. More precisely, the article investigates the extent to which perceived corruption influences two different journalistic domains: (1) journalists’ perceptions of the impact of the media accountability instruments (MAI) as any non-State means of making media responsible toward the public; (2) journalists’ evaluations of governmental pressures. This article rests on two main causal assumptions: (1) journalists’ perceptions of their autonomy affect the way they do their work and, as a consequence, the content that they produce; (2) media coverage of corruption shapes public perceptions of corruption. Given the clandestine nature of corrupt interactions, it is particularly hard to provide objective measures of this phenomenon, especially in a cross-sectional empirical design. This is why corruption is often measured through the perceptions of experts as well as common citizens. While it is acknowledged that perceptions may not necessarily reflect the actual extent of corruption in a country, empirical studies have demonstrated that citizens’ opinions on corruption are not disconnected from those of experts (Pellegata & Memoli, 2016) and more in general from the reality (Charron, 2016). More importantly, citizens’ perceptions of corruption are per se particularly important because they strongly affect their attitudes toward parties and institutions and their electoral behavior (Clausen, Kraay & Nyiri, 2011). Citizens draw on various sources to build their corruption perceptions and, among them, the news media play a crucial role. It is therefore of great importance to shift scholarly attention to journalists’ attitudes, given their role as news producers. If corruption regarded as an extra-media context affects how journalists perceive the efficacy of those instruments developed specifically to increase transparency in their profession and how they perceive the level of government pressures, this most likely also has consequences on how they cover the issues of corruption and political influence in the media. We therefore postulate that, in countries plagued by higher levels of corruption, the latter pervades not only important sectors of the public system but also the society in general, fostering a “culture of distrust” in certain institutions and weakening the rule of law (Blais, Gidengil, & Kilibarda, 2015; Melgar, Rossi & Smith, 2010). This overall perception of clientelism, malpractices, and uncertain rules negatively affects citizens’ attitudes toward the elites and the society. The journalistic profession—which performs a crucial role in the good functioning of democracy and is also important for providing representations of corruption—is also affected by corruption. The results of an empirical analysis conducted on a sample of more than 1,700 journalists resident in 12 European countries and two Arab states confirm our hypotheses. Even when controlling the impact of alternative explanatory factors at the professional, organizational, and societal levels, corruption is detrimental to media accountability and the journalists’ perceived professional autonomy from governmental pressures. The article is structured as follows. The first section discusses media accountability instruments, explaining the role that they play for journalism, and how journalists’ perceptions of governmental pressures influence their profession. The second section sets out our theoretical argument, which relates corruption to journalists’ perceptions of the impact of MAI and governmental pressures, and states the research hypotheses. The third section describes the data and the variables used in the empirical analysis, whose results are discussed in the fourth section. A final section presents the main conclusions of the study. Journalists’ Perceptions of MAI and Governmental Pressures Within journalism studies, there are many theoretical analyses (Hallin & Mancini 2004; Shoemaker & Reese, 1996, 2014) and an increasing number of empirical inquiries that explore the factors and contexts that influence journalists’ practices and self-conceptions (Fengler, Eberwein, Mazzoleni, Porlezza, & Russ-Mohl, 2014; Fengler et al., 2015; Hanitzsch & Berganza, 2012; Hanitzsch & Mellado, 2011; Reich & Hanitzsch, 2013). Nonetheless, few attempts have been made to evaluate the role played by perceived corruption in affecting journalists’ practices and perceptions. MAI are considered journalistic tools intended to enhance transparency, to weaken corruption and malpractices in journalism, and to provide remedies for unethical conduct by journalists (Bertrand, 2000; Fengler et al., 2014; McQuail, 2003). They include well-established and discrete instruments (e.g., codes of ethics, press councils, ombudsman, media criticism in the mass media, correction boxes, letters to the editor) as well as innovative tools emerging online (e.g., editorial weblogs, webcasts of internal critique sessions, media-critical activities on Twitter and Facebook; see Bertrand, 2000). Such practices should attempt to balance freedom of the press with social responsibility at the levels of both media structure and media performance. McQuail (2005) refers to media accountability as “voluntary and involuntary processes by which the media answer directly or indirectly to their society for the quality and/or consequences of publication” (p. 2007). Overall, MAI should be understood as instruments intended to increase journalism’s capacity to enhance an informed public opinion (see Bertrand, 2000; McQuail, 1992). Several scholars have already investigated the concept of “media accountability cultures” (see Eberwein, Fengler, Lauk, & Leppik-Bork, 2011), studied how journalistic cultures shape journalists’ attitudes toward media accountability (Fengler et al., 2014, 2015), or measured the impact of MAI in terms of general cultures (Mazzoleni & Splendore, 2014). However, these studies fail to pursue a comprehensive and systematic organization of the three groups of determinants of journalists’ practices and self-conceptions—at the individual, societal, and media system levels—which are generally considered as separate from one another. More importantly, this literature does not empirically test the influence of factors external to media systems, such as corruption. Political influence is usually regarded as a decisive obstacle to journalists’ autonomy (Deuze, 2005; Hallin & Mancini, 2004). The recent literature comprises an increasing number of studies that comparatively investigate the determinants of journalists’ autonomy, focusing on journalists’ perceptions rather than on aspects that lie beyond their consciousness (Hanitzsch & Mellado, 2011; Reich & Hanitzsch, 2013). According to these studies, perceived influences concern journalists’ perceptions of the forces that shape the news production process. To operationalize political influence, this literature mainly relies on survey questions intended to assess journalists’ opinions on government officials, politicians, censorship, and business representatives. Several cross-sectional empirical analyses of the determinants of political autonomy at the individual, organizational, and societal level have been conducted. These studies demonstrate that journalists perceive a higher impact of political influence in contexts characterized by strong traditions of political parallelism and low levels of democratic performance and press freedom (see Hanitzsch & Mellado, 2011; Reich & Hanitzsch, 2013). Only few recent attempts have focused on the negative consequences of corruption on journalists’ perceptions of their role. However, these studies mainly focus of nondemocratic systems or newly established democracies characterized by developing economies. Some of them are qualitative analyses conducted in Asian and Middle Eastern political systems (Pintak & Ginges, 2009; Pintak & Nazir, 2013). Örnebring (2012) investigates the interconnection between clientelism and corruption in the media system of Central and Eastern Europe. Onyebadi & Alajmi (2014) instead consider a direct form of corruption (gift solicitation) among Kuwaiti journalists. Only Relly, Zanger, & Fahmi (2015), drawing on Shoemaker & Reese (1996, 2014), have considered corruption as pertaining to the extramedia dimension of influences on journalism. They accordingly use a quantitative approach to study the determinants of professional role perceptions among Iraqi Kurdish journalists. Hanitzsch & Berganza (2012), applying a similar approach, specifically demonstrate that more widespread corruption in a country’s public sector is associated with a lower level of journalists’ trust in public institutions. The two dependent variables considered here have a certain degree of association. It is plausible that, at least in some media systems, the more journalists perceive the impact of MAI, the more they are professionally autonomous. Nonetheless, we have decided to keep the two variables separate for three reasons. First, we have a sample of different countries with different media systems and journalistic cultures. MAI are particularly well established in liberal and corporatist models (Hallin & Mancini, 2004). Their impact and perceived importance can vary among journalistic cultures (i.e., their influence can be also perceived as political control instead of a source of autonomy). Second, while the influence of a “political” phenomenon like corruption on journalists’ perceptions of governmental pressures appears more plausible, its impact on MAI is much less clear and it has yet to be tested. The third reason is a statistical one: The degree of correlation between the two indicators used to operationalize the dependent variables (see Methods section) is low and not significant according to common statistical standards. Corruption as a Determinant of MAI Impact and Governmental Pressures Corruption is a complex phenomenon that is difficult to capture in a single definition (Andersson & Heywood, 2009; Lancaster & Montinola, 1997). There are many different activities that can be defined as corrupt, and they vary across countries and regions. A narrow definition, by now standard in comparative political studies, essentially considers corruption to be the misuse of public office for private financial gain (see Andersson & Heywood, 2009; Rose-Ackerman, 1999). This definition refers to a “political” concept of corruption because it indicates that each type of corrupt activity typically involves elected officials and/or appointed bureaucrats who abuse the power and authority vested in them for private gain. Such activities take place at the expense of the collective community, thereby violating the norms that regulate the public office. There is a large body of literature that analyzes the detrimental consequences of corruption on diverse aspects of political and economic performance, such as economic growth (Mauro, 1995), economic competition (Emerson, 2006), and quality of governance (Kaufmann, Kraay, & Mastruzzi, 2006). Particularly important are those studies that investigate the impact of corruption on citizens’ attitudes toward democratic institutions. They demonstrate that in countries with higher levels of corruption, citizens express lower confidence in national institutions such as the parliament and the government (Clausen et al., 2011; Mishler & Rose, 2001), and a lower degree of satisfaction with how the democratic process works (Anderson & Tverdova, 2003). In this article, we follow the above-outlined stream of literature but shift our attention to journalists responsible for depicting corruption itself. We postulate that corruption negatively affects journalists’ perceptions of journalistic practices—by their evaluations of the impact of MAI—and professional self-conceptions, in particular in regards to pressures from the government. When corruption is perceived as endemic, it pervades diverse important sectors of a society, from the highest levels of government to the desk-level bureaucrats and street-level public officials (Rose-Ackerman, 1999). Furthermore, corruption spreads also into the society, generating a “culture of distrust” in political authorities and reducing the quality of governance, and in particular the rule of law (Melgar et al., 2010; Morlino, 2011). Given the pivotal role of journalists in connecting politics and public opinion, they cannot be considered immune from the negative effects of corruption. From a field perspective (Benson & Neveu, 2005; Bourdieu, 1998), journalism is considered to be a semiautonomous field with its own logics and practices, but, at the same time, closely intertwined with the political and economic fields (Willig, 2012). One of the key analytical concepts of the field perspective is journalistic doxa. This concept refers to a set of implicit, tacit presuppositions in the journalistic field, not least the practical schemes that editors and reporters take for granted (Bourdieu, 1998). It represents the site where professional self-conceptions take shape. A context of widespread corruption affects the journalistic doxa. Like members of other organizations, journalists may adopt rationalization tactics that let them “accept” corruption and, in some sense, “take it for granted” and deny its negative impact (Anand, Ashforth, & Joshi, 2004). As a consequence, corruption should have detrimental consequences on how journalists perceive the effectiveness and the impact of MAI instruments, that is, precisely those instruments intended to improve the quality of their work by increasing transparency. This does not mean that journalists deliberately reject MAI, but rather that they consider them less effective. This expectation is in line with a more general argument, which states that where corruption is more widespread, anticorruption policies are less effective (Mungiu-Pippidi, 2006; Persson, Rothstein, & Teorell, 2013). We therefore expect that journalists who live and work in a political and social environment plagued by corruption have less confidence in the impact of MAI. Our first research hypothesis is as follows: H1: The higher the level of corruption in a country, the lower the journalists’ perception of MAI impact. When corruption perception is high and the rule of law is weak, politicians, in particular those with executive powers, tend to feel themselves allowed to act above the rules. Therefore, a corrupt political environment is a fertile ground for politicians to interfere in those spheres of the society in which they are not allowed by the law to impose control (Rose-Ackerman, 1999). As stressed by a plethora of reports on the threats to media freedom, this problem should be particularly evident in journalism, given its important role in reporting corruption and scandals affecting government authorities. Corruption is usually flanked by, and strictly connected to, clientelism and political parallelism. In his detailed study on the relationship among clientelism, corruption, and media in Central and Eastern Europe, Örnebring (2012) argues that media are also used as “elite-to-elite” communication tools. Informal clientelistic networks involve politics, business, as well as the media in many different exchanges of favors (Örnebring, 2012). Echoing Roudakova (2008), Örnebring (2012) assumes that clientelism and corruption contribute to producing a stable system where loyalties and relationships between media and politics are seen as predictable. Whether politicians effectively exert pressures on journalists, widespread corruption induces journalists themselves to perceive higher levels of governmental pressures that reduce the quality of their work. Therefore, our second research hypothesis is the following: H2: The higher the level of corruption in a country, the higher the journalists’ perception of government pressures. Although our causal argumentation mainly focuses on extramedia determinants of journalistic quality, in the analyses now described, we considered other possible dimensions of influence at both the individual and organizational level (see Hanitzsch & Berganza, 2012; Reich & Hanitzsch, 2013; Shoemaker & Reese, 1996, 2014). Methods To conduct our empirical analysis, we used data gathered from the MediaAct data set (Fengler et al., 2014). This is a comparative survey of 1,764 journalists in 12 European countries and two Arab states, conducted between May 2011 and March 2012 by the European Union (EU)-funded research project “Media Accountability and Transparency in Europe” (MediaAct). Countries included in the survey are Austria, Estonia, Finland, France, Germany, Italy, Jordan, the Netherlands, Poland, Romania, Spain, Switzerland, Tunisia, and the United Kingdom. As the components of the MediaAct consortium stated, a “most different system” design was followed and these countries were chosen to include different journalism cultures and media systems (Eberwein, Fengler, Philipp, & Ille, 2014, p. 70). It is for this reason that three Eastern European countries (Estonia, Poland, and Romania) and two Arab countries (Jordan and Tunisia) were included in the sample. The aim of the survey is to investigate journalists’ and media professionals’ opinions on media accountability. Moreover, the data set also includes information on sociodemographic and professional aspects of the respondents. The MediaAct study was as a quasi-experimental design, that could not aim to create entirely representative samples (Eberwein et al., 2014, p. 70). Nonetheless, using a theory-based sampling scheme, the research consortium was able to recruit comparable populations, despite the lack of reliable statistical data on journalists in some of the countries. To identify journalistic populations, the project considered as journalists all professional actors occupied full time or almost full time with the collection, description, and publication of relevant information in journalistic media. Each team involved in the project collected statistical data on the basic population of journalists in their country. The national sample size was determined from the response behavior in the pretests. Supplementary Appendix Table A1 shows the sizes of the populations and of the country samples. Journalists were selected through a two-step sampling scheme, which was adapted to cope with the idiosyncrasies of the different countries to reach a meaningful sample in each national survey (Eberwein et al., 2014, p. 72). The first step differentiated journalists along nine different types of media, while in the second step the quota criteria of journalistic hierarchy was applied.1 To fulfill the quotas and avoid biased results, the journalists in each category were randomly selected. The survey was conducted using an Internet-based method Computer Assisted Web Interviewing (CAWI). To test our research hypotheses, we developed two dependent variables measuring (1) the journalists’ perception of the impact of MAI and (2) the journalists’ perception of governmental pressures. Given the nature of the survey, our first dependent variable (MAI impact) does not measure the objective effects of the established media accountability instruments, but rather the journalists’ perception of their impact in terms of accountability. More precisely, the question administered to respondents was, “Which of the following have most impact on journalists’ behavior in [country]? Please rate on a scale from 1 (No impact at all) to 5 (Very high impact).” The 19 MAI instruments that respondents rated are the following: company editorial guidelines, ombudsman/reader’s editor, in-house media blog, legal department/media lawyer, user comments, laws regulating the media, professional codes of ethics, press council, regulatory authority, journalism trade journals, media criticism in the news media, online media criticism by journalist bloggers, print or broadcast satire/comedy about the media, blogs about the media, viewers’ associations, criticism on social media, journalism education, academic analysis of journalism, and NGOs/foundations. MAI impact is an additive index developed by aggregating the scores that respondents attributed to these instruments. The Cronbach’s alpha of MAI impact is equal to .86, indicating a rather high level of internal consistency of the elements composing the index. Considering the cultural and political differences among the countries composing our sample, we conducted a multigroup confirmatory factor analysis to test the invariance of factor loadings across countries. Results confirm the equivalence in the measurement of the perceived impact of MAI across countries. Considering how the question is phrased, the higher the value of MAI impact, the higher the journalists’ perception of the impact of these instruments on media accountability. The second dependent variable measures the journalists’ perceived governmental pressures (Government pressures). This variable was developed from the following question: “What do you regard to be a problem about journalism in [country]? Please rate on a scale from 1 (This is not a problem) to 5 (This is a major problem): Governmental pressure damages journalistic quality.” Thus, the higher the level of government pressures (meaning that pressure from government is considered a major problem), the higher the journalists’ perceptions of the presence of interferences by the government. For the sake of simplicity and to increase their comparability, the two dependent variables were normalized to vary on the same 0–1 scale. Supplementary Appendix Figure A1 provides a descriptive overview of country differences in the distribution of our dependent variables. The graph in the upper panel displays the mean and the standard deviation of MAI impact in each of the 14 countries included in the analysis, while the graph in the lower panel shows the mean level and the standard deviation of government pressures. Regarding the MAI impact (upper panel), we cannot detect a general pattern in the distribution of the data. Finnish journalists are those who perceive the highest impact of MAI, followed by journalists in Jordan and Spain. At the other extreme are Tunisia, Italy, and the Netherlands. Countries in Eastern Europe (Poland and Romania) and the Arab world (Jordan and Tunisia), where MAI are less consolidated, display higher standard deviations. Turning to government pressures (lower panel), a general pattern emerges whereby Western European countries (with the exception of Estonia)—usually associated with more transparent public sectors—record average levels of pressures lower than those of Southern and Eastern European countries and Jordan and Tunisia. Our main independent variable is the level of corruption of a country’s public sector. It was measured at the aggregate level. To operationalize the level of corruption, we relied on one of the indexes most commonly used in comparative research: the 2011 Corruption Perceptions Index (CPI) developed by the international network Transparency International. CPI is an annual index, released since 1995, that measures experts’ perceptions of corruption in the public sector. To make the interpretation of this index more intuitive, we inverted its original scale, making CPI range between 0 (lower corruption) and 10 (higher corruption). CPI is particularly useful for the purpose of our study because it provides a measure of the perceived level of corruption present in a country’s public sector, directly following the narrow definition of corruption that we have adopted, as explained in the previous section (Lambsdorff, 2006). We included in the analysis a set of country-level and individual-level control variables represented by potential explanatory factors of media accountability and perceptions of governmental pressures. At the individual level, we identified and tested objective individual factors referring to journalists’ profiles and careers. Among them, two variables related to sociodemographic aspects of respondents and two others referred to their profession. Gender is a dummy variable that assumes value of 1 for females and 0 for males. Education is an ordinal variable that measures the journalists’ educational level on a 6-point scale. Journalism training measures the formal training in journalism of respondents. This is an ordinal variable with three different modalities: “no formal training,” “informal training” (journalists who have received in-house training in a news outlet or possess a vocational diploma or a certificate in journalism), and “formal training” (journalists who attended a journalism school or had a university degree in journalism or a journalism-related postgraduate degree). Professional experience is an ordinal variable that defines the level of respondents’ experience in journalism in six different categories. The lowest category identifies journalists with <1 year of experience, while the highest category groups those with >20 years of experience. It is plausible that education, journalism training, and professional experience are positively associated with MAI impact. More educated and better trained journalists, as well as professionals with more experience, generally have greater competence with and better knowledge of the dynamics present in newsrooms. Higher experience and competence should induce respondents to perceive a greater impact of self-regulation instruments. On the contrary, although the existence of a causal relationship between these control variables and government pressure is plausible, we are agnostic about its sign because of the presence of opposite explanations. Another variable measures respondents’ opinions regarding the impact of newsroom organization pressure to produce quality journalism. Management role estimates the media professionals’ perceptions of the role played by the newsroom. This is gathered from the answers given to a question that specifically asked: How does your organization seek to maintain high journalistic standards? (1) The managers encourage newsroom debate about quality issues; (2) The management encourages and/or reacts to audience complaints; (3) The company Web site highlights my organization’s commitment to high standards; (4) The management contributes financial support to the press council/media regulator; (5) I would be called in by my editor if the integrity of my work were challenged by members of the public; (6) My supervisors acknowledge if members of the newsroom maintain high standards even under difficult circumstances. Please rate the following statements on a scale from 1 (I totally disagree) to 5 (I fully agree). The variable is built by summing the scores that journalists attributed to each of the six items proposed and normalizing the value to range on a 0–1 scale. The Cronbach’s alpha of management role is .74. The higher the level of management role, the higher the journalists’ perceptions of the impact of the newsroom organization. Following Fengler et al.’s (2015, p. 260) reasoning that newsrooms are key factors in encouraging or discouraging media accountability, we expect higher values of Management Role to be associated with higher values of MAI impact and lower values of government pressure. Journalists that operate in better-organized newsrooms that support their work and induce them to produce high-quality products should be more conscious of the effects of MAI and feel themselves more protected from governmental pressures. At the societal level, the effects of corruption are controlled for the role played by an important explanatory factor of different aspects of the quality of journalism. Freedom press is an index developed by Freedom House, which since 1980 has annually measured, through experts’ surveys, media freedom in all countries and territories around the world. It varies from 0 (the most free) to 100 (the least free). This index is widely used in the field and we expect it to be negatively associated with MAI impact and positively with government pressure. Although freedom press and CPI are highly and significantly correlated, they do not overlap conceptually. We decided to include freedom press in the analysis because it is an important and widely used indicator in this field of studies and, consequently, a serious potential source of spuriousness in our results. Variance Inflation Factor test excludes that multicollinearity biases the results of the empirical models including both CPI and freedom press. Nevertheless, we also ran our regressions substituting the continuous (0–100) freedom press score with a dummy variable, which assumes value 1 for countries rated as “Free” and 0 for countries rated as “Partly Free” or “Not Free.” (log)GDP per capita (taken from The World Bank’s World Development Indicators) measures the 2011 per capita GDP of each country included in the sample and indicates its level of economic performance. Supplementary Appendix Table A2 provides the descriptive statistics for all the variables included in the analysis. To appropriately address the hierarchical nature of the data set, which combined individual observations nested with information at the level of different countries, we applied a multilevel modeling procedure. We realize that with only 14 countries studied the number of highest level units is relatively small, but the literature suggests that a low number (even lower than 14) of highest level units does not necessarily affect regression coefficients (Maas & Hox, 2005). Moreover, the value of intraclass correlation coefficients indicates that the proportion of journalists’ perceived autonomy from governmental pressure that varies at the country level is around 24%, suggesting the use of multilevel modeling. Thus, considering the continuous nature of our dependent variables, we ran Generalized Least Squares (GLS) regressions. However, as the small number of the highest-level units might produce biased standard errors, we checked the robustness of multilevel regression models on replicated samples of data through the bootstrapping approach. Furthermore, we also decided to run our regressions also using pooled Ordinary Least Squares (OLS) with the errors clustered at the country level. By applying both these checks, our empirical results do not change.2 Results Table 1 presents the empirical results of six multilevel regressions. The first three regressions (1, 2, 3) have MAI impact as the dependent variable, while regressions 4, 5, and 6 has government pressure as the dependent variable. The first specification (1 and 4) regresses the two dependent variables, respectively, on CPI and all the controls except freedom press. The second specification (2 and 5) includes freedom press but excludes CPI. Finally, the last specification (3 and 6) regresses the dependent variables on both CPI and freedom press and all the other control variables. This empirical strategy enabled us to disentangle the net effect of CPI while keeping all the other alternative explanatory factors, in particular freedom press, under control. Table 1 Multivariate Regressions of media accountability instruments (MAI) Impact and Government Pressures on Corruption (Corruption Perceptions Index, CPI) and Controls Predictors MAI impact Government pressure Null (1) (2) (3) Null (4) (5) (6) Individual level     Gender (female) .033*** (.007) .033*** (.007) .034*** (.007) .036** (.015) .037** (.015) .037** (.015)     Education .001 (.003) .001 (.003) .002 (.003) −.008 (.007) −.007 (.007) −.007 (.007)     Journalism training:         Informal .025** (.012) .022* (.012) .024** (.012) −.020 (.024) −.019 (.024) −.0212 (.024)         Formal .030** (.012) .027** (.012) .030** (.012) −.023 (.023) −.022 (.023) −.024 (.023)     Prof. experience .001 (.002) .001 (.002) .001 (.002) .003 (.005) .003 (.005) .003 (.005)     Organization role .211*** (.015) .212*** (.015) .211*** (.015) −.021 (.030) −.022 (.030) −.0211 (.030) Country level     CPI −.014*** (.005) −.020*** (.005) .095*** (.021) .070*** (.021)     Freedom press .001 (.001) .002** (.001) .014*** (.004) .008** (.003)      (log)GDP per capita −.018* (.011) .09 (.017) −.001 (.012) .065 (.044) .100 (.062) .137*** (.047)     _const. .436*** (.126) .113 (.192) .233* (.139) −.452 (.498) −.840 (.703) −1.292** (.543)     Observations 1,640 1,640 1,640 1,585 1,585 1,585     Groups 14 14 14 14 14 14     Sigma_u .034*** (.007) .019*** (.006) .025*** (.006) .014*** (.005) .161*** (.031) .089*** (.019) .102*** (.021) .074*** (.016)     Sigma_e .155*** (.002) .141*** (.002) .141*** (.002) .141*** (.002) .279*** (.004) .273*** (.005) .273*** (.005) .273*** (.005)     Prop. of explained variance     Individual level .202 .198 .210 .228 .239 .247     Country level .758 .713 .804 .899 .899 .899     AIC −1535.9 −1732.1 −1726.6 −1734.5 552.6 444.9 448.1 441.8     BIC −1519.4 −1672.7 −1667.1 −1669.7 568.9 503.9 507.2 506.3 Predictors MAI impact Government pressure Null (1) (2) (3) Null (4) (5) (6) Individual level     Gender (female) .033*** (.007) .033*** (.007) .034*** (.007) .036** (.015) .037** (.015) .037** (.015)     Education .001 (.003) .001 (.003) .002 (.003) −.008 (.007) −.007 (.007) −.007 (.007)     Journalism training:         Informal .025** (.012) .022* (.012) .024** (.012) −.020 (.024) −.019 (.024) −.0212 (.024)         Formal .030** (.012) .027** (.012) .030** (.012) −.023 (.023) −.022 (.023) −.024 (.023)     Prof. experience .001 (.002) .001 (.002) .001 (.002) .003 (.005) .003 (.005) .003 (.005)     Organization role .211*** (.015) .212*** (.015) .211*** (.015) −.021 (.030) −.022 (.030) −.0211 (.030) Country level     CPI −.014*** (.005) −.020*** (.005) .095*** (.021) .070*** (.021)     Freedom press .001 (.001) .002** (.001) .014*** (.004) .008** (.003)      (log)GDP per capita −.018* (.011) .09 (.017) −.001 (.012) .065 (.044) .100 (.062) .137*** (.047)     _const. .436*** (.126) .113 (.192) .233* (.139) −.452 (.498) −.840 (.703) −1.292** (.543)     Observations 1,640 1,640 1,640 1,585 1,585 1,585     Groups 14 14 14 14 14 14     Sigma_u .034*** (.007) .019*** (.006) .025*** (.006) .014*** (.005) .161*** (.031) .089*** (.019) .102*** (.021) .074*** (.016)     Sigma_e .155*** (.002) .141*** (.002) .141*** (.002) .141*** (.002) .279*** (.004) .273*** (.005) .273*** (.005) .273*** (.005)     Prop. of explained variance     Individual level .202 .198 .210 .228 .239 .247     Country level .758 .713 .804 .899 .899 .899     AIC −1535.9 −1732.1 −1726.6 −1734.5 552.6 444.9 448.1 441.8     BIC −1519.4 −1672.7 −1667.1 −1669.7 568.9 503.9 507.2 506.3 Note. Robust standard errors in parentheses. *** p < .01; **p < .05; *p < .1. Table 1 Multivariate Regressions of media accountability instruments (MAI) Impact and Government Pressures on Corruption (Corruption Perceptions Index, CPI) and Controls Predictors MAI impact Government pressure Null (1) (2) (3) Null (4) (5) (6) Individual level     Gender (female) .033*** (.007) .033*** (.007) .034*** (.007) .036** (.015) .037** (.015) .037** (.015)     Education .001 (.003) .001 (.003) .002 (.003) −.008 (.007) −.007 (.007) −.007 (.007)     Journalism training:         Informal .025** (.012) .022* (.012) .024** (.012) −.020 (.024) −.019 (.024) −.0212 (.024)         Formal .030** (.012) .027** (.012) .030** (.012) −.023 (.023) −.022 (.023) −.024 (.023)     Prof. experience .001 (.002) .001 (.002) .001 (.002) .003 (.005) .003 (.005) .003 (.005)     Organization role .211*** (.015) .212*** (.015) .211*** (.015) −.021 (.030) −.022 (.030) −.0211 (.030) Country level     CPI −.014*** (.005) −.020*** (.005) .095*** (.021) .070*** (.021)     Freedom press .001 (.001) .002** (.001) .014*** (.004) .008** (.003)      (log)GDP per capita −.018* (.011) .09 (.017) −.001 (.012) .065 (.044) .100 (.062) .137*** (.047)     _const. .436*** (.126) .113 (.192) .233* (.139) −.452 (.498) −.840 (.703) −1.292** (.543)     Observations 1,640 1,640 1,640 1,585 1,585 1,585     Groups 14 14 14 14 14 14     Sigma_u .034*** (.007) .019*** (.006) .025*** (.006) .014*** (.005) .161*** (.031) .089*** (.019) .102*** (.021) .074*** (.016)     Sigma_e .155*** (.002) .141*** (.002) .141*** (.002) .141*** (.002) .279*** (.004) .273*** (.005) .273*** (.005) .273*** (.005)     Prop. of explained variance     Individual level .202 .198 .210 .228 .239 .247     Country level .758 .713 .804 .899 .899 .899     AIC −1535.9 −1732.1 −1726.6 −1734.5 552.6 444.9 448.1 441.8     BIC −1519.4 −1672.7 −1667.1 −1669.7 568.9 503.9 507.2 506.3 Predictors MAI impact Government pressure Null (1) (2) (3) Null (4) (5) (6) Individual level     Gender (female) .033*** (.007) .033*** (.007) .034*** (.007) .036** (.015) .037** (.015) .037** (.015)     Education .001 (.003) .001 (.003) .002 (.003) −.008 (.007) −.007 (.007) −.007 (.007)     Journalism training:         Informal .025** (.012) .022* (.012) .024** (.012) −.020 (.024) −.019 (.024) −.0212 (.024)         Formal .030** (.012) .027** (.012) .030** (.012) −.023 (.023) −.022 (.023) −.024 (.023)     Prof. experience .001 (.002) .001 (.002) .001 (.002) .003 (.005) .003 (.005) .003 (.005)     Organization role .211*** (.015) .212*** (.015) .211*** (.015) −.021 (.030) −.022 (.030) −.0211 (.030) Country level     CPI −.014*** (.005) −.020*** (.005) .095*** (.021) .070*** (.021)     Freedom press .001 (.001) .002** (.001) .014*** (.004) .008** (.003)      (log)GDP per capita −.018* (.011) .09 (.017) −.001 (.012) .065 (.044) .100 (.062) .137*** (.047)     _const. .436*** (.126) .113 (.192) .233* (.139) −.452 (.498) −.840 (.703) −1.292** (.543)     Observations 1,640 1,640 1,640 1,585 1,585 1,585     Groups 14 14 14 14 14 14     Sigma_u .034*** (.007) .019*** (.006) .025*** (.006) .014*** (.005) .161*** (.031) .089*** (.019) .102*** (.021) .074*** (.016)     Sigma_e .155*** (.002) .141*** (.002) .141*** (.002) .141*** (.002) .279*** (.004) .273*** (.005) .273*** (.005) .273*** (.005)     Prop. of explained variance     Individual level .202 .198 .210 .228 .239 .247     Country level .758 .713 .804 .899 .899 .899     AIC −1535.9 −1732.1 −1726.6 −1734.5 552.6 444.9 448.1 441.8     BIC −1519.4 −1672.7 −1667.1 −1669.7 568.9 503.9 507.2 506.3 Note. Robust standard errors in parentheses. *** p < .01; **p < .05; *p < .1. Regressions 1, 2, and 3 confirm the validity of H1. As expected, CPI is always negatively and significantly related to MAI impact, meaning that lower perceived corruption is associated with higher journalists’ perceptions of the impact of MAI. To be stressed is that the significance of the CPI regression coefficient is not affected by the inclusion of freedom press in the analysis. Even when controlling for the impact of media freedom, corruption remains a significant explanatory factor of journalists’ perceived impact of MAI. Moreover, as the Akaike Information Criterion (AIC) and Bayesian information criterion (BIC) values presented at the bottom of Table 1 show, the specifications including CPI among the regressors are those of a higher quality given the data distribution. In particular, a likelihood ratio test confirms that the full specification (Model 3) is the one that best fits our data. Besides this main result, other control variables display a significant effect on MAI impact. Management role positively and significantly impacts on journalists’ perceptions of the impact of MAI. The more positive the journalists’ opinions of a well-organized newsroom, the higher their perceptions of MAI impact. Furthermore, females, as well as journalists with formal training, perceive a stronger impact of MAI than do males and journalists with no professional training or who are only informally trained. Freedom press is significantly associated with MAI impact only in regression 3. Surprisingly, the regression coefficient of freedom press has a positive sign, which is in contrast with our expectations. However, given the high level of correlation between freedom press and CPI, this result could be an “effect” of multicollinearity. On substituting the continuous score of press freedom with the dummy described in the previous section, freedom press turns out to be nonsignificant. The proportion of variance explained by this model specification (Model 3) is 21% at the individual and 80% at the country level.3 Regressions 4, 5, and 6 highlight a positive and significant relationship between CPI and government pressure which confirms H2. Also, freedom press is significantly (and positively, as expected) related to government pressure. However, even on introducing this variable as a control in Regression 6, the coefficient of CPI maintains the same level of significance. Moreover, also in these regressions, if we substitute the continuous score of press freedom with the dummy indicating countries rated as “Free,” the effect of freedom press turns out to be no significant. Not surprisingly, the proportion of variance in journalists’ perceptions of government pressure explained at the country level (79%) is much higher than that explained at the individual level (25%). Finally, the values of AIC and BIC show that regressions that include CPI among the explanatory factors (4 and 6) are better specified according to our data. A likelihood ratio test confirms the best goodness of fit of Model 6, which includes both CPI and freedom press. At the individual level, only gender is significantly related to government pressure. Its regression coefficient displays a positive sign, indicating that females tend to perceive more governmental pressures. By contrast, management role, a variable strongly associated with the other dependent variable, does not display a significant effect. At the country level, though only in Regression 6, (log)GDP per capita displays a positive and significant regression coefficient, which is in contrast with our expectation. However, on plotting the distribution of countries according to the average values of these two variables, it seems that this effect is mostly driven by countries such as Austria, Italy, and Spain, which present good economic performances but, at the same time, high levels of perceived governmental pressures (see Mazzoleni & Splendore, 2014). This conundrum should be the subject of future, more in-depth analyses. To increase the robustness of our results, we carried out a set of statistical checks. First, we built an alternative index measuring journalists’ perceptions of MAI impact, choosing only the most traditional and institutionalized instruments and not considering the online ones. Even when substituting MAI impact with the alternative indicator in the regressions presented above, empirical analyses confirmed the validity of the hypotheses advanced. Second, we tested the effects of alternative indicators of corruption: The World Bank’s Control of Corruption Index (CC) and the Transparency International’s Global Corruption Barometer (GCB). Like CPI also CC is developed by aggregating a large number of surveys on experts, and the two indicators are almost perfectly correlated (r = .98, p < .001). The GCB instead is a large mass survey that, among other aspects, monitors citizens’ experience of corruption. Thus, GCB measures the proportion of respondents who in 2011 reported having paid a bribe in the last 12 months. Whereas CPI and CC are considered subjective indicators because they are developed on the basis of experts’ perceptions, various scholars refer to GCB as an objective indicator of the real spread of corruption in a country because it is developed on the basis of experiences reported by common citizens (Clausen et al., 2011; Treisman, 2007). Both CC and GCB display the same effects on the two dependent variables tested in the empirical analysis, confirming the hypotheses postulated above. This check is particularly important because it demonstrates that our findings are not affected by the method that we used to measure the level of corruption, an issue that is much debated in comparative analysis (Treisman, 2007). Third, even on including a dummy variable that assumed value 1 for nondemocratic regimes, such as Jordan and Tunisia, or on excluding these two countries from the analyses, our main empirical results did not substantially change. However, it is worth noting that while CPI regression coefficients remain statistically significant, freedom press is never significantly associated with either MAI impact or government pressure. Finally, considering that our dependent variables were both left- and right-censored, because in their normalized form they cannot assume values lower than 0 and higher than 1, we reran our models adopting Tobit regressions developed to explicitly address censored dependent variables (Takeshi, 1984). The results obtained from the Tobit regressions were the same as the main results obtained running GLS regressions. Discussion In this study, we have focused on the role played by corruption in affecting journalists’ perceptions of accountability and autonomy in their profession. Our empirical findings have confirmed our research hypotheses, demonstrating that journalists who live and work in countries with a more widespread corruption perceive a lower impact of MAI and more pressures from the government. Journalism is not immune from the detrimental consequences generated by corruption. Corruption, in fact, seems to affect the journalistic doxa, the set of taken-for-granted understandings that guide journalistic practice. The presence of endemic corruption in a system affects journalists’ performances and, consequently, weakens their resolve to produce transparent and accountable journalism. Although the proportion of variance explained by our regression models at the individual level is only around 20%, it is between 70% and 80% at the country level. These findings show how much important are external factors in affecting journalists’ perceptions of accountability and transparency in their profession, in particular, extramedia factors, such as corruption, that do not pertain to journalistic culture, but influence anyway journalistic practices. Nonetheless, there are significant differences in the relationships between corruption and the two dependent variables used in our study. The graph in Figure 1 plots the 14 sample countries according to their CPI score and the country respondents’ average value of MAI impact. The solid line indicates the best linear fit between the two variables according to the data distribution, and shows the negative relationship between corruption and journalists’ perception of MAI impact in producing good-quality journalism. At the same time, the graph highlights the presence of some outliers: in particular Jordan, which displays a particularly high average value of MAI impact. In that country, data were gathered during the so-called Arab Spring, and Jordanian journalists may have reflected that experience, particularly in regard to online MAI (see Fengler et al., 2014). It is not by chance that if we focus on the alternative measure of MAI impact, which takes only the most traditional and institutionalized instruments into account, the association with corruption becomes stronger, with Jordan (and also Spain) displaying lower perceived impact of MAI, in line with countries recording similar corruption scores. Figure 1 View largeDownload slide Linear relationship between perceived corruption (Corruption Perceptions Index, CPI) and journalists’ perception of media accountability instruments (MAI) impact (country average values). Note: Perceived corruption (CPI) on the X-axis. The original 0–10 scale of CPI was inverted, so that higher values of CPI represent higher levels of perceived corruption. Journalists’ perception of MAI impact on the Y-axis. Values of MAI impact are standardized on a 0–1 scale. Higher values of MAI impact refer to journalists’ perception of a stronger impact of MAI on the profession. Figure 1 View largeDownload slide Linear relationship between perceived corruption (Corruption Perceptions Index, CPI) and journalists’ perception of media accountability instruments (MAI) impact (country average values). Note: Perceived corruption (CPI) on the X-axis. The original 0–10 scale of CPI was inverted, so that higher values of CPI represent higher levels of perceived corruption. Journalists’ perception of MAI impact on the Y-axis. Values of MAI impact are standardized on a 0–1 scale. Higher values of MAI impact refer to journalists’ perception of a stronger impact of MAI on the profession. As stressed by the Figure 2, associating the country average CPI scores and Government pressures levels, the relationship between these two variables is now positive and stronger than that between corruption and MAI impact. Journalists who live and work in countries plagued by high levels of corruption tend to perceive higher pressures from governmental authorities. More precisely, though with some distinctions, the graph marks a clear difference between two groups of countries. On the one hand, there are Northern and Western European countries (plus Estonia) characterized by low levels of both corruption and perceived government pressures; on the other, Southern and Eastern European countries and non-European ones with high levels of both corruption and government pressures. Figure 2 View largeDownload slide Linear relationship between perceived corruption (CPI) and journalists’ perception of government pressures (country average values). Note: Perceived corruption (CPI) on the X-axis. The original 0–10 scale of CPI was inverted, so that higher values of CPI represent higher levels of perceived corruption. Journalists’ perception of government pressure (Gov. pressure) on the Y-axis. Values of Gov. pressure are standardized on a 0–1 scale. Higher values of Gov. pressure refer to journalists’ perception of a higher pressure by the government on their job. Figure 2 View largeDownload slide Linear relationship between perceived corruption (CPI) and journalists’ perception of government pressures (country average values). Note: Perceived corruption (CPI) on the X-axis. The original 0–10 scale of CPI was inverted, so that higher values of CPI represent higher levels of perceived corruption. Journalists’ perception of government pressure (Gov. pressure) on the Y-axis. Values of Gov. pressure are standardized on a 0–1 scale. Higher values of Gov. pressure refer to journalists’ perception of a higher pressure by the government on their job. In his seminal work on media accountability, Bertrand (2000, p. 151) affirms that “ethics are not enough,” and that not all journalists are “endowed with a moral sense” (p. 41). His suggestion is therefore that media need accountability instruments to enforce the norms adopted in a given journalistic culture. Our empirical findings demonstrate that corruption jeopardizes journalists’ perceptions of the impact of any instruments that the journalistic field may implement to make journalists accountable. What this article firmly suggests is that considering journalists’ perceptions of their role and their place within the society as a whole (McQuail, 2013) depends also on the society itself, its functioning, and its overall level of corruption. If we consider journalists as a particular category of professionals who provide information about corruption, but who—regardless of the role they cover—are themselves influenced by corruption, this analysis sheds new light on how citizens’ opinions on corruption are linked to those of experts (Pellegata & Memoli, 2016). In environments plagued by corruption, journalists appear to be structurally unable to perform their normative role as watchdogs. Like common citizens, journalists also take corruption for granted, thus reinforcing public opinion perceptions in a “spiral of cynicism.” There are several caveats concerning this study that should be stressed and that should be subject to future research. Although the analysis involved 12 European countries and two Arab states, and although several robustness checks seem to indicate that the empirical results are indifferent to the inclusion or exclusion of the latter, our findings can hardly be generalized to many non-European countries. Media accountability and journalistic autonomy from governmental pressures are mainly conceived in Western and liberal terms. MAI in particular may not have been understood and adopted in the same way across our sample countries. To generalize our conclusions to areas different from Europe, future analyses should cover more non-European countries. Another limitation of this study concerns the potential problem of reciprocal causality between corruption and our dependent variables. Considering that our measure of corruption is developed on the basis of experts’ perceptions, and given the crucial role played by journalists in representing the issue of corruption, reversing our causal argumentation could also be plausible. It may be that experts’ (and even more, citizens’) perceptions of corruption in the public sector are affected by the fact that journalists consider the instruments developed to increase transparency in their work to be ineffective and feel pressures by the government authorities. Whether corruption causes a lower perceived impact of MAI and higher perceived pressures from the government or vice versa is almost impossible to determine with survey data. Further research using more “objective” indicators rather than subjective perceptions is certainly necessary to investigate this relation. Nonetheless, in conclusion we believe that our study offers important contributions to the literature on journalistic accountability and on corruption. First, the article highlights the importance of considering the extramedia dimension when journalists’ perceptions and practices are studied. It is now important to investigate how corruption and diverse journalistic cultures intertwine and are related to each other. Furthermore, our study shows that a category of professionals who play a crucial role in representing corruption is affected by the spread of corruption in the country where they operate. This finding should induce scholars of corruption to reflect on how the relationship between corruption and journalists’ perceptions and representations of it may affect citizens’ opinions of the spread of corruption itself in the public sector. Accordingly, in light of that relation between corruption and journalism also experts’ evaluations, on which the most common measures of perceived corruption are build, should be reassessed. Supplementary Data Supplementary Data are available at IJPOR online. Alessandro Pellegata is a research fellow at the Department of Social and Political Sciences of the University of Milano (Italy). His main research interests are in the field of comparative politics, in particular electoral competition, political representation, public opinion, and corruption. He has published articles in several academic journals, such as Democratization, International Political Science Review, Journal of Information Technology & Politics, and Social Indicators Research. Sergio Splendore is an assistant professor at the Department of Social and Political Sciences of the University of Milano (Italy) His research centers on media sociology with a deep interest in journalism studies and media production. He has authored or coauthored papers in these areas for Journalism, Media, Culture and Society, Journalism Practice, and Oxford Bibliographies in Communication. 1The nine media types considered are daily newspapers, weekly newspapers, magazines, public service radio, private commercial radio, public service television, private commercial television, online news media, and news agencies. Regarding journalistic hierarchy, MediaAct distinguishes between a management level and an operational level. 2Results of all the robustness checks discussed here and at the end of the next section are available on request from the authors. 3The proportion of variance explained at different levels was computed using the formula provided in Snijder and Bosker (1999: p. 101–105). Acknowledgement The authors thank Prof. Susanne Fengler for inspiration and the MediaAct consortium that allowed us to use its data set. We also thank the three referees of the International Journal of Public Opinion Research for their most instructive comments.  This study is based on survey data developed by the project entitled “Media Accountability and Transparency in Europe (MediaAct),” funded by the European Union Seventh Framework Programme (2007-2013. Grant Agreement number 244147. Project Coordinator: Susanne Fengler). References Anand V. , Ashforth B. E., , Joshi M. ( 2004 ). Business as usual: The acceptance and perpetuation of corruption in organizations . Academy of Management Executive ,ssss 18 , sssss – 53 . doi: 10.5465/AME.2004.13837437 Google Scholar CrossRef Search ADS Andersson S., , Heywood P. M. ( 2009 ). The politics of perception: Use and abuse of transparency international's approach to measuring corruption . Political Studies , 57 , 746 – 767 . Doi: 10.1111/j.1467-9248.2008.00758.x. Google Scholar CrossRef Search ADS Anderson C. J., , Tverdova Y. V. ( 2003 ). Corruption, political allegiances and attitudes toward government in contemporary democracies . American Journal of Political Science , 47 ( 1 ), 91 – 109 . doi: 10.1111/1540-5907.00007 Google Scholar CrossRef Search ADS Bertrand C. J. ( 2000 ). Media ethics & accountability systems . New Brunswick, London : Transaction Publishers . Benson R., , Neveu E. ( 2005 ). Introduction: Field theory as a work in progress. In Benson R. , Neveu E. (Eds.), Bourdieu and the journalistic field (pp. 1 – 29 ). Cambridge : Polity Press . Blais A. , Gidengil E., , Kilibarda A. ( 2015 ). Partisanship, information, and perceptions of government corruption . International Journal of Public Opinion Research , 29 , 95 – 110 . doi:10.1093/ijpor/edv037 Bourdieu P. ( 1998 ). Practical reason. On the theory of action . Cambridge : Polity Press . Brunetti A., , Weder B. ( 2003 ). A free press is a bad news for corruption . Journal of Public Economics , 87 , 1081 – 1824 . Google Scholar CrossRef Search ADS Camaj L. ( 2013 ). The media’s role in fighting corruption: Media effects on government accountability . The International Journal of Press/Politics , 18 , 21 – 42 . doi: 10.1177/1940161212462741 Google Scholar CrossRef Search ADS Charron N. ( 2016 ). Do corruption measures have a perception problem? Assessing the relationship between experiences and perceptions of corruption among citizens and experts . European Political Science Review , 8 ( 1 ), 147 – 171 . doi: 10.1017/S1755773914000447 Google Scholar CrossRef Search ADS Clausen B. , Kraay A., , Nyiri Z. ( 2011 ). Corruption and confidence in public institution: Evidence from a global survey . The World Bank Economic Review , 25 , 212 – 249 . doi: 10.1093/wber/lhr018 Google Scholar CrossRef Search ADS Deuze M. ( 2005 ). What is journalism? Professional identity and ideology of journalists reconsidered . Journalism , 6 , 442 – 464 . doi: 10.1177/1464884905056815 Google Scholar CrossRef Search ADS Emerson P. M. ( 2006 ). Corruption, competition and democracy . Journal of Development Economics , 81 ( 1 ), 193 – 212 . Google Scholar CrossRef Search ADS Eberwein T. , Fengler S. , Lauk E. , Leppik-Bork T. (Eds.). ( 2011 ). Mapping media accountability—In Europe and beyond . Berlin : Herbert von Halem Verlag . Eberwein T. , Fengler S. , Philipp S. , Ille M. ( 2014 ). Counting media accountability—The concept and methodology of the mediaact survey. In Fengler S. , Eberwein T. , Mazzoleni G. , Porlezza C. , Russ-Mohl S. (Eds.), Journalists and media accountability. An international study of news people in the digital age (pp. 65 – 79 ). New York, NY : Peter Lang . Fengler S. ( 2012 ). From media self-regulation to ‘crowd-criticism’: Media accountability in the digital age . Central-European journal of communication , 2 , 175 – 188 . Fengler S. , Eberwein T. , Alsius S. , Baisnée O. , Bichler K. et al. ( 2015 ). How effective is media self-regulation? Results from a comparative survey of european journalists . European journal of communication , 30 , 249 – 266 . doi: 10.1177/0267323114561009 Google Scholar CrossRef Search ADS Freille S. , Haque E. M., , Kneller R. ( 2007 ). A contribution to the empirics of press freedom and corruption . European Journal of Political Economy , 23 , 838 – 862 . doi: 10.1016/j.ejpoleco.2007.03.002 Google Scholar CrossRef Search ADS Hallin D. C. , Mancini P. ( 2004 ). Comparing media systems. Three models of media and politics . Cambridge : Cambridge University Press . Hanitzsch T., , Berganza R. ( 2012 ). Explaining journalists’ trust in public institutions across 20 countries: Media freedom, corruption, and ownership matter most . Journal of Communication , 62 , 794 – 814 . doi: 10.1111/j.1460-2466.2012.01663.x Google Scholar CrossRef Search ADS Hanitzsch T., , Mellado C. ( 2011 ). What shapes the news around the world? How journalists in eighteen countries perceive influences on their work . The International Journal of Press/Politics , 16 , 404 – 426 . doi: 10.1177/1940161211407334 Google Scholar CrossRef Search ADS Kaufmann D. , Kraay A., , Mastruzzi M. ( 2006 ). Measuring governance using cross-country perceptions data. In Rose-Ackerman S. (Ed.) International handbook on the economics of corruption (pp. 52 – 85 ). Northampton, MA : Edward Elgan Publishing . Lambsdorff J. G. ( 2006 ). Measuring corruption: The validity and precision of subjective indicators. In Sampford C. , Shacklock A. , Connors C. , Galtung F. (Eds.), Measuring corruption (pp. 81 – 100 ). Burlington, VT : Ashgate . Maas C. J., , Hox J. J. ( 2005 ). Sufficient sample sizes for multilevel modeling . Methodology , 1 , 85 – 91 . doi: 10.1027/1614-1881.1.3.86 Google Scholar CrossRef Search ADS Mauro P. ( 1995 ). Corruption and growth . The quarterly journal of economics , 110 , 681 – 712 . doi: 10.2307/2946696 Google Scholar CrossRef Search ADS Mazzoleni G. , Splendore S. ( 2014 ). Cultures of accountability'™ – or global trends in media accountability? The Hallin and Mancini model revisited. In: S. Fengler , C. Porlezza , T. Erbenwein , G. Mazzoleni , S. Russ-Mohl (eds.), Journalists and Media Accountability. An International Study of News People in the Digital Age ( 145 – 159 ), London : Peter Lang . McQuail D. ( 1992 ). Media performance . London : Sage . McQuail D. ( 2003 ). Media accountability and freedom of publication . Oxford : Oxford University . McQuail D. ( 2005 ). McQuail’s mass communication theory ( 5th ed.). London : Sage . McQuail D. ( 2013 ). Journalism and society . London : Sage . Melgar N. , Rossi M., , Smith T. S. ( 2010 ). The perception of corruption . International Journal of Public Opinion Research , 22 ( 1 ), 120 – 131 . doi:10.1093/ijpor/edp058 Google Scholar CrossRef Search ADS Mishler W., , Rose R. ( 2001 ). What are the origins of political trust? Testing institutional and cultural theories in post-communist societies . Comparative Political Studies , 34 ( 1 ), 30 – 62 . doi: 10.1177/0010414001034001002 Google Scholar CrossRef Search ADS Morlino L. ( 2011 ). Changes for democracy. Actors, structures, processes . Oxford : Oxford University Press . Mungiu-Pippidi A. ( 2006 ). Corruption: Diagnosis and treatment . Journal of Democracy , 17 , 86 – 99 . doi: 10.1353/jod.2006.0050 Google Scholar CrossRef Search ADS Onyebadi U., , Alajmi F. ( 2014 ). Gift solicitation and acceptance in journalism practice: An assessment of kuwaiti journalists perspective . Journalism , 17 , 348 – 365 . doi: 10.1177/1464884914557924 Google Scholar CrossRef Search ADS Örnebring H. ( 2012 ). Clientelism, elites and the media in central and eastern Europe . The International Journal of Press/Politics , 17 , 497 – 515 . doi: 10.1177/1940161212454329 Google Scholar CrossRef Search ADS Pellegata A., , Memoli V. ( 2016 ). Can corruption erode confidence in political institutions among european countries? Comparing the effects of different measures of perceived corruption . Social Indicators Research , 128 ( 1 ), 391 – 412 . doi:10.1007/s11205-015-1036-0 Google Scholar CrossRef Search ADS Persson A. , Rothstein B., , Teorell J. ( 2013 ). Why anticorruption reforms fail—Systemic corruption as a collective action problem . Governance , 26 , 449 – 471 . doi: 10.1111/j.1468-0491.2012.01604.x Google Scholar CrossRef Search ADS Pintak L., , Ginges J. ( 2009 ). Inside the Arab newsroom . Journalism Studies , 10 , 157 – 177 . doi: 10.1080/14616700802337800 Google Scholar CrossRef Search ADS Pintak L., , Nazir S. J. ( 2013 ). Pakistani journalism: At the crossroads of Muslim identity, national priorities and journalistic culture . Media, Culture and Society , 35 , 640 – 665 . doi: 10.1177/016344371348365 Google Scholar CrossRef Search ADS Reich Z., , Hanitzsch T. ( 2013 ). Determinants of journalists' professional autonomy: Individual and national level factors matter more than organizational ones . Mass Communication and Society , 16 ( 1 ), 133 – 156 . doi: 10.1080/15205436.2012.669002 Google Scholar CrossRef Search ADS Relly J. , Zanger M., , Fahmy S. ( 2015 ). Professional role perception among Iraqui Kurdish journalists from a ‘State within a State’ . Journalism , 16 , 1085 – 1106 . doi: 10.1177/1464884914550973 Google Scholar CrossRef Search ADS Rose-Ackerman S. ( 1999 ). Corruption and government. Causes, consequences and reforms . Cambridge : Cambridge University Press . Roudakova N. ( 2008 ). Media-political clientelism: Lessons from anthropology . Media, Culture and Society , 30 ( 1 ), 41 – 59 . doi: 10.1177/0163443707084349 Google Scholar CrossRef Search ADS Shoemaker P. J., , Reese S. D. ( 1996 ). Mediating the message: Theories of influence on mass media content . New York, NY : Longman . Shoemaker P. J., , Reese S. D. ( 2014 ). Mediating the message in the 21st century. A media sociology perspective . Routledge : London . Snijders T. A., , Bosker R. J. ( 1999 ). Multilevel analysis. An introduction to basic and advanced multilevel modeling . London : Sage . Takeshi A. ( 1984 ). Tobit Models: A Survey . Journal of Econometrics , 24 , 3 – 61 . Google Scholar CrossRef Search ADS Treisman D. ( 2007 ). What we have learned about the causes of corruption from ten years of cross-national empirical research? Annual Review of Political Science , 10 , 211 – 244 . doi: 10.1146/annurev.polisci.10.081205.095418 Google Scholar CrossRef Search ADS Willig I. ( 2012 ). News ethnography in a field perspective . Journalism , 14 , 372 – 387 . doi: 10.1177/1464884912442638. Google Scholar CrossRef Search ADS © The Author 2017. Published by Oxford University Press on behalf of The World Association for Public Opinion Research. All rights reserved.

Journal

International Journal of Public Opinion ResearchOxford University Press

Published: May 29, 2017

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off