Media as the missing link? Demonstrating the electoral profitability of issue ownership

Transcription

1 Media as the missing link? Demonstrating the electoral profitability of issue ownership Gunnar Thesen, Christoffer Green-Pedersen and Peter B. Mortensen Paper prepared for the DPSA annual meeting in Vejle, October 23-24, POLIS-research unit Department of Political Science Aarhus Uniersity

2 Abstract The theory of issue ownership posits a positie relationship between electoral support and the leel of public attention to issues that a party owns, i.e. is perceied to be best at. Parties therefore work to increase attention to owned issues in order to boost their electoral support. So far, studies hae documented seeral empirical implications of the ownership perspectie with regards to electoral outcomes, ote choice, election campaigns and political agenda-setting. Howeer, an implicit but crucial assumption in this influential theory still remains practically unexplored. In mediatized politics, the media is the most important source of information and channel of communication between oters and parties. Thus, if the ownership mechanism in fact works - linking parties issue strategies, public issue agendas and electoral support - it should be possible to trace a significant leel of changes in party support back to changes in the media s issue attention. Based on this idea, and using a comprehensie data set combining monthly opinion polls with more than radio news items from in Denmark, we show that more media attention to owned issues increase party support. The results proide support for the ownership theory, at the same time contributing significantly to our understanding of media influence on democratic politics. 1

3 Introduction Do the media influence party support and electoral outcomes? Despite increasing scholarly attention to the relationship between media and politics, political science has yet to produce a clear and conincing answer to this basic but crucial question. The paper addresses this gap through studying how the issues that appear in the news affect the support of political parties. The theory of issue ownership (Petrocik 1996) ascribes electoral adantages to those parties that are perceied to be best at salient issues. Parties therefore work to increase attention to owned issues in order to boost their electoral support. Studies hae documented seeral empirical implications of the ownership perspectie with regards to electoral outcomes, ote choice, election campaigns and political agenda-setting. Howeer, a crucial aspect of the ownership mechanism remains practically unexplored. In mediatized politics, the media is the most important source of information and channel of communication between oters and parties (Strömbäck 2008). Gien that the media exert substantial influence on the issue priorities of the public (cf. McCombs 2004), and that issue ownership works to the electoral benefit of parties (cf. Belanger & Meguid 2008), the media hold the potential to influence party competition and affect electoral outcomes. In other words, if the ownership mechanism in fact works it should be possible to trace a significant leel of changes in party support back to changes in the media s issue attention. Based on this idea, and using a comprehensie data set combining monthly opinion polls with more than radio news items from in Denmark, we show that more media attention to owned issues increase party support. The results contribute significantly to our understanding of issue ownership, party competition and mediatized politics. The media agenda is if not exactly a missing link an indispensable supplementary factor in understanding electoral politics and party support. 2

4 The mediatization of politics and the meagre response of political science The term mediatization of politics (cf. Mazzoleni & Schulz 1999; Strömbäck 2008; Esser & Strömbäck 2014) reflects the increasing influence of media in political life. Deelopments like the decline of party membership and end of the party press, and of course the increasing spread and use of different communication media, means that media institutions of today are the most important arenas for communication between the public and political sphere, as well as autonomous political actors (Strömbäck 2008). The possibilities of political parties of coneying political messages to the electorate other than through the mass media play a largely symbolic role. Combined with the increasing tendency of oters to switch parties (Mair et al. 2004), the media s role in politics would seem to warrant considerable research attention. The response of contemporary political science to these changes has neertheless been careful. Looking at for instance partisan research, the literature clearly acknowledges the importance of the media (Semetko 2006). But in terms of substantial focus, partisan research has mostly concentrated on the organizational consequences of these changes, studying for instance how parties hae become campaign organizations (Mair et al. 2004). The emphasis on organizational aspects is often coupled with an increasing focus on describing party actiities during election campaigns (Farrell & Schmitt-Beck 2002). The response from electoral research offers some supplements, but not a lot considering the debate about and need for exogenous ariables in this field of study (cf. Wlezien et al. 1997). For instance, the literature on economic oting mostly ignores the media when examining how parties perform at the polls, concentrating on determining the effect that different economic indicators hae on ote choice. In a notable exception to the literature s oersight, Hetherington (1996) finds that the media s constant negatie framing of the economy contributed to Bush s defeat in the 1992 US presidential campaign. Related to this literature is also the concept of cost of ruling, documenting how goernments generally lose about 2 percent of the ote (Steenson 2002). Een though some studies contain ideas about the significance of media attention in this mechanism (cf. Rose and Mackie 1983), there is no attempt at exploring this in theoretical models or empirical tests. 3

5 Not surprisingly, more responses are to be found at the intersection between political science and communication research. Mediatization studies hae spearheaded efforts to make sense of the mediatization of politics, defining it as a long-term process through which the importance and influence of the media in political processes and oer political institutions and actors hae increased (Esser & Strömbäck 2014). Analyses indicate that journalists retain the most power oer the content and framing of news (Strömbäck & Nord 2006) and that mediatization has affected the organization of European political parties (Donges 2008; referenced in Esser and Matthes 2013) and the (media) behaior of Members of Parliament (Elmelund-Præstekær et al. 2011). Neertheless, there is still a remarkable dearth of systematic empirical research on the mediatization of politics (Strömbäck 2011: 423), in particular when it comes to the issue content of the media and its effect on politics. In this regards political agenda setting offers systematic empirical studies of the media s role in promoting social problems on the political agenda (Thesen 2014). A key assumption in this literature is that news stories are the central attention input to politicization process (Green-Pedersen 2011). This does not mean that the media necessarily determine which issues that will receie attention from political parties. Logics of party and issue competition are crucial when predicting whether media-generated opportunities of politicization are seized. Accordingly, parties tend to respond to news that draw attention to own strengths, such as owned issues (Green-Pedersen & Stubager 2010; Vliegenthart & Walgrae 2011; Thesen 2013). While being empirically strong, political agenda-setting work still remains undertheorized (an Aelst et al 2014). Discussions of what the empirical findings signify are ery often lacking or insufficient (Thesen et al 2013). For instance, although we hae learned that the media influence the issue attention of political parties, this still doesn t answer the question about how the media through agenda-setting processes affect party support. Other literatures in the field of media and communication studies examines many different ways in which media content influences the attitudes, opinions and behaiour of the public. Public agenda-setting, for instance, has repeatedly documented the media s ability to change the issue priorities of the public (see oeriew in McCombs 2004). Yet, what is generally missing are studies that concentrate on how the issues in the media affect the public s ote. 4

6 The media agenda, issue ownership and party support Our approach to this gap is based on a few oerarching assumptions that desere to be spelled out. First, from our perspectie, a central part of the shift from class politics to media politics is the increased importance of agenda-setting in politics. When oters increasingly ote on the basis of issues instead of social structure (Thomassen 2005), the need to understand how issue attention rise and fall and how it affects electoral politics becomes eer more pressing. In this perspectie, the implications of mediatization are straightforward: If the electorate is olatile, their issue agenda is crucial for their ote, and political parties can only reach the electorate through the media then the interaction between the public, the media and political parties should be a core research theme for partisan and electoral research. Second, studies of political communication as well as party and electoral research hae been strongly biased towards the study of election campaigns. Quite understandably, this relates to the central importance of elections in representatie democracies and the strong research infrastructure deeloped around election sureys. Howeer, to understand the continuous process of media influence on party competition, we cannot look at election campaigns in isolation. In other words, there s a lack of systematic knowledge about mediatized politics during routine times. The study therefore examines seeral full electoral cycles, hopefully contributing to the increasing interest among scholars to study the eolution of electoral sentiment oer time (Jennings and Wlezien 2014). Third, we take issue with the zero-sum game interpretation of mediatization where gains for some actors come at the expense of others (Strömbäck and Dimitroa 2011). Mediatization of politics, thus understood, implies an a apriori conclusion of decreasing political influence (cf. Strömbäck 2008). We wish to challenge this notion. The effect of mediatization on different political actors is an empirical question; some might lose, but others could win (Thesen 2014). Reintroducing a key question put forward more than three decades ago (Hernes 1978), but which has remained largely unanswered, our goal should be to find out how the media influence the distribution of power between different actors in politics. Our research question is arguably one of the most important ones in this regard: Do the media, through their coerage of different issues, affect the electoral competition between parties? 5

7 In answering this question we take our starting point in the theory of issue ownership (Petrocik 1996), which proides a link between the issue content of the media and oter s choice of parties/candidates. The idea is that political parties are perceied to be best at certain issues, and that they stand to gain electoral support when political debates address these issues. The bulk of electoral research that has dealt with issue oting only examines how the issue priorities of oters affect their ote, leaing the question of where their agenda comes from unanswered (cf. an der Brug 2004; Aardal & an Wijnen 2005). Howeer, two studies of the Netherlands (Boomgarden and Vliegenthart 2006) and Belgium (Walgrae and De Swert 2004) respectiely, indicate that media coerage of immigration and crime to some extent could explain the increased support for anti-immigration and extreme right parties. And Kleinnijenhuis et al. (2007) find that during the 2003 campaign, also in the Netherlands, news about a party s position on an owned issue increased the probability of oting for that party. Ansolabehere and Iyengar s (1994) experiments in a US setting corroborate these findings, showing that news on crime boost Republican support while news on unemployment faor the Democrats. The most recent exception to the lack of interest in the relationship between the media s issue agenda and party support is another experimental study set in a multiparty context. In two surey experiments of Norwegian oters, Beyer et al. (2014) draw attention to immigration and enironment whereupon the parties that own these issues increased their ote share. While sketching the rudiments of an interesting research agenda, these studies are of clearly limited reach. We need to compare different parties and campaigns, coer a wider range of issues and most importantly, capture the deelopment of electoral sentiment in between election campaigns. Still, although the research reiewed is limited and scattered it is a reasonable assumption that the media affect the support of political parties through highlighting some issues at the cost of others. Put differently, we expect that increased media attention for owned issues lead to an increase in public support. 6

8 Research design and data In examining the link between the media agenda, issue ownership and party support we utilize a comprehensie data set combining monthly opinion polls with more than radio news items from in Denmark. Denmark is a multiparty system sharing characteristics with seeral European PR-systems. It has been noted that issue ownership is more fluid and contested in multiparty systems (Karlsen 2004). The extent to which Danish experiences are releant for two-party systems is therefore debatable. Howeer, despite a multiparty system with clear consensual features, Danish party competition is characterized by two blocs of parties proiding clear goernment alternaties (Green-Pedersen & Thomsen 2005). This means that the opposition-goernment diide is organized around a bloc of left-wing and a bloc of right-wing parties. We therefore approach the question of issue ownership from a bloc-perspectie. As a consequence, the instability in issue ownership that characterizes multiparty systems (relatie to two-party systems) due to within-bloc party competition is more or less ruled out. The results of the study should in other words also bear releance in a two-party context. The left-wing or Social Democratic led bloc consists of the Social Democrats, the Socialist People s Party and more extreme left-wing parties (eg. the Red Green Alliance). Typically, the Social Liberals hae also supported the left-wing bloc, although in certain periods, it has supported the right-wing bloc. This was for instance the case from 1982 to The right-wing or bourgeois bloc consists of the Liberals and the Conseraties and the radical right-wing parties (Progress Party, Danish People s Party). The two small center parties, Centre Democrats and Christian People s Party, hae typically also supported the right-wing bloc. Note that the Danish tradition for minority goernments implies that not all parties from a bloc actually take part in the goernment. Howeer, all parties clearly belong to a bloc when it comes to the question of goernment formation. The bloc definition applied in the analyses therefore includes both parties in office (for the opposition: alternatie goernment parties) and the support parties that are part of the goernment s parliamentary basis (for the opposition: support parties for the alternatie goernment). In terms of media system, Denmark is a typical example of Hallin and Mancini s (2004) democratic corporatist model (Allern & Blach-Ørsten 2011). Still, the simplicity of our argument 7

9 linking parties ownership of the news agenda to their public support suggests that the findings should be generalizable outside the North European context. Different media systems generally fulfill the key assumption of our argument: Politics is mediatized in the sense that oters get their information about parties, positions and policies from the media. We do acknowledge that the issue content of the media might differ across systems, due to for instance ariation in political parallelism, state interention and the emphasis put on commercial news alues. For instance, one could expect Liberal media systems to be een more focused on crime which in turn would mean that the partisan bias introduced by the news agenda would be different compared to the Danish context. Future comparisons will reeal whether such expectations hold true. For now, and from the perspectie of our case selection, the most important aspect is that the general mechanism neertheless would be the same although the issue content may ary. Seeral data sources are integrated in this study, combining information related to media content (the issues on the news agenda), public perception of parties issue competence (issue ownership), political context ariables (eg. bloc in office, time in office), economic deelopment and party support (ote intention from polls). Below we proide more details to each of the sources and to our measures of the argument proposed in the paper. Starting with the media agenda, we draw on data from Danish radio news in the period The basic dataset, which contains more than 114,000 news features 1, proides a unique opportunity to study media content and its effects oer a long time frame and in different political contexts. The radio news was produced by the Danish Broadcasting Corporation (DR), which enjoyed a de-facto monopoly on broadcasting radio news nationally in this period. Studies of the Danish media system (Lund 2002) indicate that many stories originate in the major national newspapers but that the radio news was the most important filter for stories raised in the newspapers to make it into the TV news in the eening. Radio news thereby constitutes the best 1 The original dataset contains more than news features. We hae excluded the large number of news features (approx. 70,000) that reports political news from abroad with no direct relation to Danish politics. This could for instance be a ciil war somewhere or a US presidential election. 8

10 single source for measuring the agenda of the mass media in general in Denmark. 2 News features were coded twice daily using the issues addressed in the afternoon and at 6.30 pm when long ersions of the hourly radio news were broadcasted. Each news feature was coded with regard to both issue content and political actors. The issue content coding was conducted using a modified ersion of the Danish policy agendas coding scheme 3. The scheme contains 27 policy issues, eg. the enironment, economy, social welfare, transportation (see full list in Table A1). The issue coding was subsequently combined with data from Danish election sureys 4 that measure issue ownership through seeral elections. The surey questions tap oters competence images of the two blocs (Social Democratic or bourgeois led coalition) for different issues, and so ownership was assigned to the media data using the existing issue content coding of the radio news database. The resulting issue ownership ariable thus indicates whether a Social Democratic or bourgeois bloc holds ownership of the issues on the media agenda, or whether instead no bloc enjoys an electoral adantage (labeled unowned issues). The electoral studies generally show consistent patterns of bloc-leel issue ownership oer time (Andersen 2003). While the bourgeois bloc owns economy, trade and industry, crime, justice and immigration, the sureys indicate Social Democratic ownership of unemployment, enironment, social welfare, health and housing. The main unowned (or contested) issues are education, foreign affairs, energy and traffic. For a complete oeriew of how the 27 issues hae been assigned to the blocs see Table A.1 in the appendix. The data on bloc support for the period from are based on monthly sureys performed by Gallup. 5 For each month the ote intention for parties belonging to the two blocs were aggregated, thus forming the dependent ariable of monthly bloc-leel ote intention. To control for the well-documented impact of economic conditions on party choice (Lewis-Beck & 2 Probably because of the limited tradition for self-made stories, the Danish radio news hae neer been subject to any regulation of their coerage during election times, such as requiring a certain balance in their coerage. They also hae no reputation for haing a particular political color. 3 See The coding scheme was originally deeloped by Baumgartner and Jones for the US (see 4 Made aailable from by The Danish Election Project. 5 Data kindly made aailable by Jacob Askham-Christensen (2012). 9

11 Stegmaier 2007), we include a ariable measuring economic deelopment in Denmark. This measure is based on the OECD s Main Economic Indicators database. More specifically, it shows the so-called Composite Leading Indicator (CLI) which is built on a wide range of key short-term economic indicators. 6 Table 1. Variable descriptions and descripties. Variable Description Mean S.D. Vote intention Support (in pct) for the Social Democratic / Bourgeois bloc in opinion polls Bloc change Change of bloc composition in 1993 (Social Liberals). Values from 6 (month of the change) to 1 (6 months after) Change in office Change of goernment. Values range from 6 (month of the change) to 1 (6 months after) Bloc in office Dummy indicating whether the bloc held office Social bloc Economic deelopment News saliency Democratic Dummy separating the bourgeois bloc (0) from the Social Democratic bloc (1) OECD-measure of short-term economic deelopment News features about this issue category (see ariable below) as pct of all news features this month Issue ownership 1=Social Democratic, 2=unowned, 3= bourgeois ownership Note:We show descripties of the original ariables, although most ariables in the models are first differenced. Table 1 describes the aboe mentioned ariables, as well as the control ariables which are included in our analytical models. The dataset coers the 20 years from 1983 to 2004 and has obserations on a monthly basis for each of the two blocs across owned, unowned and opponent owned issues, meaning that we hae: 240 months X 2 blocs X 3 issue categories = 1440 units. 6 See for details. The CLI is a times series consisting of seeral indicators which show a reasonably consistent relationship with a reference series (e.g. industrial production, GDP) at turning points. It proides qualitatie information on short-term economic moements, especially at the turning points, meaning that its main message is the increase or decrease, rather than the amplitude of the changes. 10

12 Results We begin by looking at the deelopment in bloc ownership of the media agenda during our period of inestigation. Figure 1 indicates that most of the news agenda could be labeled unbiased from an ownership perspectie, as on aerage approximately 38 % of news features fall into the category of unowned issues. Comparing the two blocs, we find a dominance of bourgeois ownership (34 %) relatie to social democratic ownership (28 %). There is considerable ariation from month to month, but no clear trend in terms of an increasing bias towards either of the blocs oer time. 100% Social democratic owned issues Unowned issues Bourgeois owned issues 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Fig. 1. Bloc ownership of the media agenda, Denmark

13 ,00 50,00 40,00 30,00 Bourgeois VI Social democratic VI 20,00 10,00 0,00 Fig. 2. Vote intention for Social Democratic and bourgeois bloc, Denmark Studying the deelopment of ote intention for the two blocs, much of the same can be said. There is substantial monthly ariation, but no unambiguous trend towards increased dominance of any of the blocs. Rather, in line with the bloc-based party competition of the Danish system, we find that the shifts in public opinion corresponds to the three bloc-alterations in office during this period. In 1993 the left-wing bloc took office for the first time in our 20 year period. Although their electoral adantage decreased quickly they were able to stay in position until Inestigating the relationship between these two sets of time series, ownership of the media agenda on the one hand and ote intention on the other, we start by noticing a positie biariate correlation of.275 (significant at the 1 percent leel). When the ownership series changes, in the sense that a bloc increases its share of owned issues in the news, the public support for this bloc increases. Moing on, we need a multiariate test of our claim that bloc ownership of the media agenda 12

14 affects bloc support, a test where the effects of all explanatory and control ariables are modeled simultaneously. Since our dataset includes seeral groups or clusters of data (blocs, months, issue categories), we estimate a set of multileel models based on restricted maximum likelihood (REML) estimations. The multileel analysis models the ote intention of both blocs (Social Democratic and bourgeois) across 240 months (from January 1984 to December 2003) repeated oer the three releant issue categories (owned, unowned and opponent owned issues 7 ). Table 2 presents the results from first differenced models. In terms of our key ariables of interests, results are similar when running leel models with a lagged dependent ariable. 8 Examining model 1 first, we find that both bloc composition and the economy exert effects on ote intention in line with what could be expected. The change in bloc composition in 1993, when the Social Liberals switched to the Social Democratic bloc, indeed turns up as a significant interaction effect in the model. Such a change increases ote intention for the left-wing bloc, naturally decreasing the support of the right-wing bloc. Regarding the economy we expect, in line with preious research (cf. an der Brug et al 2007), parties in office to be adantaged by good economic times and disadantaged by bad economic times, while opposition parties should experience the opposite. The economic deelopment ariable was therefore interacted with the bloc in office dummy producing a significant and positie coefficient in the models. Controlling for all the other factors in the model, a bloc that holds office will shrink during negatie turns in the economy and grow when the economy improes. Estimating predicted changes in ote intention (based on Model 2, Table 2), we find that the bloc in office will increase its support by 0.41 percentage points when the OECD measure of economic deelopment rise from one month to the next by 0.6 (corresponding to the sample maximum increase). When the economy worsens by -0.2 (corresponding to the sample maximum decrease), ote intention for the bloc in office is reduced by percentage points. The difference between the most optimistic and most pessimistic economic scenarios thus account for a 0.69 percentage point change in the support for the goerning bloc. 7 The results reported in the paper are reproducable also when we leae out the issue category of unowned issues. 8 Not reported here. Note that in leel models the lagged dependent ariable has a strong and positie effect. That is, preious month s leel of public support for a bloc will always be the best predictor of that bloc s performance the present month. 13

15 Bloc ownership of the media agenda Controls and economic deelopment Table 2. Models explaining Vote Intention (support for Social Democratic bloc and support for Bourgeois bloc). Monthly data, Denmark, First-difference models. (1) (2) Vote Intention Vote Intention Social Democratic bloc (0.089) (0.089) Bloc change *** *** (0.174) (0.174) Social Democratic bloc X Bloc change 2.598*** 2.557*** (0.202) (0.202) Change in office (0.100) (0.101) Bloc in office (0.117) (0.117) Economic deelopment (0.421) (0.420) Bloc in office X Economic deelopment 1.528* 1.511* (0.611) (0.610) News saliency 0.025* (0.011) Issue ownership* Owned issues s Unowned issues (0.103) Owned issues s Opponent owned (0.103) Issue ownership X News saliency* Owned issues s Unowned issues * (0.016) Owned issues s Opponent owned ** (0.016) Constant (0.100) (0.117) N 1,434 1,434 Wald chi *** 188.8*** Log restricted-likelihood Note: Standard errors in parentheses. Owned issues used as reference category. *** p<0.001, ** p<0.01, * p<0.05, p<0.10 Model 2 in Table 2 is supplemented with our key interest in this paper, which is of course bloc ownership of the news agenda and its effect on ote intention. In order to capture this mechanism we interact the news saliency of an issue category a gien month with the issue ownership ariable. 14

16 Predicted change in ote intention Note that owned issues is the reference category for this interaction coefficient, and that the two rows at the bottom of model 2 thus test the difference between owned issues on the one hand, and unowned issues or opponent owned issues on the other hand. In both cases the difference is significant and negatie, indicating that media attention to owned issues produce a positie effect on ote intention both in comparison to unowned issues and opponent owned issues. In other words, the issue ownership balance in the media affects the relatie strength of the two blocs. 1,0 0,8 0,6 0,4 0,2 0,0-0,2-0,4-0,6-0,8-1,0 Owned issues Unowned issues Opponent owned issues Change in news saliency Fig 3. Predicted change in ote intention (bloc support) when news saliency changes for owned, unowned and opponent owned issues. Figure 3 spells out this interaction effect, showing how predicted change in ote intention differs for owned, unowned and opponent owned issues as the monthly change in news saliency aries. 9 When media attention for bloc-owned issues rise from one month to the next by 20 percentage points (corresponding to the sample maximum increase), ote intention for the bloc increases by 9 See appendix for separate illustrations of owned, unowned and opponent owned issues respectiely. These illustrations also include confidence interals. 15

17 0.51 percentage points. Conersely, when news about bloc-owned issues fall by 25 percentage points (corresponding to the sample maximum reduction), bloc support is reduced by 0.62 percentage points. In other words, a difference in ote intention of more than 1 percentage point is at stake when comparing the two scenarios. Interestingly this exceeds the oerall change (0.69) related to changes in economic conditions discussed aboe. An opposite pattern is eident for changes in media attention to issues owned by the opposing bloc, while changes in news saliency for unowned issues seem unconsequential. Thus, it matters whether media attention to owned issues is replaced by media attention to unowned or opponent owned issues. In the latter case, the ote loss due to a decrease in news on owned issues is supplemented by the negatie effects accompanying increased attention to the opposing bloc s issues. Picture a situation where the media agenda changes so that the share of owned issues is reduced by 6.5 percentage points (corresponding to the sample standard deiation change) and where all of this attention is shifted to opponent owned issues. The total ote loss for the bloc equals 0.27 percentage points. This might not seem that much, but remember that this is the effect of a standard deiation change in the ownership composition of the news agenda. Furthermore, this negatie scenario should also be ealuated against its positie counterpart where the opposite agenda changes take place. The estimated gain in ote intention is 0.30 percentage points, which in turn approximates a 0.6 percentage point difference from the ote loss scenario. Een without considering the more abrupt agenda changes that there are many examples of in the period under inestigation, it is clear that bloc ownership of the media agenda has an effect on public support. When elections are competetie, this effect could mean the difference between office and opposition. 16

18 Conclusion Aiming to redress the meagre political science response to the mediatization of politics, we set out to examine whether the media, through their issue agenda, affect party support and electoral competition. The argument was that the news agenda is not politically neutral, as has been demonstrated in the literature on political agenda-setting (cf. Thesen 2013), but comes with specific political biases related to the issue ownership (Petrocik 1996) structure of a political system. When the share of owned issues in the news increases, parties stand to increase their support in the public. Our empirical test of this idea in the context of Denmark, coering all issues on the media agenda, shows a fairly strong and conincing effect of changes in bloc ownership to the news on bloc support oer a 20 year period. This news ownership and party support mechanism is especially interesting in the light of recent decades deelopments in both political and media systems. We already know that the public s agenda is influenced by the issue agenda of the media (cf. McCombs 2004), and that the process of mediatization hae made the media increasingly important as a source of information and channel of communication between oters and politicians (cf. Strömbäck 2008). At the same time, both the share of oters that switch parties (cf. Mair et al. 2004) and issue oting (cf. Thomassen 2005) hae risen. This combination of deelopments make the news agenda eer more crucial. Not only in shifting political attention, but also in shifting otes. To the extent that our test is generalizable, the conclusion is that party competition - and in consequence also electoral results is dependent on which issues receie media attention. No wonder then, that parties put a lot of effort into controlling the issue content of campaigns and public debate. Our study suggests that it is also ital to affect the agenda in between elections and that parties that succeed in this work, or for other reasons are fortunate in terms of the news agenda, will profit electorally. The conclusion ineitably raises the question whether the media-sceptic thesis of much mediatization research (cf. Meyer 2002) holds true: hae the media become too strong, rendering political parties a weaker force in democratic politics? Our approach has mostly ignored this question focusing instead on how the media contributes to the competition between political parties. We do acknowledge that the present study indicates a substantial role for the media in electoral politics. But in itself, this does not necessarily imply a weakening of political parties or 17

19 politics. On the one hand, one could argue that increased importance of commercial news alues (Landerer 2013) means that the mechanism examined in this study becomes a tool by which a typical media incentie structure or media logic (Esser and Mathes 2013) could increase its political influence. On the other hand, we know that the leel of political attention to an issue impacts the amount of news coerage it gets (cf. Bennett 1990). In other words, implications for the media-politics balance are not straightforward but subject to the forces that drie the news agenda (cf. Boydstun 2013). Due to our bloc-leel approach we hae argued that the Danish case bears releance for not only similar multiparty systems, but also a two-party context. Of course, although we beliee in the generalizability of the findings it is likely that the exact nature of the mechanism will ary across systems. As touched upon earlier, issue ownership is more fluid and contested in multiparty systems when looking at the party leel (Karlsen 2004). Consequently, the ownership effect on indiidual parties is likely to stand out more clearly in a two-party context. Another example of cross-country ariation relates to how mediatized politics interacts with the way different political systems handles the demands of organized interests. In more corporatist systems for instance, it has been argued that the negotiated consensus reached by interest groups through corporatie institutions work to constrain their party political counterparts in parliament (Nørgaard and Klemmensen 2009). Thus, the leel of party competition and conflict on typical corporatist issues (i.e. work enironment, labour market policies, pensions and benefits) is reduced. Does this mean that such issues receie less media attention and are less likely to affect party competition in corporatist systems compared to pluralist systems? If so, can we discern any electoral implications for the Social Democrats who typically own these issues? This brings us back to our own case, where there are still seeral interesting contingencies of the relationship between news ownership and party support left to explore. Additional models suggest that the bourgeois bloc could win, and lose, more than the Social Democratic bloc through the ownership effect. Further research is needed in order to inestigate this, and other contingencies, more closely. In other words we need to examine whether the media agenda introduces different (party) political biases into electoral competition, and we need to examine how this comes about. The latter question is not only about which issues dominate the news and which parties own these 18

20 issues. It is certainly also about how these issues, and the political actors that try to promote them, are framed in the news. The mediatization of politics, which one might define differently but which no one disputes, suggests that the media matter for politics. To push our understanding of mediatized politics we will hae to continue the search for the right research designs, leaning ourseles to Zaller s (1996: 18) claim that: «the true magnitude of the persuasie effect of mass communication is closer to massie than to small to negligible and ( ) the frequency of such effects is often. Exactly as common intuition would suggest, mass communication is a powerful instrument for shaping the attitudes of citizens who are exposed to it, and it exercises this power on an essentially continuous basis. The ery large media effects that, as I claim, are all around us are obiously somewhat difficult to obsere, since many carefully designed studies hae failed to detect them. But this, by itself, is no embarrassment to my thesis. ( ) What matter in the end is whether, under appropriate circumstances, a putatie effect can be made to stand out clearly against whateer background noise may also be present.» We hae tried to identify a particular set of appropriate circumstances relating to news agendas, issue ownership and party support. Hopefully the results could contribute to how media influence on democratic politics is understood. 19

21 References Aardal, B., & an Wijnen, P. (2005). Issue oting. The European oter. A comparatie study of modern democracies, Allern, S., & Blach-Ørsten, M. (2011). The news media as a political institution: A Scandinaian perspectie. Journalism Studies, 12(1), Andersen, J. G Partiernes image: De borgerlige bedst til at sikre elfærden, in Jørgen G. Andersen & Ole Borre, eds, Politisk Forandring, Århus: Systime, Ansolabehere, S. & Iyengar, S Riding the wae and claiming ownership oer issues the joint effects of adertising and news coerage in campaigns. Public Opinion Quarterly 58 (3): Askham-Christensen, Jacob (2012): The Logic of Political Competition and Goernment Responsieness in Parliamentary Systems. Odense: Uniersity Press of Southern Denmark Bélanger, Eric & Bonnie M. Meguid (2008). "Issue Salience, Issue Ownership, and issue based ote chocie ", Electoral Studies, 27, pp Bennett, W. L. (1990). Toward a theory of press state relations in the United States. Journal of communication, 40(2), Beyer, A., Knutsen, C. H., & Rasch, B. E. (2014). Election Campaigns, Issue Focus and Voting Intentions: Surey Experiments of Norwegian Voters. Scandinaian Political Studies. Boomgaarden, H. G., & Vliegenthart, R "Explaining the rise of anti-immigrant parties: The role of news media content." Electoral studies 26.2: Boydstun, A. E Making the News: Politics, the Media, and Agenda Setting. Uniersity of Chicago Press. Elmelund-Præstekær, C., Hopmann, D. N., & Nørgaard, A. S Does Mediatization Change MP-Media Interaction and MP Attitudes toward the Media? The International Journal of Press/Politics, 16(3), Esser, F. & Strömbäck, J Mediatization of Politics: Understanding the Transformation of Western Democracies, Basingstoke: Palgrae Macmillan. Esser, Frank, and Jörg Matthes Mediatization Effects on Political News, Political Actors, Political Decisions, and Political Audiences. In Challenges to Democracy in the 21st Century. Democracy in the Age of Globalization and Mediatization, ed. Hanspeter Kriesi, Daniel Boschsler, Jörg Matthes, Sandra Laenex, Marc Bühlmann, and Frank Esser, Hampshire, UK: Palgrae Macmillan. Farrell, D. & R. Schmitt-Beck (eds.) Do Politcal Campaigns Matter?, London: Routledge. Green-Pedersen, C. (2011). Partier i nye tider: den politiske dagsorden i Danmark (Doctoral dissertation, Aarhus Uniersitet, School of Business and Social Sciences, Department of Political Science and Goernment). 20

22 Green-Pedersen, C. & R. Stubager The Political Conditionality of Mass Media Influence. When Do Parties Follow Mass Media Attention? Brit J Polit Sci. 40 (3): Green-Pedersen, Christoffer & Lisbeth H. Thomsen (2005). Bloc Politics s. Broad Cooperation. The Functioning of Danish Minority Parliamentarism, The Journal of Legislatie Studies, 11, 2: Hallin, D. C., & Mancini, P Comparing media systems: Three models of media and politics. Cambridge Uniersity Press. Hernes, G Det mediaridde samfunn. In Forhandlingsøkonomi og blandingsadminstrasjon, ed. Hernes, Bergen: Uniersitetsforlaget. Hetherington, M "The media's role in forming oters' national economic ealuations in 1992." Am Journ of Pol Sci: Jennings, W. & Wlezien, C The Timeline of Elections: A Comparatie Perspectie. Working paper. Karlsen, R. (2004). Valgkamp i flerpartisystem aktualisering a saker og kamp om eierskap. Tidsskrift for samfunnsforskning, 45(4), Kleinnijenhuis, J, et al "A Test of Rialing Approaches to Explain News Effects: News on Issue Positions of Parties, Real-World Deelopments, Support and Criticism, and Success and Failure." Journal of Communication 57.2: Landerer, Nino Rethinking the Logics: A Conceptual Framework for the Mediatization of Politics. Communication Theory 23 (3): Lewis-Beck, M. S., & Stegmaier, M.. Economic Models of Voting. In The Oxford Handbook of Political Behaior. Edited by Russell Dalton and Hans-Dieter Klingemann, Oxford: Oxford Uniersity Press, Mair, P., W. Müller & F. Plasser, eds Political Parties and Electoral Change. London: Sage. Mazzoleni, G. & W. Schulz Mediatization of Politics: A Challenge for Democracy. Pol. Communication 16 (3): Meyer, Thomas Media Democracy: How the Media Colonize Politics. Cambridge, UK: Polity Press. McCombs, M. (2004). Setting the agenda: The mass media and public opinion. Polity. Nørgaard, A. S. & Klemmensen, R "Horfor stemmer oppositionen for regeringens loforsfag? Korporatisme og parlamentariske forlig i Danmark, " Politica: Tidsskrift for Politisk Videnskab 41.1 Petrocik, J Issue-ownership in Presidential Elections with a 1980 Case Study. Am J Polit Sci 40 (3): Rose, R. & Mackie, T. T Incumbency in Goernment: Asset or Liability?, in Daalder, H. P. & Mair, P., eds., Western European Party Systems: Continuity and Change. California: Sage Publications. 21

23 Semetko, H. (2006). Parties in the Media Age. In R. Katz & W. Crotty (eds.), Handbook of Party Politics, London: Sage. Steenson, R. T "The cost of ruling, cabinet duration, and the median-gap model." Public Choice (2002): Strömbäck, Jesper Mediatization and Perceptions of the Media s Political Influence. Journalism Studies 12 (4): Strömbäck, J Four Phases of Mediatization: An Analysis of the Mediatization of Politics. International Journal of Press/Politics 13 (3): Strömbäck, J, & Nord, L Do Politicians Lead the Tango? A Study of the Relationship between Swedish Journalists and Their Political Sources in the Context of Election Campaigns. European Journal of Communication 21 (2): Strömbäck, J., & Dimitroa, D. V Mediatization and Media Interentionism: A Comparatie Analysis of Sweden and the United States. International Journal of Press/Politics 16 (1): Thesen, G When Good News is Scarce and Bad News is Good. Goernment Responsibilities and Opposition Possibilities in Political Agenda-Setting. Eur J Polit Res, 52 (3): Thesen, G., an Aelst, P., Walgrae, S., Vliegenthart, R So What, Why Not and What Now? Agenda-Setting and the Concept of Media Influence on Politics, paper for Comparatie Policy Agendas Conference, Antwerp, Thesen, G. 2014, "Political agenda-setting as mediatized politics? Media-politics interactions from a party and issue competition perspectie", International Journal of Press/Politics, Vol. 19, No. 2, pp Thomassen, J., ed The European Voter. Oxford: Oxford Uniersity Press. an Aelst, P., Thesen, G., Walgrae, S. & Vliegenthart, R. 2014: Mediatization and Political Agenda Setting: Changing Issue Priorities?, in Mediatization of Politics: Understanding the Transformation of Western Democracies, edited by Frank Esser & Jesper Strömbäck, published by Basingstoke: Palgrae Macmillan. Van der Brug, W., Van der Eijk, C., & Franklin, M. (2007). The economy and the ote: Economic conditions and elections in fifteen countries. Cambridge Uniersity Press. Van der Brug, W Issue ownership and party choice. Electoral Studies, 23(2), Vliegenthart, R., & Walgrae, S "When the media matter for politics: Partisan moderators of the mass media's agenda-setting influence on parliament in Belgium". Party Politics 17 (3): Walgrae, S. & P. an Aelst The Contingency of the Mass Media s Political Agenda Setting Power. Towards a Preliminary Theory. J Commun 56 (2): Walgrae, S., & De Swert, K. (2004). The making of the (issues of the) Vlaams Blok. Political Communication, 21(4),

24 Wlezien, Christopher, Mark N. Franklin and Daniel Twiggs "Economic Perceptions and Vote Choice: Disentangling the Endogeneity" Political Behaior 67: Zaller, J The Myth of Massie Media Impact Reied: New Support for a Discredited Idea, in Mutz, Sniderman and Brody, eds, Political Persuasion and Attitude Change (Ann Arbor: Uniersity of Michigan Press,, pp

25 Appendix Issue category Economy Immigration Agriculture Law and order, crime Business, industry and consumer issues Defence Ciil liberties State church Education Culture, sports Energy Traffic, transport, infrastructure Research, technology, communications Foreign affairs EU Greenland and the Faroe Islands Central local relations, regional/local policy Politics, general The Monarchy Planning, resources, ciil preparedness Health Labour market and unemployment Enironment Social welfare and family Housing Public sector Danish foreign aid Table A1. Issue categories and issue ownership Bourgeois ownership 24 Unowned issues Social democratic ownership

26 25

27 Predicted change in ote intention Predicted change in ote intention 1,0 0,8 0,6 0,4 0,2 0,0-0,2-0,4-0,6-0,8-1,0 Owned issues Change in news saliency Fig A1. Predicted change in ote intention (bloc support) when news saliency changes for owned issues. 1,0 0,8 0,6 0,4 0,2 0,0-0,2-0,4-0,6-0,8-1,0 Unowned issues Change in news saliency Fig A2. Predicted change in ote intention (bloc support) when news saliency changes for unowned issues. 26

28 Predicted change in ote intention 1,0 0,8 0,6 0,4 0,2 0,0-0,2-0,4-0,6-0,8-1,0 Opponent owned issues Change in news saliency Fig A3. Predicted change in ote intention (bloc support) when news saliency changes for opponent owned issues. 27