Media and Corruption: We Need to Talk about Television

Tania Gosselin (Université du Québec à Montréal)

Marina Nistotskaya (University of Gothenburg)

Victor Lapuente (University of Gothenburg)

Paper prepared for presentation at the Annual Conference of the European Consortium

for Political Research in Montreal, Quebec, August 26-29, 2015. Preliminary draft;

please do not quote without authors’ permission.

2

Introduction

A free, independent and diverse media scene has long been regarded as key ingredient of

democracy and good governance. Greater media presence allows for the diffusion of

information, which enables citizens to hold office holders accountable; in turn, politicians, aware

of media scrutiny, are less likely to engage in corrupt behavior (eg. Adsera et al. 2003; Pellegrini

and Gerlagh 2008). Media freedom, also shows a robust relationship with corruption even when

other determinants are taken into account (Brunetti and Weder 2003; Chowdhury 2004).

However, readership figures do not tell us which kind of information circulates, and freedom

indices typically reveal little about the nature of pressures the media are subjected to. As Lindst

and Naurin (2010) note, media systems are connected to politics and economics in ways that

challenge simple models. The “power of information” is, among other factors, shaped by the

sanctioning capacity of citizens. For example, recent studies find that the impact of press

freedom on corruption is conditioned by regime/level of democracy (Kalenborn and Lessmann

2013; Bhattacharyya and Hodler 2015). Yet, the findings with respect to which political

institutions allow citizens to sanction corrupt behavior are contradictory (see review in Chang et

al 2010.) Corruption studies are just beginning to investigate the interplay between media and

institutions such as electoral systems or presidentialism vs parliamentarism (Camaj 2013).

Comparative investigations of linkages between media characteristics - other than

presence and freedom - and corruption are still largely missing1, though some findings point to

the importance of the relationship between government/political parties and the media.

Attempting to disentangle the effects of various forms of media control, Freille et al. (2007) find

that the political dimension of Freedom House’s Media Freedom Index (rather than its legal

control dimension) is what limits media capacity to act as a check on corruption. Public/party

media ownership has been associated with lesser government performance (Djankov et al. 2003;

Besley and Prat 2006; Besley and Burgess 2002). However, the data exploited in these studies

hardly reflect the variations in characteristics of public broadcasters such as Rai Uno in Italy and

1 The gap is nowhere more glaring than in wealthy democracies, where access to free media and citizens’

sanctioning capacity is typically high. The majority of previous comparative studies involved a larger number of

countries; however, OECD countries almost all range in the highest quintile when it comes to measures of media

access or freedom.

3

the BBC in the UK, for example. A more complete picture of the media is needed to further our

understanding of the link between information and corruption.

We follow Freille et al. (2007) in trying to identify the characteristics of media that bear

most on corruption. We use recent data on media systems in 33 European countries. The expert

survey allows to gauge the extent to which outlets are influenced by politics, as well as the

quality of content measured by argument diversity and information accuracy. Our results show

that greater influence of politics and lower information quality of television news content – not

newspapers - is associated with higher levels of corruption; the link is particularly strong in the

case of public television.

The media, politics and information quality

The importance of the ‘political factor’ resonates with the large literature on the impact of

partisanship on voter evaluation of politicians’ performance (measured by the state of the

economy or corruption, see for example Eggers 2014 for a brief review) as well as with the work

of scholars of media systems. The latter focus on links between politics and media in the form of

political parallelism, ownership and regulation as well as journalists’ professionalism (eg. Hallin

and Mancini 2004; Norris 2000; Blumler, McLeod and Rosengren, 1992). However, comparative

research is constrained by the shortage of standardized measures of the influence of politics on

media. As a result, the bulk of comparative research is qualitative and focused on a limited

number of cases. Quantitative works, often concerned with democratic output such as citizens’

information level, largely uses ownership categories such as public, mixed or private media

systems, differentiated by the importance of public broadcasting services.2

Public service broadcasters (PSB) have been found to contribute to information levels

because of the higher quality public affairs programming (Holtz-Bacha and Norris 2001; Aarts

and Semetko 2003; Toka and Popescu 2009). Shehata et al. (2015) find that in Sweden,

inadvertent learning from exposure to public television was particularly relevant for the least

interested viewers. Curran et al. (2009) find that PSB systems offer more hard and international

news, fostering public knowledge, thus contributing to greater news consumption and narrowing

the knowledge gap between the more and less-educated citizens than market driven or mixed

2 Quantitative studies also feature market size but they are largely focused on the US due to detailed data available at

the local market level in that country.

4

media systems. Yet not all public broadcasters offer the same type of content, nor promote

knowledge about politics equally well; it depends on how well PSBs are insulated from

government interference and market forces (Soroka et al. 2013; Hanretty 2010). Looking at the

diversity of sources of stories aired by televisions channels in 11 countries, Tiffen et al. (2014)

do not find meaningful differences in all contexts. Jenssen (2009) finds that exposure to public or

private television makes no difference for political knowledge in Norway.

Private ownership on the other hand is often associated with ‘commercialism’,

sensationalism, and ‘horse race coverage’ of electoral campaigns (Aalberg et al. 2011). Such

characteristics are generally not deemed as conducive to the type of news that citizens need)3.

But again, this expectation is not borne out in a systematic manner. Operationalizing horse race

coverage by the number of references to polls is EU elections coverage in 27 countries, Banducci

and Hanretty (2014) find that such references are more frequent in broadsheets (typically ‘quality

newspapers’) than on commercial television. They also find greater reliance on polls in less

competitive markets and where journalists are more professionalized.4

A number of authors argue that newspapers are more important than television and radio

when it comes to effects on citizens’ information (Delli Carpini and Keeter 1996; Newton 1999;

Price and Zaller 1993), based on the idea that broadsheets offer more or higher quality content.

However, it has been argued that television can more easily attract attention and facilitate

information retention (Graber 2001, Neuman, Just and Crigler 1992, Kwak 1999). At the system

level, Shehata and Stromberg (2011) find that media environments dominated by newspapers are

associated with a greater impact of class on the news consumption gap between those are

interesting in politics and those who are not. Overall, while many micro- and macro-level studies

associate a specific media format(s) with content and effects, there is a need for further

investigation.

Experimental studies (including those exploiting ‘natural’ or quasi-experimental

conditions) shed light on the media-corruption relationship from another angle. Politicians

exposed as corrupt in the media may lose some votes, but are not necessarily thrown out of

3 Horse race type of coverage is based on considerations about the strategy of parties and candidates as well as

opinion polls, as opposed to more substantial coverage of candidates’ views and issue positions. Elenbaas and de

Vreese (2008) finds that it fosters political cynicism in young voters. 4 Another factor that matters in their study is party systems’ polarization (greater polarization is associated with

greater use of polls in the media). ‘Politics as a game’ may well be a characteristics of coverage which attracts

partisan readers (see Iyengar et al. 2004).

5

office. Chang et al.’s (2010) study of Italy after the Second World War demonstrate that media

coverage about corrupt behavior did not lead to the electoral demise of politicians until the Mani

pulite (Clean Hands) investigation received unparalleled media attention in the mid-1990s. Yet

Ferraz and Finan (2008) find that Brasilian mayors’ reelection rates did decrease in the wake of

media exposure, especially where corruption was more severe. In both cases, the effects were

more pronounced where there were press circulation was higher/radio outlets more present.

Studies that directly manipulate media content also show mixed results. Reinnika and

Svensson (2004, 2005) finds that an information campaign in national and local newspapers in

Uganda contributed to curb graft by local officials; the publication of data pertaining to central

grants resulted in schools located closer to a newspaper outlet receiving a greater share of the

grant, and seeing student enrollment and test scores increase. In contrast, Humphreys and

Weinstein (2012) find that Uganda politicians did not respond to country-wide dissemination of

information about their performance; transparency did not impact their chances of reelection in

the subsequent election either. Banerjee et al. (2010) found the same lack of impact for a pre-

election campaign urging villagers in India not to vote for corrupt candidates. Chong et al.

(2015) found only a small negative effect for information dissemination about corrupt mayors in

Mexico; interestingly, when high levels of corruption were reported, challengers’ electoral scores

were even more adversely affected than incumbents’; turnout also took a significant hit.

These results suggests that media effects on corruption are sensitive to a number of

factors, one of them partisanship. In their analysis of the Italian case, Chang and his colleagues

(2010) propose that a threshold in media coverage has to be reached to insure that corruption is

electorally salient5; they also speculate that the information must be credible, which might not be

the case if it is disseminated by partisan media. Individual experimental and survey data also

indicate that partisan attitudes influence how media information is received. Anduiza et al.

(2013) find that partisanship colors citizens’ judgement, making them more tolerant of corrupt

acts committed by members of their preferred parties. Larcinese and Sircar (2014) find the same

moderating role of partisanship in the UK in the wake of the coverage received by a scandal

about MPs’ expenses. In their operationalization of corruption scandals in Spain, Fernandez-

Vazquez et al. (2014) do not consider accusations made by partisan actors, including “partisan-

leaning newspapers[s].”

6

Few macro-level comparative studies of corruption have focused on media ownership.

Observing that state ownership often leads to bias coverage in favor of the government, Djankov

et al. (2003) investigate the link between governance and state/public media ownership in a cross

section of nearly 100 countries. They find that state ownership, especially for newspapers,

translate into more corruption. According to Gentzkow, Glaeser, and Goldin (2004), the

development of a nonpartisan press, allowed by increasing sales and competition, ensured that

newspapers published more and less biased reports about corruption scandals in the United

States in the 20th century.

These suggest that partisan news is bad news when it comes to holding politicians

accountable, in part or mostly because it provides less, or incomplete, or inaccurate information

to citizens. But again this may apply differently based on media systems’ characteristics. In the

highly partisan Spanish media system, Baumgartner and Chaqués-Bonanfont (2015) find that

two newspapers with partisan audience devote significantly greater space to cover corruption

scandals and other faults of the rival party, in an otherwise similar agenda. As noted by Dunaway

et al. (2015, 17), ‘media bias’ has come to designate anything between the “willful distortion of

fact to hard and critical yet ‘earnest’ coverage”. Yet the two are conceptually quite different.

This means that, whenever possible, we should treat political influence and other aspects of

information quality as distinct factors. For example, Dunaway and her colleagues find that

slanted stories offer more information about issues in campaigns news in US local, mainstream

newspapers. That is not to say that political influence and information quality cannot be

connected. As the results below will show, in the real world, they are often related.

Data and hypotheses

We rely on the Control of Corruption Index (CCI) from the World Bank Governance

Indicators (2009) as our dependent variable. It builds on a large number of data sources, which

include expert and citizen perceptions; it thus gauges a wide range of aspects of corruption (see

Bauhr and Grimes 2014; these authors also provide useful remarks on the validity of the CCI).6

6 It is common for measures of governance/corruption to include elements linked to the media. It is also the case of

the CCI, with a survey question about corruption in the media taken from the Transparency International Global

Corruption Barometer Survey). The advantage of an index combining a larger number of factors is to render the

weight of the media element much less significant, rendering it less likely that the measure is directly connected to

7

The independent variables come from the EMSS expert survey conducted in 33 countries in 2009

and 2010 (Popescu, Gosselin and Santana Pereira 2010). The survey queried over 650 experts on

aspects such as media partisanship, information accuracy, diversity of arguments available,

journalist professionalism. Most questions pertained to media systems as a whole; others asked

respondents to evaluate up to ten media outlets offering political coverage on a daily basis in

2009 in each country (3 to 5 TV channels and 3 to 5 newspapers with at least a 5%

audience/readership in each country).

We use the battery of questions pertaining to specific outlets, more precisely the

questions aimed at evaluating partisanship, advocacy, accuracy and diversity of views in 289

individual outlets in the 33 countries included in the survey. The question about media

partisanship read thusly: “How far is the political coverage of each of the following media

outlets influenced by a party or parties to which it is close?” (see Table A1 in the Appendix for

question wording and reliability scores). Following Popescu and Toka (2012) we combine these

four questions into two scales: the Political Commitment scale measures the extent to which

party and ideology colors public affairs coverage; the Information Commitment scale gauges the

quality of information based on the factual accuracy as well as the diversity (balance) of political

views available in each outlet’s coverage (see Table 1 for correlations between the scales). The

Political/Information Commitment labels reflect the previously mentioned idea that if

misinformation can be due to partisanship, complete and accurate information and politically

neutral coverage need not go hand in hand. For example, tabloids may provide a poorly

researched yet ‘neutral’ coverage of politics; similarly, the content of a reference newspaper can

be politically colored, yet provide multiple points of view and be the product of serious fact-

checking process). These two dimensions take into account that partisanship and information

quality are key elements in the causal mechanisms proposed by macro and micro level studies of

media effects on corruption. While cross national data does not allow a direct test of either of

these mechanisms (experimental data is better suited for that purpose), the EMSS offers the

possibility to deepen our understanding of the role of the media environments with respect to

corruption; the media remain an overwhelmingly national affair, structured by language, focus

day-to-day variations in media content. We also used Transparency International’s index as a dependent variable

and obtained similar results to those presented below.

8

on national politics/events, distribution networks, as well as the influence they have on each

other’s coverage).

Unlike Popescu and Toka however, we distinguish between newspapers and television, as

well as between private and public television channels. The media and corruption literature has

almost entirely focused on the press; yet television remains an important, if not the main source

of political news for most citizens. While we make no formal hypothesis pertaining to the links

between specific outlet types and corruption control, the works reviewed in the previous section

showed that we can certainly refine our knowledge by examining both types of information

sources (TV\newspapers, public\private) separately.

The EMSS data also includes system-level evaluations of journalistic professionalism, as

well as external diversity (the extent to which it is possible to find a variety of views in the

media), as well as evaluations of information depth, contextualization, sensationalism,

personalization, and the presentation of politics as a game or a horse race. While these

evaluations do not pertain to specific outlet type or ownership, they provide a useful

complementary check on our interpretations of the results for the Political and Information

Commitment dimensions.

Finally, as controls, we use level of democracy (an average of the Freedom House and

Polity measures of democracy) and wealth (GDP in 2000)7, which have been systematically

linked to corruption in previous macro studies. They are entered in OLS regression equations

with each characteristic of media systems taken separately given the limited number of countries

for which data is available. 8

7 Along with the CCI index, level of democracy and GDP are taken from the Quality of Government dataset (Teorell

et al. 2012). 8 We also conducted analyses with party system polarization and competition (we used Banducci and Hanretty’s

(2014) measure of competition in media market to assess whether having more effective players on the media scene

made a difference). The results were not significant (including the coefficients of various interactions terms). This

was to insure that the results we observe for the media variables are not entirely or mostly attributable to

characteristics of the party system. Obviously, media and politics can be more or less tightly enmeshed, depending

on media systems; the relationship is often anchored in traditions and historical developments (see Hallin and

Mancini 2004). The data we use aimed precisely at attempting to take this relationship into account in a quantitative

manner.

9

Based on the existing literature, we hypothesize that:

1) a stronger information commitment will be associated with better control of corruption;

2) a stronger political commitment will be associated with lower control of corruption.

Results

Table 1 displays the bivariate correlations between the Political and the Information

Commitment scales (Table A2 in the Appendix displays a fuller set of correlations). Like

Popescu and Toka (2012) who proposed the two dimensions, we note that certain characteristics

of media tend to go together. Presenting information for television and newspapers separately

yields useful insight: first, political bias often affects both television and newspapers, thus

coloring whole media systems. Second, more politically biased media television outlets – but

interestingly not newspapers - typically display lower commitment to information quality gauged

by factual accuracy and viewpoints diversity. Such linkages between media aspects (as well as

with those of the wider political environment) characterize the ideal types proposed by Hallin

and Mancini (2004) to compare media systems in the western world.

< Table 1 >

In Table 2, the Baseline model shows that the relationship between level of democracy

and wealth on the one hand and corruption on the other is strong, in line with previous works on

the determinants of corruption. Then, the two dimensions of media performance assessed in this

paper, Political and Information Commitments, are subsequently added to the Baseline model to

compare their contribution to overall variance (Tables 2 and 3). The results show that both

dimensions are important for corruption control. A larger influence of parties and an advocacy

stance is linked to less control of corruption, while a more robust commitment to presenting

accurate information and all sides of an argument is linked to better corruption control. However,

the media outlets that matter in this respect are only television channels, and not newspapers.

This confirms the importance of considering these types of outlets separately when trying to

assess the impact of media contexts. The different outcomes for broadcasting outlets and the

press may be due to a number of reasons. One might be historical traditions; after all, newspapers

10

started out as partisan affairs and in many media systems, have kept a partisan/ideological

coloring. There are also typically more newspapers than television channels in a media system;

the audience of the former is likely to be more segmented than the latter’s, conferring greater

weight to television news content. Public television, which in some countries commands a very

significant audience, has yet a distinct status, imbued (at least in theory) with a civic orientation.

Because of this status of public broadcasting services, we also present the analysis for

public and private TV channels separately in the right side columns of Tables 2 and 3. Political

Commitment of public TV channels adds the most to the level of explained variance in our

measure of corruption; private channels also matter, albeit less so. Where information quality is

concerned, it is public television outlets alone that drive the positive association between TV

Information Commitment and control of corruption; the coefficient for private television outlets

is not significant. These results highlight the key relationship between the performance of public

broadcasting and holding corruption in check.

< Tables 2 and 3 >

Other questions featured in the EMSS allow us to further check on the robustness of the

association found between the political and informational aspects of media performance. The

measures are available at the system level only (and not for each outlet like the variables behind

Political and Information Commitments).9 When journalists’ political orientations are public

knowledge, control of corruption is significantly lower. It is also the case when “media outlets in

general succeed in serving as watchdogs scrutinizing the actions of government official on behalf

of citizens”. External diversity (the presence of all major political opinions) in broadcasting

outlets is significantly associated with better control of corruption control; however, external

diversity in newspapers is not.

Still at the system level, whether tabloid and quality newspapers cover the news

differently or not did not display a significant coefficient; nor did the question gauging the

reliance of the media in general on a more or less wide range of specialists to present information

and analyses, nor questions about the depth/richness of media information. The EMSS includes a

9 The results pertaining to these complementary, system level media characteristics are not shown but are available

from the authors.

11

Commercialisation Index gauging personalization, sensationalism and gamification of the

depiction of politics in national media systems. Taken separately, these three variables do not

pass the significance threshold. However, the index displays a significant negative coefficient,

albeit only at the .1 level.

Potentially anchored as much in the political and the information dimensions, journalists’

professionalism (gauged with three questions about journalists’ ethical motivations, excellence

criteria and training quality) and overall media credibility in a country make the strongest

contributions to overall explained variance in corruption control than any other individual media

characteristics considered in this analysis (always in addition to the Baseline model). Overall, the

information gleaned at the system level supports the results shown in Tables 2 and 3:

politicization matters most (in a bad way), and information quality too (in a good way), but to a

lesser extent. And it is television that stands behind the relationships, not newspapers.

Conclusion

In this paper, we investigate in further details the aspects of media that might be at work

behind the oft noted relationship between media and corruption in macro studies. The expert

survey data we used provides comparable measures of characteristics of 289 media outlets in 33

countries in Europe. In this preliminary analysis, we relied on two dimensions of media

performance for which up to ten outlets were individually evaluated in each country. The

analysis shows that politicized news content spells bad news for corruption control, while

information quality is associated with less corruption. However, this applies only to television

(and especially public television), and not to newspapers.

It is hard to tell whether it is because of political commitment per se in television that

corruption is less under control, or because political commitment is often the companion of less

accurate and less diverse information in the 33 countries involved in the analysis. Despite the

fact that this cross national study cannot reveal much about causality, it does suggest citizens

need both credible (not overly partisan) and quality information to hold politicians accountable.

The results are compatible with the criticism addressed to the principal-agent mechanism; the

idea that once the media tells about corruption citizens will activate sanctioning mechanisms is

too simple. This mechanism supposes that citizens consider all corruption information in the

12

same manner, no matter the media source or their own beliefs; we know from survey and

experimental studies that it is not the case.

Our next step is to bring together a quantitative analysis and a more detailed qualitative

examination of how media characteristics – partisanship and information quality as well as other

aspects covered by the EMSS - combine in patterns that may be linked to corruption in different

settings.

13

Table 1. Bivariate correlations – Political and Information Commitments

Political

commitment

TV

Political

Commitment

Newspapers

Information

Commitment

TV

Information

commitment

Newspapers

Political Commitment

TV .466*** -.622*** .045

Political Commitment

Newspapers -.119 -.254

Information Commitment

TV .106

Information Commitment

Newspapers

14

Table 2. Political Commitment and Control of Corruption (OLS)

Baseline

(controls only)

Political

Commitment

TV

Political

Commitment

Newspapers

Political

Commitment

Public TV

Political

Commitment

Private TV

beta Sig beta Sig beta Sig beta Sig beta Sig

Constant

.000

.024

.004 .028 .006

Democracy

(FH/Polity

2009)

0.314 .002 .326 .001 .312 .002 .329 .001 .342 .001

Real GDP

(2000) 0.694 .000 .496 .000 .679 .000 .457 .000 .542 .000

Variable of

interest

-.267 .002 -.086 .291 -.304 .001 -.306 .014

Adjusted Rsq

.804

.860

.805

.869

.843

N 33

15

Table 3. Information Commitment and Control of Corruption (OLS)

Information

Commitment

TV

Information

Commitment

Newspapers

Information

Commitment

Public TV

Information

Commitment

Private TV

beta sig beta sig beta sig beta sig

Constant

.000 .001 .000 .000

Democracy

(FH/Polity

2009)

.374 .000 .323 .002 .319 .004 305 .003

Real GDP

(2000) .526 .000 .000 .538 .000 714 .000

Variable of

interest .226 .006 .037 .669 .200 .032 076 .357

Adjusted Rsq .852 .798 .835 .803

N 33

16

References

Aalberg, Toril, Jesper Strömbäck, and Claes H . de Vreese. 2011. The framing of politics as

strategy and game: A review of concepts, operationalizations and key findings. Journalism 13

(2) : 1–17.

Aarts, Kees, and Holli A. Semetko. 2003. The Divided Electorate: Media Use and Political

Involvement. Journal of Politics 65 (3): 759-784.

Adsera, Alicia, Carles Boix, and Mark Payne. 2003. Are You Being Served? Political

Accountability and Quality of Government. Journal of Law, Economics and Organizations 19

(2): 445-490.;

Anduiza, Eva, Aina Gallego, and Jordi Muñoz. 2013. Turning a Blind Eye Experimental

Evidence of Partisan Bias in Attitudes Toward Corruption." Comparative Political Studies 46:

1664-1692.

Susan Banducci and Chris Hanretty (2014). Comparative determinants of horse-race coverage.

European Political Science Review, 6, pp 621-640

Banerjee, Abhijit, Selvan Kumar, Rohini Pande and Feliz Su. 2010. Do Informed Voters Make

Better Choices? Experimental Evidence from Urban India. Unpublished manuscript

(http://www.povertyactionlab.org/node/2764).

Bauhr, Monika, and Marcia Grimes. 2014. Indignation or Resignation: The Implications of

Transparency for Societal Accountability. Governance 27 (2) :291–320)

Baumgartner, Frank R., and Laura Chaqués Bonafont. 2015. All News is Bad News: Newspaper

Coverage of Political Parties in Spain, Political Communication 32 (2) : 268.

Besley, Timothy, and Andrea Prat 2006. Handcuffs for the Grabbing Hand? Media Capture and

Government Accountability. American Economic Review 96 (3) : 720-736.

Besley, Timothy, and Robin Burgess. 2002. The Political Economy of Government

Responsiveness: Theory and Evidence from India. Quarterly Journal of Economics 117 (40):

1415-1452.

Bhattacharyya, Samit, and Rolland Hodler. 2015. Media freedom and democracy in the fight

against corruption. European Journal of Political Economy 29: 13-34.

Blumler, McLeod, & Rosengren, 1992

Brunetti, A., and Weder, B. 2003. A free press is bad news for corruption. Journal of Public

Economics 87 (7–8) : 1801–1824.

Camaj, Lindita. 2013. The Media’s Role in Fighting Corruption: Media Effects on Governmental

Accountability. International Journal of Press/Politics 18 (1) : 21–42.

Chang, Eric C., Miriam A. Golden and Seth Hill. 2010. Legislative Malfeasance and Political

Accountability. World Politics 62 (2): 177-220.

Chong, Alberto, Ana L. De La O, Dean Karlan and Leonard Wantchekon. 2015. Does

Corruption Information Inspire the Fight or Quash the Hope? A Field Experiment in Mexico on

Voter Turnout, Choice, and Party Identification. Journal of Politics 77 (1) : 55.

17

Chowdhury, S. K. (2004). The effect of democracy and press freedom on corruption: An

empirical test. Economics Letters 85 (1) : 93–101.

Curran, James, Shanto Iyengar, Anker Brink Lund and Inka Salovaara-Moring. 2009. Media

System, Public Knowledge and Democracy: A Comparative Study. European Journal of

Communication 24: 5-26.

Delli Carpini, Micheal X., and Scott Keeter. 1996. What Americans Know About Politics and

Why it Matters. New Haven: Yale University Press.

Djankov, Simeon, Carolee McLeish, Tatiana Nenova and Andrei Shleifer. 2003. Who Owns the

Media? Journal of Law and Economics 46 (2): 341-382

Johanna L. Dunaway, Johanna Ll, , Nicholas T. Davis, Jeremy Padgett and Rosanne M. Scholl.

2015. Objectivity and Information Bias in Campaign News. Journal of Communication.

doi: 10.1111/jcom.12172

Eggers, Andrew C. 2014. Partisanship and electoral accountability: evidence from the UK

expenses scandal. Quarterly Journal of Political Science 9 (4): 441-472.

Elenbaas and de Vreese 2008

Fernandez-Vazquez, Pablo, Pablo Barberá and Gonzalo Rivero. 2014. Rooting Out Corruption or

Rooting for Corruption? The Heterogeneous Electoral Consequences of Scandals. Forthcoming

in Political Science Research and Methods. (available at

http://pablobarbera.com/static/rooting_out_corruption.pdf).

Ferraz, Claudio, and Frederico Finan. 2008. Exposing Corrupt Politicians: The Effects Of

Brazils Publicly-released Audits on Electoral Outcomes. Quarterly Journal of Economics 123

(3): 703–745.

Freille, S., Haque, M. E., and Kneller, R. (2007). A contribution to the empirics of press freedom

and corruption. European Journal of Political Economy 23 : 838–862.

Hallin, Daniel C., and Paolo Mancini. 2004. Comparing Media Systems. Three Models of Media

and Politics. Cambridge: Cambridge University Press.

Hanretty, Chris. 2010. Explaining the de facto independence of public broadcasters. British

Journal of Political Science 40 (1): 75-89.

Holtz-Bacha, Christina, and Pippa Norris. 2001. To Entertain, Inform, and Educate.Still the Role

of Public Television. Political Communication 18: 125-140.

Humphreys, Macartan, and Jeremy Weinstein. 2012. Policing Politicians: Citizen Empowerment

and Political Accountability in Uganda. Preliminary Analysis. ICG working paper (available at

http://www.theigc.org/wp content/uploads/2015/02/Humphreys-Weinstein-2012-Working-

Paper.pdf).

Iyengar, Shanto, James Curran, Anker Brink Lund, Inka Salovaara-Moring, Kyu S. Hahn and

Sharon Coen. 2010. Cross-National versus Individual-Level Differences in Political Information:

A Media Systems Perspective. Journal of Elections, Public Opinion and Parties 20 : 291–309.

Iyengar, Shanto, Helmut Norpoth, and Kyu S. Kahn. 2004. Consumer Demand for Election

News: The Horserace Sells. Journal of Politics 66 (1): 157-175.

18

Jenssen, Anders Todal. 2009. Does Public Broadcasting Make a Difference? Political

Knowledge and Electoral Campaigns on Television. Scandinavian Political Studies 32 : 247–71.

Kalenborn, Christine, and Christian. Lessmann. 2013. The Impact of Democracy and Press

Freedom on Corruption: Conditionality Matters. Journal of Policy Modeling 35 : 857–886

Kwak, Nojin. 1999. Revisiting the Knowledge Gap Hypothesis: Education, Motivation, and

Media Use. Communication Research 26: 385-413.

Larcinese, Valentino, and Indraneel Sircar. 2014. Crime and Punishment the British Way:

Accountability Channels Following the MPs’Expenses Scandal. Working Paper n. 517, (June).

Innocenzo Gasparini Institute for Economic Research, Bocconi University (available at

http://www.igier.unibocconi.it/files/517.pdf).

Lindstedt, Catharina, and Daniel Naurin. 2010. Transparency Is Not Enough: Making

Transparency Effective in Reducing Corruption. International Political Science Review 31: 301–

322.

Neuman, W. Russell, Marion R. Just and Ann N. Criggler. 1992. Common Knowledge: News

and the Construction of Political Meaning. Chicago: University of Chicago Press.

Norris, Pippa. 2000. A Virtuous Circle: Political Communication in Post-Industrial Societies.

New York: Cambridge University Press.

Pellegrini, L., and Gerlagh, R. 2008. Causes of corruption: A survey of cross-country analyses

and extended results. Economics of Governance 9 (3). 245-263

Popescu, Marina, Tania Gosselin and Jose Santana Pereira. 2010. European Media Systems

Survey 2010. Data set. Colchester, UK: Department of Government, University of Essex.

URL: www.mediasystemsineurope.org.

Popescu, Marina, and Gabor Toka. 2012. How can mass media help citizens make sense of the

political world? Media systems and citizens’ cognitive political engagement in Europe . Paper at

the workshop Advancing comparative political communication research: new frameworks,

designs and data, ECPR Joint Sessions of Workshops, Antwerp, April.

Price, Vincent, and John Zaller.1993. Who Gets the News? Alternative Measures of News

Reception and Their Implications for Research. Public Opinion Quarterly 57: 133-164.

Prior, Markus. 2013. Media and Political Polarization. Annual Review of Political Science 16 :

101-127.

Reinikka , RItva, and Jakob Svensson . 2005. Fighting Corruption to Improve Schooling:

Evidence from a Newspaper Campaign in Uganda. Journal of the European Economic

Association 2/3 : 259-267. Papers and Proceedings of the Nineteenth Annual Congress of the

European Economic Association.

Soroka, Soroka, Blake Andrew, Toril Aalberg, Shanto Iyengar, James Curran, Sharon Coen,

Kaori Hayashi, Paul Jones, Gianpetro Mazzoleni, June Woong Rhee, David Rowe and Rod

Tiffen. 2013. Auntie Knows Best? Public Broadcasters and Current Affairs Knowledge. British

Journal of Political Science 43 : 719-739.

Shehata ,Shehata, David Nicolas Hopmann , Lars Nord , and Jonas Höijer. 2015.

Political Communication 32 (3) :

19

Tiffen, Rodney, Paul K. Jones, David Rowe, Toril Aalberg, Sharon Coen , James Curran, Kaori

Hayashi, Shanto Iyengar, Gianpietro Mazzoleni, Stylianos Papathanassopoulos, Hernando Rojas,

and Stuart Soroka. 2014. Sources in the News. Journalism Studies 15 (4) : 374-391.

Teorell, Jan, Marcus Samanni, Sören Holmberg and Bo Rothstein. 2012. The Quality of

Government Basic Dataset made from The QoG Standard Dataset version 6Apr11. University of

Gothenburg: The Quality of Government Institute, http://www.qog.pol.gu.se.

Toka, Gabor, and Marina Popescu. 2009. Public Television, Private Television and Citizens’

Political Knowledge. EUI Working paper series, RSCAS 2009/66. Florence: European

University Institute, Robert Schuman Centre for Advanced Studies.

20

Appendix

Table A1. Media Characteristics – question wording and reliability scores (EMSS 2009-10)

Variable name Question wording Item

reliability

Partisanship How far is the political coverage of each of the following

media outlets influenced by a party or parties to which it is

close?

(0=not at all; 10=strongly)

0.92

Advocacy To what extent does each advocate particular views and

policies?

(0=never; 10=always)

0.90

Accuracy To what extent do these media provide accurate information

on facts backed by credible sources and expertise?

(0=never; 10=always)

0.94

Diversity To what extent does each present equally well the arguments

of all sides in political debates?

(0=never; 10=always)

0.91

Political Commitment: the weighted national mean of the sum of the Partisanship and Advocacy

variables, themselves the mean expert rating of (all TV, public TV, private TV and newspapers)

outlets on the questions in Table A1 above. The weight of each outlet was proportional to its

2009 audience size reported in European Audiovisual Observatory (2010) and, for newspapers,

circulation as reported in Zenith (2010). In the EMSS data, the corresponding variables are

‘partall’, ‘partpbtv’, ‘partprtv’ and ‘partnews’.

Information Commitment: the weighted national mean of the sum of the Accuracy and Diversity

variables, themselves the mean expert rating of (all TV, public TV, private TV and newspapers)

outlets on the questions in Table A1 above. Weighting was similar to Political Commitment. In

the EMSS data, the corresponding variables are ‘qualall’, ‘qualpbtv’, ‘qualprtv’ and ‘qualnews’.

21

Table A2. Bivariate correlations - Media Characteristics

Note: Only significant coefficients appear in the table.

Partisans

hip TV

Partisans

hip

Newspap

ers

Advocac

y

TV

Advocac

y

Newspap

ers

Argumen

t

Diversity

TV

Argumen

t

Diversity

Newspap

ers

Accuracy

TV

Accuracy

Newspap

ers

Partisanship

TV

Partisanship

Newspapers

.636

Advocacy

TV

.825 .460

Advocacy

Newspapers

.666

Argument

Diversity TV

-.599 -.641

Argument

Diversity

Newspapers

-.348 -.294

Accuracy

TV

-.459 -.521 .734

Accuracy

Newspapers

.876