THE PUBLIC OR PARTIES IN THE MEDIA?
A STUDY OF PUBLIC, PARTY AND MEDIA ISSUE AGENDAS IN FIVE DANISH ELECTION CAMPAIGNS
DAVID NICOLAS HOPMANN
CHRISTIAN ELMELUND-PR^STEK^R
ERIK ALB^K
CLAES DE VREESE
oo
Abstract
As the media plays a central role in the way modern democracies function, it is important to study whether the media engages in a top-down or a bottom-up mode of covering election news. The article studies whether the media agenda is congruent with the parties' and the public's agenda by analysing the three agendas in five national Danish election campaigns. Theoretically, increased professionalisation of politicians' efforts to influence media coverage suggests convergence between the media and the party agendas, while increased commercialisation of the media suggests convergence between the media and the public agendas. However, since both the professionalisation of the parties and the commercialisation of the media are ongoing processes, convergence between all three agendas may be expected. Results show that the media agenda is slightly more similar to the agenda of the parties, but in general the media seems to be rather good at balancing their obligations to represent the issues of the elite on the one hand and to give voice to public concerns on the other. During the past two decades no convergence between the agendas is found, i.e. the interaction of the three different agendas is rather stable.
David Nicolas Hopmann (dnh@sam.sdu.dk), Christian Elmelund-Pr^stek^r, and Erik Alb^k are researchers in the Centre for Journalism, Department of Political Science, University of Southern Denmark.
Claes de Vreese is Professor in Amsterdam School of Communications Research, University of Amsterdam; e-mail: C.H.deVreese@uva.nl.
a a
ro
O
Cfi O O rN
£
I 0 £
II) G C

While it is safe to state that modern democratic societies can hardly be imagined without some form of mass media, it is anything but given what role the media should play in a democracy (McQuail 2005; Street 2001; Strömbäck 2004). Different models of democracy call for different media roles. If one adheres to Schumpeterian competitive democracy, the media should provide, in a fair, balanced, accurate and comprehensive manner, information to the voters on "political elites, both ... what they have done, what they promise to do, and whether they have done what they promised when elected" (Strömbäck 2005, 339; see also Asp 2006). If, on the other hand, one adheres to a participatory or a deliberative model of democracy the news media is expected to "let ordinary people set the agenda" rather than "focus on the real actors - the political elites" (Strömbäck 2005, 342), i.e. the media should voice the public's concerns in such a manner that they can be resolved politically (Hoffmann-Riem 2003).
But what in fact determines the media's coverage of politics? The present study focuses on the issue agendas of the public, the media and the political parties in five Danish national elections to find out whether, in times of an election campaign, the media reflects the public (bottom-up coverage) or the party (top-down coverage) agenda - or both. Substantial evidence suggests that the public and the media agendas are rather similar (Dearing and Rogers 1996), while other studies have found a mixed picture with respect to the similarity of party and media agendas (Ridout and Mellen 2007). Few studies, however, investigate all three agendas at the same time (e.g., Asp 1983; Soroka 2002b). Furthermore, most studies cover a limited time period (typically one election campaign).
Inspired by this literature, the present article breaks new ground in agendasetting analysis in two important ways by studying all three agendas in the same study and by analysing changes over time. Thus, we investigate whether the media primarily reflects the issue concerns of the public or the parties. Further, we study whether the congruence between the media agenda and the agendas of the public and the parties have changed over time.
Empirically, we draw on data covering five national elections from 1994 to 2007 in Denmark. We include three datasets covering: television news during election campaigns; several channels of party election campaign communication; and surveys of what the general public sees as the country's most important political J problems. In the following section a number of hypotheses on the similarities of the three agendas and possible convergence over time are developed. This is followed by a presentation of our case and the design and methods applied. Next, the empirical findings are reported, and the results are summarised and discussed in the final section of the paper.
^	Theoretical Framework and Hypotheses
^	Drawing on exiting agenda-setting and political communication literature we
Q put forward a number of hypotheses on the connection between the media agenda on the one hand and the public and the party agendas on the other and how this connection may have changed over time.
Driven by Political Elites
Similar to most western democracies, the case under study, Denmark, is a representative democracy that only rarely asks its citizens to be actively involved
0 fi
in political decisions (Ismayr 2003; Petersson 2000). That is, the public is asked to elect a number of parliamentarians who then decide who forms the government; 00 referenda are only held rarely and most often in relation to European integration. ^ Moreover, extant research has shown that the media tends to focus on the powerful stakeholders in society (Cook 2006). In their study on agenda congruence in the US Ridout and Mellen (2007, 45) hypothesise that "coverage of campaigns tends to be reactive" as journalists simply report what is happening. Taken together, the media can be expected to focus on the political elites, and thus our first hypothesis suggests "catering to the political elites": The media agenda is similar to the agenda of the parties (H1).
However, we can also assume changes over time. Politicians have professionalised their approach toward the media and have become mediatised in the sense that they increasingly adhere to a media logic when communicating to the electorate (Blumler and Kavanagh 1999; Mazzoleni and Schulz 1999; Neveu 2002; Strömbäck 2008). The increased professionalisation and mediatisation of political parties is empirically well documented across most of Western Europe (Farrell and Webb 2000; Norris 2002).
It is also true for Denmark: even though political advertising is prohibited on television, the campaign budgets of the Danish political parties are increasing rapidly, and the parties hire still more campaign, news media and communication experts (Elmelund-Prastek^r and Hopmann 2008; Jonsson 2006). Danish parties have experienced a dramatic professionalisation since the mid-1990s, in particular after the turn of the millennium (Andersen and Pedersen 1999; Bille, Elklit and Jakobsen 1992; Jonsson 2006): in 1995 the parliament decided to almost double the annual direct public economic support for political parties, which led to an increase in the number of professionals employed by the parties. In the late 1990s, the major parties hired spin doctors, and strategic communication became an issue of interest to both academics and pundits in the 2001 election (Elmelund-Prastek^r and Hopmann 2008; Jonsson and Larsen 2002; Sarup 2004). As political parties increasingly hire journalists and media experts in order to understand the logic of the news media and to set their political solutions on the media agenda, the party and the media agendas can be expected to converge over the years. In sum, we can formulate this second hypothesis as "increasing catering to the political elites": The media's issue coverage is increasingly closer to the party agenda (H1a).
Driven by Public Demand
The broadcasting media, which is the case in our study, is operating in an environment shaped by many other factors than the political system. In most European countries, broadcasting started as public service: state-owned or heavily stateregulated (Starkey 2007). As, over the years, the media markets were gradually liberalised, more and more private, commercially oriented actors emerged (Machill 1999; Mazzoleni and Schulz 1999; Pfetsch 1996).
Denmark is no different: until 1988 the public service broadcaster (Danmarks Radio, DR) enjoyed a monopoly over the airwaves. In 1986, the parliament had decided to establish a new, commercially run alternative named TV2/Danmark (Powers, Kristjansdottir and Sutton 1994). Until 2004, this channel received some public license fees but was always mainly financed by commercial advertisements.
At first, the news bulletins on commercial television (TV2) were no great success. The popularity, however, changed after only a few years. In terms of viewers, commercial television's news bulletins have been the most watched ever since the beginning of the 1990s, although the lead over the public service (DR) news bulletins is not great. In 2006, public service television implemented several changes to their news coverage, announcing that these were driven by market logic and a need to increase its audience share (Claudi, S0lund and Thorsen 2006; Holm, Svith and Kartveit 2008). Thus, the time period covered in this study saw competition between the public service and commercial news bulletins, continuously spurring a process toward finding new ways to gain an edge and increase audience shares. This process affected both the public service and commercial television news bulletins (Hjarvard 2006).
This commercialisation and increasing competition of the media may have consequences for media content. Generally speaking, "a profit-seeking model of the media suggests that campaign news is not defined by the intrinsic importance of campaign activities but by what attracts large numbers of viewers" (Ridout and Mellen 2007, 45). Increased commercialisation of the media markets and the new commercial competitors to public service broadcasting therefore lead to a "demand market, whereby the assumed wishes and desires of the public have become more decisive for what the media select and provide" (Brants and van Praag 2006, 30). In the words of Swedish scholar Djerf-Pierre (2000), televised political journalism (at least in Sweden) no longer supplies the audience with news it "needs." Instead, it provides what the public "wants."
Although it has not automatically led to an "Americanised" modus of news journalism, at least in countries with strong public service broadcasting (Blumler and Gurevitch 2001; Mazzoleni and Schulz 1999; Strömbäck 2007) - for example, the trend towards strategy and spin framed news journalism has been rather mute in most West European countries (Binderkrantz and Green-Pedersen 2008; Brants and van Praag 2006; Esser and D'Angelo 2006; Hopmann, Alb^k and de Vreese 2009) - increased competition and commercialisation supposedly still have had a significant impact on some aspects of media content (Hjarvard 1999, 2006). Both major Danish broadcasters emphasise that they today, as different from earlier periods, see their role as a platform for ordinary citizens to gain public attention (Hjarvard 1999; Holm 2007; Svith 2007). For example, it is DR's goal to "define news as what is on the public agenda [and] news stories are chosen so they can create action and involvement amongst the audience" (Holm 2007, 80). In the words of TV2, the goal is to "inform the political system what is happening amongst the Danes rather than telling the Danes what is happening amongst politicians" (Hjarvard
£ 3
a
2006, 125). To achieve such goals, both Danish broadcasters use opinion polls to guide their selection of which issues to have on their agenda (van der Brugge and Voss 2003, 132; Holm 2007, 80; de Vreese 2003). At the same time it is an explicit goal for both broadcasters to reach as many Danes as possible and not only certain groups of the population (Holm 2007; Svith 2007).
Furthermore, extant research on agenda-setting has also shown that the media has a large impact on the public agenda (McCombs and Shaw 1972; Weaver 1996;
^ Weaver, McCombs and Shaw 2004). That is, we assume that not only does the me-
dia aim at covering issues deemed necessary by the public, the media clearly also impacts on which issues the public deems important. On this basis we formulate LH our second major hypothesis as "catering to the public": The media's issue coverage is similar to the public agenda (H2).
Since the commercialisation has increased in the period covered by this study (i.e., the early 1990s), we also expect that the media agenda has become more similar to the public agenda over time. Therefore, we formulate an additional, dynamic hypothesis as "increasing catering to the public": The media's issue coverage is increasingly closer to the public agenda (H2a).
Driven by Interaction
In a similar analysis on the media, the party and the public agendas in the US, Dalton et al. (1998, 465) suggest that agendas are a "result from the interaction of social actors; each actor is constrained by the others and by the flow of actual political events." Rather than assuming that the media's efforts to cater to the public or that the parties' efforts to influence the media are dominating aspects, an interaction model assumes that all these efforts take place and shape the different agendas simultaneously (cf. Soroka 2002b). Moreover, one can also assume that not only is the media agenda similar to the party and the public agendas, but also that the party and the public agendas are congruent. Hence, we formulate this final hypothesis as an "interaction model": All three agendas (the media, the party and the public agendas) are similar to one another (H3).
Data and Method
To test our hypotheses, we draw on three different datasets covering five consecutive national elections from 1994 to 2007. Since television is the public's most frequently used source of information (Lund 2001; Mazzoleni and Schulz 1999), we begin by a content analysis of the major news bulletins on both major Danish news broadcasters, public service (DR) and commercial (TV2). The analysis covers three weeks prior to each election day, the typical length of Danish election campaigns. For each news story, the dominant issue was coded, applying 29 different issue codes.1
Second, a content analysis of the political parties' campaign communication in the election periods studied was conducted. There is no precise guidance in the literature on how party data can be pooled, and operationalisations differ widely (Soroka 2002b, 271; Walgrave and van Aelst 2006). However, we try to gauge the party agendas in a broader manner than seen in earlier studies (cf. Ridout and Mellen 2007). The empirical material of our study includes the parties' election manifestos, their ads and letters-to-the-editor in the five biggest national newspapers,2 the parties' televised so-called "presentation programmes" (shown on DR) and two televised party leader debates (on DR and TV2). For the "small" units of analysis (i.e., the letters-to-the-editor and the newspaper ads), one issue per item was coded - and the frequency of the issues is calculated for each channel of communication. For the "large" units of analysis (i.e., the televised programmes and the manifestos), one issue per message is coded; a "message" can consist of several sentences or just one depending on how much the speaker elaborates on a specific
point. Hence, messages are semantic entities delimited by a change of meaning of the text or speech.3 Again, the total salience of each issue is computed separately for each source. Here, a coding scheme with 25 different issue codes was used.4 Finally, the agendas of the five different sources are pooled in the following steps: first, for each party an agenda was calculated for each channel of communication; second, a 'summery' agenda was computed for each party at each election; third, theses party agendas were pooled to one aggregated party agenda for each election. To account for the fact that Danish parties have different electoral sizes we weighted the individual party agendas by parties' share of votes in the election campaign investigated. This way we have one (weighted) party agenda for each of the five election campaigns studied.
Finally, following previous research (Asp 1983; Dalton et al. 1998; Soroka 2002b) we draw on representative survey data from the Danish Election Study to map the public agenda in the different elections. In the aftermath of each national election, a representative sample of the Danish population is asked "which problems do you think are the most important ones politicians should take care of?" Each respondent could give several answers (for more details, see Andersen 2008).5 The answers given were originally grouped into nine issue categories.
As the coding of issues is different in the three datasets - and to further improve the reliability of the data - all issues were grouped in the same nine categories as used by the Danish Election Study. These categories are: employment; taxes; other economic issues; environment; immigration; welfare; EU & foreign affairs; other or unclear issues. In addition, in both the media and the party data we find attention to non-substantive campaign issues such as the presentation of opinion polls or campaign trail stories (cf. D'Angelo, Calderone and Territola 2005; Esser and D'Angelo 2006). As this category is not used (and most likely is not relevant) in the data for the public agenda, the non-substantive issue category is excluded from the analysis.
In a first analytic step we describe the media, the party and the public agendas during the five elections studied. This is done on the basis of tables 1, 2 and 3 in the following section. Next, we turn to a test of our hypotheses. Here the relative sizes of the issue categories are compared election by election. This comparison is done by utilising an indexing method similar to several previous studies (e.g., Asp 1983, 2006; Brandenburg 2005; Ridout and Mellen 2007). The computed index values tell Q us how large proportions that need to be reallocated within one agenda in order to make it completely similar to another agenda. The larger the index value, the more dissimilar the two analysed agendas are. The index of dissimilarity ranges ^ from 0 (= perfect overlap) and to 100 (= no overlap).6
Findings
®	Starting with investigating the media agenda and the changes within the dif-
ferent issue categories, we see a rather stable picture (see Table 1). Some minor changes from one election campaign to the next are found, e.g. the somewhat larger focus on foreign affairs in 2001 following the attacks on the World Trade Centre in New York City, but overall the issue attention is rather stable. If one compares the different issue categories, it is evident that the welfare issue dominates the media
£ 3
a
c >
agenda followed by foreign affairs. During the 1990s more attention was invested in economic and employment issues than in the 2000s, which can be explained by the economic challenges Denmark was facing at that time.
Table 1: Media Agenda in Election Campaigns 1994-2007 (percentages of news stories)
	1994	1998	2001	2005	2007
Employment	3	0	1	7	1
Economy	16	12	9	7	4
Taxes	2	3	2	3	6
Environment	8	14	8	10	19
Welfare	38	35	31	45	35
Immigration	4	7	13	5	7
EU & foreign affairs	25	22	32	19	20
Other	3	6	5	4	8
Note: The percentages shown are the average for both broadcasters included in the study.
Second, the party agenda is dominated by welfare issues, even more so than is the media agenda (Table 2). Also the employment and economy issue categories follow the same trend as found in the media data. The party agenda first and foremost differs from the media agenda by having a higher salience of immigration
-	and a lower salience of the EU and foreign affairs. The party agenda, in other words, has been more domestically oriented - with a special focus on immigration
-	than the media agenda has been.
Table 2: Party Agenda in Election Campaigns 1994-2007 (percentages)
	1994	1998	2001	2005	2007
Employment	16	4	7	9	4
Economy	21	13	2	7	7
Taxes	4	5	11	7	6
Environment	7	6	4	5	9
Welfare	33	49	46	49	34
Immigration	10	12	18	16	12
EU & foreign affairs	6	6	7	5	10
Other	4	5	5	3	18
Note: See data and method section for a detailed explanation of the computation of the party agenda.
Third, the public agenda has also been dominated by welfare issues (Table 3),7 as well as immigration, which has been heavily rising through the 1990s. In accordance with the parties but in contrast to the media, the public is not very concerned with the EU and foreign affairs - this issue category is among the lowest ranked categories on the public agenda.
Table 3: Public Agenda in Election Campaigns 1994-2007 (percentages of answers by voters)
	1994	1998	2001	2005	2007
Employment	24	7	3	16	1
Economy	15	7	4	3	3
Taxes	2	5	4	5	3
Environment	8	9	3	4	7
Welfare	38	47	55	53	60
Immigration	8	14	23	13	14
EU & foreign affairs	3	5	6	3	2
Other	2	6	4	3	5
Note: Respondents could give several answers.
Whose Agenda Are the Media Covering?
Table 4 shows similarities and dissimilarities between the media agenda, the party and the public agendas, respectively, during the election campaigns from 1994 to 2007.
The first hypothesis (H1), "Catering to the political elites," states that the media and the party agendas are congruent. As indicated in the above description of the different agendas, we do see a substantial difference between the agendas with respect to the issue of the EU and foreign affairs: in every included election campaign the media focuses more on this issue than the parties. At the same time, the media focuses somewhat less on e.g. employment and immigration than do the political parties.
That said, looking at the overall differences, we find similarities, not dissimilarities: on average, the index of dissimilarity equals 27. That is, just a bit more than one quarter of the issue attention on one agenda has to be relocated in order to y render the two agendas perfectly similar. Only a single year appears to be a minor outlier, namely the 2001 election campaign. Presumably, this difference is caused by the media's attention to the September 11 attacks which occurred a few weeks prior to the election campaign. In sum, partial support for H1a is found, the media agenda is rather similar to the party agenda - the question remains, however, whether it is more similar to the party than to the public agenda.8 We return to this question in the Discussion.
Our next hypothesis (H1a), "Increasing catering to the political elites," is not supported: the media and the party agendas do not converge over time. Comparing the index of dissimilarity for 1994 and 2007, we find a somewhat increasing similar-jß ity, but given the inevitable measurement error we are not confident in speaking of a substantial trend of conversion. Again the campaign of 2001 stands out, which reflects the special attention of the media in this election discussed above.
Table 4 also allows us to test our second main hypothesis (H2), "Catering to the public." Again, we see that the media focuses more on foreign affairs and somewhat less on immigration than does the average voter. In fact, the overall figures for the agenda congruence of the media and the public are very similar to the figures comparing the media and the party agendas: on average, 31 percent of the issue
0
fi ¥
I
C >
Table 4: Congruence of Media Agenda with Party and Public Agendas in Election Campaigns 1994-2007	^
	1994	1998	2001	2005	2007
Employment	-8	-3	4	-7	3
Economy	6	6	-2	4	4
Taxes	2	0	7	2	-2
Environment	-1	-3	1	1	2
Welfare	-5	2	-9	-4	-26
Immigration	2	-2	-5	3	-2
EU & foreign affairs	3	1	1	2	8
Other	2	-1	1	0	13
Dissimilarity	14	9	14	12	30
Note: The sign indicates whether an issue category is more (+) or less (-) salient on the party agenda than on the public agenda. The overall dissimilarity is computed by adding all percentage point differences (absolute values) divided by 2 (since we are dealing with two agendas, see Asp 1983, 352f.).
attention needs to be reallocated within the one agenda to make it perfectly similar to the other. Once again the 2001 campaign appears to be an outlier as the public agenda was not greatly concerned about foreign affairs.9 In sum, our results partially support hypothesis H2a: the media agenda is close to the public agenda.10
The next hypothesis (H2a), "Increasing catering to the public," finds no empirical support. The results shown in Table 4 do not indicate a trend of increasing similarity between the media and the public agendas. Despite the increasing commercialisation of the media market we, if anything, find increasing dissimilarity over the years. Even though the commercial broadcaster, TV2, presumably is keener to cater to the public than the public service broadcaster, DR1, both follow the same pattern over the years (data not shown).
A Result of Interactions?
The final hypothesis (H3), "Interaction model," assumes that the media, the party and the public agendas are the results of an interaction between the actors in these three spheres. That is, it may very well be that the media is concerned with catering to the public, but at the same time follows the political elites closely. With the available data, we unfortunately cannot determine the exact causal mechanisms behind the shaping of the three agendas (cf. Soroka 2002b). Nevertheless, similar to Dalton et al. (1998) we find indications of mutual influence of the three agendas.
In every election campaign both the party and the public agendas are rather similar to the media agenda which indicates that the media does not have an autonomous agenda-setting power. The one instance with slightly larger differences, the 2001 election, was not caused by independent actions of the media but by external events.
Moreover, as shown in table 5, there is high congruence between the party and the public agenda. Given modern campaigning techniques this finding hardly comes as a surprise: "the candidates are systematically monitoring public opinion and attempting to persuade the public" (Dalton et al. 1998, 476) and therefore parties are "adjusting to their voters" (Asp 1983, 351).
*** Table 5 about here *** - TABLE NOT INCLUDED IN ORIGINAL TEXT!
Similar to the findings presented by Asp (1983) and Dalton et al. (1998), our findings therefore indicate that the media, the party and the public agendas are shaped by mutually influencing one another. As mentioned above, we know that the media undertakes opinion polls to gauge the public's interests, and previous research has shown that the media can have an influence on the public agenda and the party agenda. In sum, we concur with Dalton et al. (1998, 476) when they state that the media does "not play the dominant agenda-setting role portrayed in some political communications literature." In other words, the media seems, on the one hand, to voice issues emphasised by the public and, on the other hand, to give a platform to the issues deemed important by the political parties, at least during election campaigns.
Discussion
In this article we analyse the similarities and dissimilarities between the issue attention on the media, the party and the public agendas during five national election campaigns. We find that the media agenda is rather similar to both the party and the public agendas. However, the media and the party agendas are slightly more similar than are the media and the public agendas; the media caters to the political elites a little more than to the public. The differences are, however, rather small and both H1 ("Catering to the political elites") and H2 ("Catering to the public") are empirically supported. This conclusion is corroborated by the findings supporting our third and final hypothesis, which suggests that the three agendas are results of an interaction between all actors and not simply one dominant source. Besides the large similarities between the media agenda and both the party and the public agendas, we find even more overlap between the latter two.
By contrast we find no empirical support for our hypotheses on trends (H1a and H2a): the media agenda is converging with neither the party nor the public agenda in the studied period of time. In other words, with respect to issue coverage it is y too simplistic to assume that increased commercialisation and competition more or less automatically leads to a "demand market" or that the professionalisation of the parties automatically enables the parties to control the media agenda.
At a more general level our results suggest that the Danish media in fact fulfils more "democratic duties" at the same time. The media has never exclusively (at least since 1994) covered the issue concerns of the parties, neither does it exclusively voice the issues seen as the most important by the general public. Rather it seems as if the Danish media has in fact succeeded in balancing a representative and a deliberative democratic ideal fairly equally in the past 13 years of election campaign coverage. Different ideals of democracy - and thus different ideals of journalism jß - continue to compete with one another. The balance between the two, found in our study, strikes us as a workable and reasonable pragmatic compromise.
The question remains, whether our results can be generalised to other settings: during election times political parties presumable have a larger influence on the media than in routine times. Walgrave and van Aelst (2006) list three reasons why political parties may be more successful in building the media agenda during election campaigns: first, during election campaigns political parties and their QQ candidates "are [even more] vigorously trying to influence the public agenda"
0
fi ¥
I
C >
(Walgrave and van Aelst 2006, 97). Second, since the media pays more attention to politics during campaigns, "there is plenty of room for parties and candidates to get their substantial message across" (Walgrave and van Aelst 2006, 98). Finally, both the media and the parties pay more attention to fairness and balance in news coverage during election campaigns, which in turn limits the media's power to shape the media agenda single-handedly (Walgrave and van Aelst 2006, 98). In other words, one may assume that the media is more prone to cater to the public than to the political elites in non-election times. Thus, it is relevant to study whether the results are valid in non-election times; extant research finds that major differences may exist between election campaign and routine times (Binderkrantz and Green-Pedersen 2008; Walgrave and van Aelst 2006).
Obviously, this study is far from conclusive. Future research on the subject is needed. First, in an effort to determine the exact causal mechanism shaping the different agendas research should aim at collecting public opinion data prior to an election campaign. Second, more detailed data would make it possible to analyse, who is following whom in the course of an election, thus treating elections as processes and not as events (e.g., Brandenburg 2002). Such an approach would also help to firmly establish causal influences from one agenda to another (cf. Soroka 2002a, 2002b). Third, since we are dealing with five time points only, it is unquestionably important to collect longitudinal data to establish more firmly whether or not the different agendas are in fact converging.
Nevertheless, our findings draw a general picture of the media, the party and the public agendas during the latest two decades, and hopefully the study can be of inspiration to future research on issue agendas.
Notes:
1.	The numbers of analysed news stories are 247, 211, 252, 253 and 218 for the election campaigns in 1994, 1998, 2001, 2005 and 2007. The reliability score for the issue coding was .81 (Krippendorff's alpha). For more details see Hopmann et al 2009.
2.	The morning papers Politiken, Jyllands-Posten, Berlingske Tidende and the tabloids EkstraBladet and B.T.
3.	The numbers of units analysed is: letters: 586; ads: 1,310; party presentation programmes: 3,873 messages; party leader debates: 2,708 messages; manifestos: 4,051 messages.
4.	This measure yielded a inter-coder reliability score of .70 (Krippendorff's alpha).
5.	The data covering the 2005 and 2007 elections are from surveys conducted during the election campaigns, the data covering the campaigns in 1994, 1998 and 2001 were collected after the election campaigns.
6.	This index has been described as the "Duncan's index of dissimilarity" (Brandenburg 2005) and the "matching index" (Asp 1983).
7.	Different from earlier, in 2007 respondents gave many unspecified answers relating to welfare rendering it difficult to decide when they gave one or more answers. Therefore, the proportion of welfare answers in 2007 may be slightly overestimated (Andersen 2009).
8.	Following Asp (1983), one could suggest that it is appropriate to differentiate between media outlets according to their political stances. The reasoning behind this suggestion is that the conservative media may be closer to right-wing parties than to left-wing parties and vice-versa. Previous research has suggested that Danish public-service broadcasting has a left-wing bias and commercial broadcasting a right-wing bias (Hallin & Mancini 2004, 170; Hjarvard 1999, 71). We therefore compared whether public broadcasting and commercial broadcasting substantially
£ 3 ft
(I) O C
(B
fN 00
differed in their coverage of the two major blocs in Danish politics: there are only minor differences between the two broadcasters (data not shown). Hence, with Asp (1983, 340) we can conclude that the "professional news criteria seem to be more important than the partisan news criteria."
9.	The slightly larger difference for 2007 can most likely be explained by the overestimated public attention to welfare (60 percent), see note 7.
10.	Following Soroka's (2002b) typology on which issues on the media agenda tend to influence the public agenda the most, most differences between the media and the public agendas are little surprising: on the one hand, the public experiences problems with employment and welfare service first-hand, they are obtrusive issues and thus limit the media's agenda setting power; on the other hand, even though governmental issues such as foreign affairs are not tangible and therefore unobtrusive, they are nevertheless "undramatic" for the public. Unobtrusive and sensational issues such as the environment and immigration seem to follow the same pattern on both the media and the public agenda. One could therefore argue that the media indeed seems to have a substantial agenda-setting power at least with respect to some issues. That said, the differences shown in table 4 are not large and rather stable. While we do not deny that the media is an important source of information for the public, we agree with Dalton et al. (1998) when they conclude that the media does not appear to be an autonomous actor as sometimes suggested: the small dissimilarities and rather stable similarities between the different agendas presented in this study seem to support their conclusion (cf. the following discussion).
References:
Andersen, Jorgen Goul. 2008. Et valg med paradokser: Opinionsklimaet og folketingsvalget 2007.
Tidsskriftet Politik 11, 3, 10-26. Andersen, Peter Werther, and Karina Pedersen. 1999. De danske partier og folketingsvalgkampen i
1998. Politica 31, 3, 297-312. Asp, Kent. 1983. The Struggle for the Agenda. Communication Research 10, 3, 333-55. Asp, Kent. 2006. Rättvisa nyhetsmedier - Partiskheten under2006 ärs medievalrörelse. Göteborg: JMG
Arbetsrapport nr. 42, Göteborgs Universitet. Bille, Lars, Jorgen Elklit, and Mikael Jakobsen. 1992. Denmark: The 1990 campaign. In S. Bowler and
D. Farrell (eds.), Electoral Strategies and Political Marketing, 63-81. New York: St. Martin's Press. Binderkrantz, Anne, and Christoffer Green-Pedersen. 2008. The Framing of Politics in the News Media: Policy or Processes in Focus? Paper presented at the XV NOPSA conference, University of Tromso, August 6-8.
Blumler, Jay G., and Michael Gurevitch. 2001. 'Americanization' Reconsidered: U.K.-U.S. Campaign Communication Comparisons Across Time. In W. L. Bennett and R. M. Entman (eds.), Mediated Politics - Communications in the Future of Democracy, 380-403. Cambridge: Cambridge University Press.
Blumler, Jay G., and Dennis Kavanagh. 1999. The Third Age of Political Communication: Influences
and Features. Political Communication 16, 3, 209-230. Brandenburg, Heinz. 2002. Who Follows Whom? The Impact of Parties on Media Agenda Formation in the 1997 British General Election Campaign. Harvard International Journal of Press/Politics 7, 3, 34-54.
Brandenburg, Heinz. 2005. Political Bias in the Irish Media: A Quantitative Study of Campaign
Coverage during the 2002 General Election. Irish Political Studies 20, 3, 297-322. Brants, Kees, and Philip van Praag. 2006. Signs of Media Logic: Half a Century of Political
Communication in the Netherlands. Javnost—The Public 13, 1, 25-40. van der Brugge, Jimmy, and Henning Voss. 2003. Mediernes dagsorden. In J. G. Andersen and O. Borre (eds.), Politisk forandring. Vxrdipolitik og nye skillelinier ved folketingsvalget 2001, 119-134. Ärhus: Systime.
Claudi, Nadia, Sune Solund, and Nils Thorsen. 2006. Slaget om nyhederne. Politiken, Kultur, 1. Cook, Timothy. 2006. The News Media as a Political Institution: Looking Backward and Looking
Forward. Political Communication 23, 159-171. Dalton, Russell J., et al. 1998. A Test of Media-Centered Agenda Setting: Newspaper Content and Public Interests in a Presidential Election. Political Communication 15, 4, 463-481.
D'Angelo, Paul, Matthew Calderone, and Anthony Territola. 2005. Strategy and Issue Framing: An Exploratory Analysis of Topics and Frames in Campaign 2004 Print News. Atlantic Journal of Communication 13, 4, 199-219.	00
Dearing, James W., and Everett M. Rogers. 1996. Agenda-setting. Thousand Oaks: Sage.
Djerf-Pierre, Monika. 2000. Squaring the Circle: Public Service and Commercial News on Swedish Television 1956-99. Journalism Studies 1, 2, 239-260.
Elmelund-Prssteksr, Christian, and David Nicolas Hopmann. 2008. Teflon og bullshit? Ti myter om politik, journalistikog valgkampe. Odense: University Press of Southern Denmark.
Esser, Frank, and Paul D'Angelo. 2006. Framing the Press and Publicity Process in U.S., British, and German General Election Campaigns. Press/Politics 11, 3, 44-66.
Farrell, David, and Paul Webb. 2000. Political Parties as Campaign Organizations. In R. Dalton and M. Wattenberg (eds.), Parties without Partisans. Political Change in Advanced Industrial Democracies, 102-129. Oxford: Oxford University Press.
Hallin, Daniel C. and Paolo Mancini. 2004. Comparing Media Systems: Three Models of Media and Politics. Cambridge: Cambridge University Press.
Hjarvard, Stig. 2006. Tv-nyheder. In S. Hjarvard (ed.), Dansk tv's historie, 105-44. Kobenhavn: Samfundslitteratur.
Hjarvard, Stig. 1999. TV-nyhederikonkurrence. Frederiksberg: Samfundslitteratur.
Hoffmann-Riem, Wolfgang. 2003. Mediedemokratie zwischen normativer Vision und normativem Albtraum. In W. Donsbach and O. Jandura (eds.), Chancen und Gefahren der Mediendemokratie, 28-39. Konstanz: UVK Verlagsgesellschaft.
Holm, Hans-Henrik. 2007. Public service journalistik? Vssentlighed, publikumsinteresse og demokratisk handling. In M. Carstensen, F. Svith and P. Mouritsen (eds.), DR og TV2 - i folkets tjeneste?, 69-91. Ärhus: Ajour.
Holm, Hans-Henrik, Flemming Svith and Kate Kartveit eds. 2008. Nar nyheder bliver til - pä DR og TV2. Ärhus: Ajour.
Hopmann, David Nicolas, Erik Albsk and Claes H. de Vreese. 2009. Incumbency Bonus in Election News Coverage Explained: The Logics of the Political System and the Media Market. Paper presented at the International Communication Association 59th Annual Conference, Chicago May 21 - 25.
Ismayr, Wolfgang ed. 2003. Die politischen Systeme Westeuropas. Opladen: Leske + Budrich.
Jonsson, Rasmus. 2006. Den professionelle politiske kommunikation krsver nye journalistiske perspektiver. In P. Bro, R. Jonsson and O. Larsen (eds.), Politiskjournalistikog kommunikation: Forandringer i forholdet mellem politik og medier, 43-66. Frederiksberg: Samfundslitteratur.
Jonsson, Rasmus and Ole Larsen. 2002. Professionelpolitisk kommunikation. Et studie af 20 dages valgkamp. Kobenhavn: Akademisk Forlag.
Lund, Anker Brink. 2001. Forskning om medier og demokrati. Kobenhavn: Danske Dagblades Forening.
Machill, Marcel. 1999. The Effect of the Commercialization of Swedish Television on Journalistic Culture. Harvard International Journal of Press/Politics 4, 2, 103-111.
Mazzoleni, Gianpietro and Winfried Schulz. 1999. "Mediatization" of Politics: A Challenge for Democracy? Political Communication 16, 247-261.
McCombs, Maxwell and Donald Shaw. 1972. The Agenda-Setting Function of Mass Media. Public Opinion Quarterly 36, 2, 176-187.
McQuail, Denis. 2005. McQuail's Mass Communication Theory. Sage Publications Ltd.
Neveu, Erik. 2002. Four Generations of Political Journalism. In R. Kuhn and E. Neveu (eds.), Political Journalism: New Challenges, New Practices, 22-43. London: Routledge.
Norris, Pippa. 2002. Campaign Communications. In L. LeDuc, R. Niemi and P. Norris (eds.), Comparing Democracies 2. New Challenges in the Study of Elections and Voting, 127-47. London: Sage Publications.
Petersson, Olof. 2000. Nordiskpolitik. Göteborg: Norstedts Juridik AB.
Pfetsch, Barbara. 1996. Convergence Through Privatization? Changing Media Environments and Televised Politics in Germany. European Journal of Communication 11, 4, 427-451.
Powers, Angela, Hildur Kristjansdottir, and Hal Sutton. 1994. Competition in Danish Television News.
3 ft
O
(I)
o
C
(B
00
Journal of Media Economics 7, 4, 21-30. Ridout, Travis N., and Rob Mellen. 2007. Does the Media Agenda Reflect the Candidates' Agenda?
International Journal of Press/Politics 12, 2, 44-62. Sarup, Thomas Sondergaard. 2004. Spindoktori. Hvad er det, og findes det i Danmark? In M. F. Nielsen (ed.), Spin, selvfremstilling og samfund - public relations som reflekterende praksis, 39-74. Frederiksberg: Samfundslitteratur. Soroka, Stuart N. 2002a. Agenda-Setting Dynamics in Canada.Toronto: University of British Columbia Press.
Soroka, Stuart N. 2002b. Issue Attributes and Agenda-Setting by Media, the Public, and Policymakers
in Canada. International Journal of Public Opinion Research 14, 3, 264-285. Starkey, Guy. 2007. Balance and Bias in Journalism: Representation, Regulation and Democracy.
Hampshire: Palgrave Macmillan. Street, John. 2001. Mass Media, Politics and Democracy. Houndmills: Palgrave. Strömbäck, Jesper. 2004. Den medialiserade demokratin: Om journalistikens ideal, verklighetoch makt. Stockholm: SNS Förlag.
Strömbäck, Jesper. 2005. In Search of a Standard: Four Models of Democracy and Their Normative
Implications for Journalism. Journalism Studies 6, 3, 331-45. Strömbäck, Jesper. 2007. Marketplace of Ideas and Marketplace of Money. A Study of
Commercialism and the Swedish Election News Coverage in 1998 and 2002. Nordicom Review Jubilee Issue, 51-61.
Strömbäck, Jesper. 2008. Four Phases of Mediatization: An Analysis of the Mediatization of Politics.
International Journal of Press/Politics 13, 3, 228-246. Svith, Flemming. 2007. Public service per stedfortrsder - Journalistik i offentlighedens tjeneste. In
M. Carstensen, F. Svith and P. Mouritsen (eds.), DR og TV2 - i folkets tjeneste?, 93-110. Ärhus: Ajour. de Vreese, Claes H. 2003. Television Reporting of Second-Order Elections. Journalism Studies 4, 2, 183-98.
Walgrave, Stefaan, and Peter van Aelst. 2006. The Contingency of the Mass Media's Political Agenda
Setting Power: Toward a Preliminary Theory. Journal of Communication 56, 88-109. Weaver, David. 1996. Media Agenda-Setting and Elections: 1970s-1990s. In D. L. Paletz (ed.), Political Communication Research:Approaches, Studies, Assessments Volume II, 211-224. Norwood: Ablex Publishing.
Weaver, David, Maxwell McCombs, and Donald L. Shaw. 2004. Agenda-Setting Research: Issues, Attributes, and Influences. In L. L. Kaid (ed.), Handbook of Political Communication Research, 25781. London: Lawrence Erlbaum Associates.