Parties,Ideology,and News Media in Central-Eastern and Western Europe

Abstract

This study assesses the degree and direction of media bias towards political parties in Central-Eastern and Western European democracies. Previous research has argued that despite policy efforts to detach the media from the political domain, journalism in former Eastern Bloc countries is still characterized by a more partisan style than in Western Europe. Our analysis employs data from the European Parliament Election Study 2009 (EES) and the European Media Systems Survey 2010 (EMSS), covering 187 parties and more than 120 media outlets in fifteen Western and ten Central and Eastern media systems across the European Union. To analyze partisan media bias, we look at (1) how well audience patterns correlate with partisan preferences of media users and (2) the extent to which media outlets favour specific parties according to experts. Contrary to our hypotheses, the results show that levels of media bias in Central and Eastern Europe are similar to those in Western Europe. We also find that left–right party ideology predicts media bias in the latter, but not in the former. Our findings question the general assumption that partisan media bias is higher in “the East” and challenge the widespread belief that a long tradition of media commercialization, as found in “the West,” leads to less political media bias.

Keywords

partisan media bias Central and Eastern Europe comparative research political communication survey

From the beginning of the twentieth century, objectivity has become a normative principle for journalists across the Western world.¹ The idea of providing neutral, fair and balanced information has come to constitute the core of mainstream journalism culture,² as opposed to the political advocacy role of earlier centuries.³

In Central and Eastern Europe (CEE), the process towards less partisan and more professional news coverage was limited during the Cold War. It was not until the twilight of the communist regimes that political journalism could start leaving behind decades of state control and strong party press. Accordingly, it appears natural to wonder if journalists in former Eastern Bloc media systems still retain today part of their former partisan role or if they have moved towards a less extreme partisan style as is claimed to be common in Western Europe (WE).⁴

Until now, media audiences or news coverage in “the East” and “the West” have rarely been compared,⁵ and there has been little research analyzing partisan media bias in post-communist countries. Studying partisan media bias is, however, highly relevant in light of its societal consequences. Previous studies have shown that, under some conditions, it reinforces political choices, informs citizens and detaches citizens from politics, and polarizes the electorate.⁶ Previous research also argues that the tone of political news plays a critical role in CEE democracies, where citizens have not had the time to develop strong partisan loyalties and affections towards the political system.⁷ This study presents a systematic comparison of partisan media bias in ten CEE countries and fourteen Western European systems.⁸ By analyzing how media usage patterns are related to the political preferences of audiences, and how experts evaluate the partisanship of media outlets, we contribute to the research on media transformation and media pluralism in CEE and WE.

Parties and Democracy: A Cornerstone Called Media Autonomy

The transformation of mass media, together with the reform of political institutions, has been considered a key step for post-communist societies to consolidate their democracies.⁹ In particular, it has been argued that the extent to which the media left behind their previous advocacy reporting style and approached higher levels of objectivity and neutrality is an indicator of the degree of democratization in former communist countries, as well as the level of convergence with Western societies.¹⁰

The Objectivity Principle in “the West”

The ideal of a clear separation between the media and the political sphere has flourished over many decades in “the West.”¹¹ In the United States, profit-oriented commercialism and technological innovations paved the way for the consolidation of objective journalism. Embracing objectivity was a way for media owners to avoid criticism and the loss of wide audiences. Also, it was the journalists’ efforts to self-govern and legitimize their profession that led to the articulation of objectivity as a cardinal professional principle.¹² The promotion of editors’ associations, the defensive attitude against any intrusion on the news media agenda by political actors and their public relations aides, and the will to impartially inform their audiences are examples of such efforts. Following Chalaby, the “secularization” of the media and the penetration of an objectivity-oriented journalism in the, by then, more coloured and partisan (West) European media landscape was a result of commercialization and globalization trends, which brought the media closer to an “Anglo-American way” of approaching news-making.¹³

But how does this ideal of objectivity apply to today’s news coverage of public affairs? Journalists ordinarily engage in objective reporting by covering news without any specific partisan view.¹⁴ Following Semetko, objectivity further implies focusing on news values such as the relevance of political actors and major parties.¹⁵ By so doing, Western media may be promoting a kind of political, though non-partisan, bias. For example, previous research has shown that incumbent politicians are treated as more newsworthy than oppositional actors¹⁶ and portrayed in more favourable ways.¹⁷ More powerful and representative political groups and parties are also shown to be more often in the news.¹⁸

Yet there are also signs of partisan bias in some WE media systems.¹⁰⁷ Media outlets in Italy, Spain, or France have been depicted as more or less openly partisan,¹⁹ while early studies on the United Kingdom and Sweden showed strong ties between the political preferences of readers and of the newspapers they read.²⁰ Goldman and Mutz showed that media in the United Kingdom, Greece, and Italy hold similar high levels of media–party parallelism²¹ to Bulgarian mainstream media outlets.²² Austria, Spain, Germany, and Greece have also served as examples for the discussion of political imbalance regarding public service broadcasting.²³ In fact, post-communist media systems have been described as resembling the so-called polarized pluralist media systems,²⁴ most specifically the Italian case,²⁵ where public service broadcasting had, until very recently, institutionalized links to political parties.²⁶ Hallin and Mancini have also suggested that CEE media systems may be closer to the younger democracies of Greece, Spain, and Portugal than to Italy, with its more perennial democracy, parties with broader social roots, and a longer tradition of consensual politics.²⁷

Even among those Western countries with a tradition of advocacy journalism, such as Germany or the Mediterranean media systems, however, partisanship is generally claimed to have declined as a result of an extended commercialization of their media markets.²⁸ The profit-oriented character of news organizations and their aim of responding to the demands of broad audiences across political trends may have contributed to the decline of partisan media.²⁹ Western public broadcasters have been detached from direct government interference for a longer time. An earlier European integration and subsequent legislative harmonization also played its part.³⁰ This legislation called for stricter regulations on advertising and (to a lesser extent) on ownership concentration, which favoured the development of news media providing more diverse and unbiased information.

Partisan Media Bias in “the East”

As in Western Europe, East-Central European countries have also developed policy and regulatory frameworks to promote media independence from political control and avoid political imbalances in news media coverage. However, the will and ability of political parties to influence the media, in addition to the advocacy role traditionally attributed to journalists, casts doubts on the extent to which political balance is promoted by news outlets in CEE. During the 1990s Public Service Broadcasters were created in most CEE countries with the ideal of serving the public interest. In a very short period of time, telecommunications and broadcasting markets were also liberalized. These efforts were coupled with substantial increases in foreign investment—especially from Western media conglomerates—and recent legislative harmonization and compliance with EU policy.³¹ CEE media systems moved towards commercialization and digitalization, began to produce a more diverse and internationalized content, and also became less partisan than in the first “political breakthrough years” after the Iron Curtain fell.³² However, the enforcement of this legislation proved to be less strict in new EU member states than in Western Europe.³³

Some authors have pointed at specific political “disabling environments,” which have affected the development of free and independent media markets in post-communist democracies.³⁴ In many cases, processes toward a new political and economic order after the fall of communism were defined by the interests of elites from the old regime together with old oppositionists and foreign investors. These “negotiated revolutions”³⁵ eventually narrowed the scope of institutional transformations as expressions of popular will. The unfolding of Public Service Broadcasting (PSB) was no exception. Communist media systems were state-owned monopolies with party “nomenklatura” as media executives, propaganda content, and systematic censorship.³⁶ After the fall of communism, the newly established public broadcasters still played a central role within these media systems and maintained almost all their employees.³⁷ The conflict between the first heads of Hungarian Radio and Television and the nationalist government, which culminated in the resignation of the two directors, is probably one of the best-known examples of the strong political control of public media during the 1990s. Throughout that decade, countries such as Slovakia, the Czech Republic, and even successive Polish governments all tried to subordinate PSB to the administration in office.³⁸

Nowadays, public service broadcasters in CEE countries are still strongly controlled by governments, and new regimes often start with a “purge” of the former administration.³⁹ The media is a means to gain and sustain political control. Weak party machines competing to gain seats in Parliament are highly dependent on the support of electorates whose political affinities are volatile and can be strongly influenced by the media.⁴⁰ Against this background, it is no surprise that the existence of tight links between current main media groups, local investors, and political parties is a feature common to all post-socialist countries in the region.⁴¹ Sparks noted that—after a period of regime changes and relatively anarchic freedom of journalists—“in country after country, there is very strong evidence, particularly with regard to the mass media, of the close interrelationship between political and economic power.”⁴² The media are used by political elites to win voters’ support and defeat political rivals in Bulgaria, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Slovakia, and Slovenia.⁴³ Furthermore, as Roudakova pointed out, in some countries clientelistic links can be clearly identified, but they are “transitory and non-binding”⁴⁴ and they change depending on financial support, among other factors.

The poor organization of journalists and their vulnerability to political instrumentalization also renders it difficult to improve journalistic professionalization and autonomy in the region.⁴⁵ Raycheva and Dimitrova noted that journalists in Bulgaria are punished for criticizing powerful political candidates.⁴⁶ Other authors suggest that a “fact-centred” style of journalism never fully replaced advocacy journalism in countries such as Poland,⁴⁷ Hungary, Bulgaria,⁴⁸ or Romania, where taking a political stance is not considered to be unprofessional and journalists see themselves as providers of opinion and contributors to democracy.⁴⁹

As with Western European countries, however, there are of course differences between countries in CEE. For example, whereas levels of media freedom in Baltic countries are comparable to those in more long-standing democracies in “the West,” and political parallelism in the press is said to have decreased,⁵⁰ press–party parallelism scores high in Romania or Bulgaria.⁵¹ Romanian and Bulgarian media work under considerable political pressure,⁵² and public service broadcasters are still tightly controlled by the government.⁵³

Nevertheless, as Gross puts it, “generalizations regarding media evolution in the region are possible.”⁵⁴ There have been recent episodes of conflict involving political actors and media instrumentalization even in very successful examples of democratization such as Poland, Hungary, or the Czech Republic.⁵⁵ The weakness of parties, electoral volatility, and the overall vulnerability of such political systems⁵⁶ have placed them, together with other countries in the region, within what Carothers called a political “gray zone.”⁵⁷ Under these circumstances, it seems unlikely that the extent of political control and partisanship in news media coverage in “the East” has decreased to the point of convergence with those found in Western media systems, where modern day journalistic practices and norms value objectivity, political detachment and neutrality in news reporting.

Leftist Parties in “the East”

If we are right to assume that partisan biases in news content are more common in CEE countries than in Western-European countries, the question arises of which political forces—left-of-centre or right-of-centre—are most favoured by the media. Historically, CEE leftist parties have been more successful in cooperating with different political groups throughout the post-communist period than their right-wing counterparts.⁵⁸ Old and new left-wing parties have joined forces with trade unions, business elites, and reformers from former communist parties, resulting in a lower fragmentation of the left side.

Communist successor parties in countries such as Poland or Hungary were particularly successful in the construction of cross-class alliances. As Orenstein argued, in these parties (1) reformers took control and tried to model their new parties by mimicking Western social-democratic parties and (2) strategic electoral blocks were built with trade unions.⁵⁹ The return of the left to power in the mid-nineties in Lithuania, Hungary, Bulgaria, or Romania was not only due to disenchantment with the austerity economic policies promoted by previous right-wing governments but also to the will of “the left” to re-integrate working-class voters together with former “nomenklatura” elites and new economic groups.⁶⁰

In Hungary, Lithuania, and Romania, the rise of left-wing parties in the 1990s was coupled with the weakness of their competitors.⁶¹ Throughout the decade, CEE left-wing parties were not only less fragmented but also counted on a more loyal electoral base than right-wing parties. Tavits and Letki’s study further showed that the “leftist parties have been, on average, electorally stronger than other parties throughout the post-Communist period” in CEE.⁶²

A number of studies—most of them case studies—have analyzed the relationship between the media’s favourability towards leftist parties and their political success. For example, Bulgarian newspapers covered claims by leftist parties supporting government spending more often than pledges of government cuts by the right.⁶³ Czech leftist press outlets had a great influence in the widespread institutional distrust among citizens by emphasizing the lack of transparency in the processes of privatization in the country,⁶⁴ while Polish national broadcasting and the members of the broadcasting regulatory body had leftist preferences all throughout Kwaśniewski’s presidency.⁶⁵

In short, there are several factors supporting the assumption that news coverage is more favourable toward left-of-centre parties. The institutional relevance and ability of Eastern leftist parties to build cross-sectional coalitions in recent decades, their membership base and local presence, their professional management and skilled PR specialists,⁶⁶ and their overall enduring electoral success should be counted among them.

To sum up, the discussion so far leads us to formulate two hypotheses:

Hypothesis 1: Partisan media bias is higher in Central and Eastern Europe than in Western Europe.

Hypothesis 2: Media bias in CEE is more favourable to left-of-centre parties than to right-of-centre parties.

Data and Method

To investigate the hypotheses this study builds on data from 187 parties and more than 120 media outlets in 15 EU WE and 10 EU CEE media systems (see Note 8). The data was retrieved from the European Parliament Election Study 2009 (EES) voter survey and the European Media Systems Survey 2010 (EMSS). The EES voter survey contains information on the political attitudes and behavior as well as media usage patterns of 27,079 individuals from 27 EU member states, collected during the three weeks following the June 2009 European Parliamentary Elections. The EMSS gathers 838 experts’ assessments of political balance and diversity of viewpoints in European mainstream media for 34 EU media systems, collected from December 2009 to May 2010.

This data is firstly used to measure partisan media bias and secondly to conduct analyses of the extent and direction of partisan media bias—that is, whether media favours leftist or rightist parties—in both WE and CEE regions.

Measurement of Partisan Media Bias

By relying on previous operationalizations of media favourability towards parties, we measure media bias towards political parties by means of two indicators. We look at (1) how well the partisan preferences of media users correlate with their reading and viewing patterns and (2) the extent to which media outlets favour specific parties, according to experts. We call the respective indicators Partisan Media Use (PMU) and Media–Party Political Agreement (MPPA).

The measurement of the extent of partisan media bias through PMU is constructed upon data from the voter survey of the EES, whereas MPPA relies on data from the EMSS. First, the EES is used to build our initial measure of media bias (PMU) for 187 parties. We do this by combining a person’s partisan preferences with information on how often the same person used three leading newspapers and two major TV news programmes in each of the twenty-four EU member states selected.⁶⁷ Second, we use EMSS expert assessments to construct MPPA for each of the aforementioned parties using reported evaluations on the political colour of each media outlet.⁶⁸

To construct the indicator Partisan Media Use (PMU), we consider the following linear regression model for each party p:

Y_{i, p} = β_{0} + \sum_{l} (β_{l} \cdot X_{i, l})

$Y_{i, p}$ is the reported probability that citizen i votes for party p (11-point scale) and $X_{i, l}$ is the frequency of use of news media outlet l by citizen i (8-point scale). The regression includes the frequency of use for two TV channels and three newspapers of the country of the given party as independent variables. The more the media use patterns predict the probability of voting for a given party, the more this party is assumed to be favourably treated by the media. For each party, the resulting adjusted R² score accounts for the amount of variance of individual party preference explained by the frequency of each news media outlet’s use and is interpreted as a measure of media bias towards each given party (see the appendix for a detailed description of the procedure applied to German parties). Previous approaches by Seymour-Ure, Hallin and Mancini, or van Kempen have pointed at the partisanship of newspaper readers as a way to operationalize the association between political parties and newspapers—press–party parallelism—or wider political trends and media outlets in a given country—political parallelism.⁶⁹ Our methodological perspective is inspired by van Kempen.⁷⁰ Nevertheless, our approach differs from hers as we construct a measure at a party level rather than at a country level.

Our second indicator, Media–Party Political Agreement (MPPA), is constructed for each party p as follows:

M P P A_{p} = \sum_{l} (Y_{l} \cdot Z_{p, l})

$Y_{l}$ is the audience share of outlet l (1-point scale) and $Z_{p, l}$ is the percentage of experts that think that outlet l favours party p. The formula above considers the percentage of expert assessments of partisan bias of the two main TV channels and the three main newspapers of the country of the given party. The resulting MPPA score accounts for the level of affinity between a party and the aforementioned mainstream media outlets of a given country. By construction, our measure ranges from 0 to 100 (see the appendix for a detailed description of the procedure applied to German parties). We consider MPPA as a measure of how (un)favourably parties are treated in the media, both in editorials and in general news coverage. Kahn and Kenney have shown that in US Senatorial elections, the information on news pages was slanted “in favor of the [incumbent] candidate endorsed on the newspaper’s editorial page.”⁷¹ The results of studies in the United Kingdom,⁷² Germany,⁷³ and Portugal⁷⁴ also corroborate spill-over effects from the editorial pages to news coverage.

Overall, the use of two different indicators of media bias increases our confidence in the results and the validity of the analysis as a means not only to detect recognizable partisan editorial lines but also to uncover more subtle imbalances within media content.

Data Analysis

To find support for hypothesis 1, we analyze the scores yielded by our indicators of partisan media bias—PMU and MPPA—for all major parties in fourteen WE and ten CEE democracies. We compare the average degree of media bias towards parties in the two regions by using t tests, and we run separate analyses for (a) TV news, (b) newspapers, and (c) both news types combined.

To find evidence for hypothesis 2 and assess the direction of media bias in CEE, that is, whether more left-of-centre or more right-of-centre parties tend to be more favourably treated by the media, we run a series of linear regression models at the party level in the region, and we compare them with the same models in “the West.” In all the regressions, the dependent variable (or DV) is the extent of partisan media bias—measured by either PMU or MPPA—and the left–right ideology of each party is the main independent variable (or IV).

The ideological position of each party—main independent variable—is measured through the average left–right placement attributed to this party by citizens surveyed by EES in each given country, on a scale from 0 to 10. Higher values of this variable indicate a more right-wing party. The variable captures the average perception of left–right party positions within each country. The approach we rely upon to construct this variable has been used in the past to investigate the determinants of voter turnout or the extent of campaign professionalization of European parties.⁷⁵

We acknowledge that the left–right continuum captures a variety of dimensions and conflicts across cultures.⁷⁶ The labels “left” and “right” may cover somewhat different meanings and policy positions across countries. However, as Bohrer et al. put it, “few would argue that post-communist nations are in a state of complete ideological disorder.”⁷⁷ As an example, citizens surveyed by EES in CEE countries consistently place parties according to their general labels, such as “social-democrat” or “conservative,” on the opposite sides of the aforementioned continuum. Furthermore, we calculated the mean standard deviation of left–right party locations in both regions. Similarly to Huber and Inglehart,⁷⁸ we could find only slight differences between standard deviations in CEE and WE countries (for WE, SD = 2.35; for CEE, SD = 2.66).⁷⁹ Therefore, the agreement among citizens about left–right party placements is hardly greater in WE than in CEE.

Comparisons between our measurements of left–right party placement and other researchers’ scales that use standardized indicators to assess ideological party locations also show that the left–right scale is valid not only in older but also in newer EU democracies. We computed several rank correlations between our left–right party scale and two other data sets based on party categorizations from two different approaches: (1) Huber and Inglehart’s analysis of assessments by experts,⁸⁰ and more recently (2) Manifesto Project Research group’s content analyses of party programs (CMP).⁸¹ In our comparison with Huber and Inglehart,⁸² we correlated the left–right placement of twenty-nine parties for CEE and fifty-nine parties for WE. The Spearman’s rank correlation coefficient between our values and theirs (perfect correlations being −1 and +1) yielded very high scores ( $ρ$ = .707 for CEE and $ρ$ = .918 for WE), both significant at the 99 percent level. The comparison with the Manifesto Project database on eighty-six parties for WE and sixty-eight parties for CE yielded lower but still significant correlation coefficients ( $ρ$ = .431 for CEE, and $ρ$ = .596 for WE), both significant at the 99 percent level. In short, the value and significance of the coefficients resulting from the comparison with two other approaches documents the robustness of our ideological party categorization in both old and new EU states.⁸³ Overall, these reliability checks, as well as the confirmation of a similar level of agreement among citizens of both CEE and WE when assessing party positions in a left–right continuum, make us confident that our left–right party scale is both valid and reliable across Europe.⁸⁴

Finally, we consider further control variables in our regression models of Partisan Media Use and Media–Party Political Agreement (DVs) on party ideology (IV).⁸⁵ First, country dummies are included in all our regression models to account for idiosyncratic factors that may have an effect on the variation of the dependent variable from one place to another. Additionally, we consider two further predictors: vote share of the party and incumbency.

As previously discussed, incumbent politicians and powerful political groups tend to be better treated by the media. We therefore expect journalistic deference to those most prominent political sources to represent a distortion to the accurate assessment of partisan media bias. Incumbency is a binary variable that refers to whether each given party was in power shortly before May 2010, as reported by the Norwegian Social Science Data Services (NSD). The size of the party is operationalized as the percentage of vote share of each given party in a parliamentary election before May 2010—when the EMSS was finished—as reported by the NSD’s European Election Database (EED). In all models where we assess the direction of partisan media bias—whether it is leftist or rightist parties that benefit the most from media coverage—we use Huber-White robust standard errors to minimize the effects of potential violations of homoscedasticity.

Results

The findings of this study are reported in two sections: the first accounts for the results of t tests used to compare the levels of partisan media bias in WE and in CEE; the second section presents the results of the series of regression models that allow us to investigate the direction of partisan media bias in both regions.

Media Bias in the “East” and in the “West”

We begin by reporting the degree of partisan media bias in WE and CEE democracies through the average values of our two indicators—PMU and MPPA. As shown in Table 1, the average PMU is low in both regions, and slightly higher in WE than in CEE countries, although non-significantly. Similarly, the average MPPA is greater in WE (M = 11.09, SD = 13.50) than in CEE countries (M = 9.63, SD = 11.00), but this difference is again not significant (t[185] = –.78, p > .05).⁸⁶

Table 1

Average Partisan Media Uses in Central–Eastern and Western Europe

	Central and Eastern Europe		Western Europe
	Mean	SD	Mean	SD	t	df	p
Partisan Media Use	0.035	0.034	0.043	0.045	−1.192	185	0.235
Partisan Newspapers Use	0.022	0.024	0.033	0.032	−2.359	185	0.019
Partisan TV News Use	0.016	0.025	0.017	0.028	−0.103	185	0.918

Note: Values are mean scores on a continuous scale (0 = none, 1 = a lot). N = 187.

Source: European Parliament Election Study 2009.

Both average PMU and MPPA for press and TV separately score higher in WE than in CEE, although this difference is only significant for readers of main national newspapers (Partisan Newspapers Use).⁸⁷ Also, as shown in Table 1, average Partisan TV News Use is weaker than partisan press use in both regions. Analyzing MPPA separately for TV and newspapers shows a similar pattern. Although the relatively lower levels of Partisan TV News Use as compared to Partisan Newspapers Use are in line with previous research on current Czech, Hungarian, and Polish TV news,⁸⁸ they do not fit well with the historical development of television in CEE, as we outlined in the first section of this study. Most specifically, public broadcasters have been subject to state and partisan intervention even after the fall of communism. With the exceptional case of Poland—with a strong PSB in a reasonable economic position⁸⁹—however, market shares of public TV channels in CEE (M = 17.81, SD = 11.98) are considerably lower than in WE (M = 37, SD = 10.39, t[22] = −4.18, p < .001), which could add to the explanation of the overall weak partisanship in TV news reflected in our data for CEE.⁹⁰ In short, our first hypothesis was not supported.

Media Bias towards Left-of-Centre and Right-of-Centre Parties

Does being a left-of-centre or right-of-centre party matter for gaining media favourability, as our second hypothesis suggested? Our findings on the direction of media bias in formerly communist countries yield weak relationships between Partisan Media Use or Media–Party Political Agreement, and party ideology. Neither left-of-centre nor right-of-centre parties are treated systematically more favourably than their opponents by mainstream media in CEE countries.

Table 2 shows the results of six models considering MPPA (Models 1a, 1b, and 1c) and PMU (Models 2a, 2b, and 2c) as dependent variables and party ideology as main independent variable, both in WE and in CEE. In Models 1a and 1b, we correlate party ideology with MPPA and PMU respectively. To confirm the robustness of such relationships, a variable accounting for party incumbency is included in Models 2a and 2b, which is coupled with a second control variable—party vote share—in Models 3a and 3b. By so doing, we can better show the level of distortion added by each control variable to the relationship between party ideology and media bias towards each party.

Table 2

Ordinary Least Squares regression of Partisan Media Bias on Party Ideology

	Model 1a (MPPA)		Model 1b (MPPA)		Model 1c (MPPA)		Model 2a (PMU)		Model 2b (PMU)		Model 2c (PMU)
	“West”	“East”	“West”	“East”	“West”	“East”	“West”	“East”	“West”	“East”	“West”	“East”
Party ideology	.334***	.216	.229***	.208	.120*	.103	.170**	−.061	.136*	−.060	.084	−.122
Incumbency			.385***	.251	.049	.043			.114	−.011	−.050	−.159
Vote share					.816***	.797***					.346**	.405**
R ²	.124	.121	.238	.174	.658	.627	.334	.367	.345	.367	.427	.497
N	113	74	113	74	113	74	113	74	113	74	113	74

Source: European Parliament Election Study 2009 and the European Media Systems Survey 2010.

Note: Entries are standardized regression coefficients. Country dummies are included in all models. PMU = Partisan Media Use; MPPA = Media–Party Political Agreement.

Significant at the 10% level; **Significant at the 5% level; ***Significant at the 1% level.

Party ideology accounts for the variation of media bias in five of our six models only in Western Europe. As party ideology increases by 1 standard deviation, MPPA in Models 1a, 1b, and 1c increases by 0.33, 0.22, and 0.12 SDs, and PMU in Models 2a and 2b increases by 0.17 and 0.13 SDs, respectively. Our results indicate growing levels of media bias towards parties in “the West” the higher their average left–right placement is. Therefore, the more rightist a party is, the greater its probability of receiving support from the media in Western Europe. By contrast, for parties in CEE we find a negative relationship between left–right ideology and PMU (Models 2a, 2b, and 2c), but this correlation is not statistically significant. Ideological placement of parties does not account for any variation of media bias in any of our models when run in “the East.”

It is worth highlighting the crucial role of the size of the party (vote share) in determining media support. For example, in Model 1c the large coefficient of vote share decreases the significance of party ideology as a predictor of MPPA in WE, as compared to Model 1b, although it is always positive and significant. Furthermore, the inclusion of vote share in the model increases dramatically the p values of the coefficients of the variable incumbency, although initially this predictor was important in explaining media favourability. Having discarded concerning levels of multicollinearity between vote share and incumbency, one may suggest that the strong correlation between vote share and MPPA (r = .74 for WE and r = .70 for CEE systems) could explain the particularly strong effects of the former in Model 1c.⁹¹ In short, our second hypothesis was also not supported.

Discussion and Conclusion

Our findings reveal similar levels of partisan media bias in both WE and CEE. Regarding the press, political imbalances are actually stronger in the former than in the latter. Our first hypothesis was therefore not supported. In addition, we could not find systematic patterns regarding the direction of any biases in CEE. Contrary to the prediction of our second hypothesis, left–right party ideology appears not to be correlated with the reception of favourable media treatment in CEE media outlets.

It has been argued that the transformation of former Eastern Bloc media systems would imply a departure from the previously partisan role of journalism to a less partisan style as is common in WE.⁹² Our results do indeed show substantial similarities between CEE and WE regarding their low degree of partisanship in news coverage. However, we also show that party ideology only explains media favourability towards (rightist) parties in WE. In spite of the vast majority of literature showing that media are still politicized in CEE, in combination our findings suggest that tracing partisan biases in the region can be particularly challenging. These results may be an indication that partiality in media coverage in “the East” does not follow partisan lines. Accordingly, the media would be neither punishing nor supporting political parties, but rather promoting or opposing political actors depending on which way the political wind is blowing. This can be rooted in CEE’s short democratic experiences in the following ways.

Firstly, media moguls in the region engage in elite networks at the service of particular political or business goals. A landscape of changing and individual-centred clientelistic exchanges can make it difficult to trace a segmentation of the media across partisan lines.⁹³ Secondly, previous research has documented higher levels of electoral volatility in CEE in comparison to WE.⁹⁴ This higher electoral volatility mirrors changing political preferences and is especially pronounced in countries where alignments along the left–right axis compete with other ideological streams, for instance those resting on the old Russian communism and the new nationalist reformism, like in the Baltic region.⁹⁵ This could have played a role when measuring the extent and direction of media bias by means of one of our indicators, Partisan Media Use, making it difficult to track any meaningful relationship between the probability of voting for a party and the frequency of reading or viewing certain media outlets. Thirdly, although previous studies have argued that media favourability towards leftist parties may have added to their political success in several CEE countries, our findings can be read as mirroring the overall instabilities of political structures, for example, frequently changing and expanding coalitions, often integrating newly created or splinter parties from different sides of the ideological spectrum.⁹⁶

In summary, the youth and mutability of CEE political systems—whether in the form of high volatility, changing governmental coalitions, or clientelism built on particularistic interests rather than mediated by political parties—might be linked not only to the thrust for politicians to utilize the media, as research has already shown, but also to the difficulty of identifying media bias along partisan lines. These instabilities may help explain the weak explanatory power of the left–right ideology of parties on media support in “the East,” and show the need for future research that accounts for the connection of the media to political contingencies when analyzing news media bias in CEE countries.

By contrast, our results indicated that the media in WE favours rightist parties. This finding shows that WE systems are characterized by more identifiable and predictable patterns in the way parties and media interact. This evidence, coupled with the tendency of the media to support incumbent parties, as our results also document, speak to the stability of Western political systems. Previous research has argued that the level of media autonomy from politics in Western democracies is high and that a long tradition of commercial media and objective journalism has brought more political balance in “the West.”⁹⁷ However, a strand of literature has argued that the economic difficulties and strong market competition faced by Western mainstream media enhance fragmentation of media outlets across partisan lines.⁹⁸ Large audiences, audience segmentation, and market competition in a context of declining profit make partisan niche locations a successful strategy of product differentiation.⁹⁹ Against this background and in the light of our results, the thesis upon which Western mainstream media’s favourability towards traditional economic right-wing claims is the result of media companies’ dependence on advertising revenues needs to be addressed by future empirical research.¹⁰⁰

Finally, our study comes with a number of limitations. Firstly, our measure of partisan media bias relies on two proxies, namely, media users’ political preferences and experts’ assessments of media partisanship. The high demands of gathering content data on media’s favourability towards parties have prevented us from analyzing news media coverage across countries. Secondly, the correlation score between both indicators of media bias shows relatively low consistency, most likely because they are neither based on the same methodology nor measuring exactly the same phenomenon. Nevertheless, previous literature has measured media bias by the same means, and we found evidence that both indicators behave similarly when related to party ideology, which makes us confident that they are capturing (parts of) the same latent construct—media bias.

Thirdly, we acknowledge that CEE is not a homogeneous region (nor is Western Europe). Our study builds upon previous theoretical and empirical research which found similarities between the countries of CEE in terms of media change, the interplay of media and politics, and the nature of party competition.¹⁰¹ Previous studies have also dealt with Western Europe in comparable terms. Some scholars have looked at the distinctive organizational and regulatory frameworks of Western European media,¹⁰² while others have pointed to shared characteristics of their political communication and media systems as a result of similar economic development and common political history.¹⁰³ Furthermore, other authors have engaged in empirical comparative analysis of the structure of political competition between WE and CEE.¹⁰⁴

Fourthly, our indicator of party ideology relies on the mean left–right placement attributed to each party by EES 2009 respondents. Some scholars have argued that conceptions of “left” and “right” may be different in WE and CEE and,¹⁰⁵ most importantly, between countries within these regions.¹⁰⁶ Our correlation tests, as well as the additional analyses of the extent of agreement among citizens on party locations, however, made us confident that citizens in both regions are, on average, capable of locating parties with different ideologies in different spaces on the left–right continuum.

Finally, the lack of longitudinal data makes it difficult to confirm whether the low levels of partisanship found within CEE media are not just mirroring unstable political and media systems but are also reflecting a consistent shift toward neutrality over time. Perhaps more importantly, our unexpected results have cast doubt upon the widespread belief that Western media exhibit less partisan bias than the media further east. Future research is needed to identify the exact reasons behind these political imbalances.

This article is, to the best of our knowledge, the first to conduct a comparative analysis of partisan media bias in CEE and WE. By showing consistent differences between CEE and WE partisan media bias, this study makes a contribution to the debate on the extent of media transformation in CEE countries and the degree of convergence between WE and CEE. The low levels of partisan media bias found in “the East,” and the lack of a systematic pattern towards left-of-centre or right-of-centre parties in the region may reflect the contribution of the media to the institutional instability of their young political systems. The indications of a slant towards rightist parties found in WE media points to more predictable patterns of how media and politics interact, and shows that a longer tradition of privately owned media is not necessarily a guarantee against imbalance and partiality in media coverage. Overall, our findings point to the need for further comparative research to investigate the effects of commercialization, media ownership, and professionalization of journalism on political balance in news media. Our results also ask for longitudinal studies able to capture the impact of electoral volatility and party system changes on partisan media bias in CEE.

Footnotes

Appendix

Acknowledgements

We thank Frank Esser and Manuel Puppis for their valuable comments and suggestions to a previous version of this paper. We are also grateful to two anonymous reviewers who contributed to improve the quality of the paper.

Notes

Laia Castro-Herrero is a research and teaching assistant at the Department of Communication and Media Research at the University of Fribourg and a PhD candidate at the National Centre of Competence in Research – Democracy in Switzerland.

David Nicolas Hopmann is a Professor (with special responsibilities), at the Centre for Journalism, Department of Political Science at the University of Southern Denmark in Odense, Denmark.

Sven Engesser is senior research and teaching associate at the Institute of Mass Communication and Media Research (IPMZ) at the University of Zurich. He received his PhD from LMU Munich. His research interests include political communication, journalism, science communication, and media systems.