The Icelandic media coverage of the constitutional assembly election

On November 27, 2010, the people of Iceland elected 25 individuals to the country’s constitutional assembly. As there were 522 candidates for the 25 seats in the assembly, the media were faced with a new dilemma, i.e. how to ensure impartiality and objectivity in their coverage of the candidates and the subject matter. The present study compares the media coverage of the constitutional assembly election to two other national elections; the general election in the spring of 2009 and the municipal election in the spring of 2010. All news stories in the 13 major print, broadcast and online news outlets in Iceland were coded two weeks prior to each election. The results indicate that the national media almost ignored the constitutional assembly election in comparison to the other elections. There were 632 news stories on the general election, 590 stories on the municipal election but only 165 stories on the constitutional assembly election. The lack of coverage of the candidates for the constitutional assembly seems to reveal that the traditional media, i.e. the print and broadcast media, and the online media did not know how to best serve and inform the public in the democratic process.


Introduction
On November 27, 2010, Icelandic voters had the unique opportunity to elect 25 of their own countrymen to a constitutional assembly whose role was to review the current constitution of Iceland and lay the foundation for a new one.Due to the unusual nature of the constitutional assembly election, it is the main objective of the present paper to examine the Icelandic news media's coverage of that particular elect ion.Furthermore, to gain a better understanding of how the coverage of the media might have differed from their regular election coverage, two other elections are studied, i.e. the general election in April of 2009 and the municipal election in May of 2010.
Voter turnout in the constitutional election was low.Only 35.95% of the eligible voters actually voted (Stjórnlagaráð 2010).In comparison, voter turnout was 85.1% for the general election in 2009 (Hagstofa Íslands 2010a).Three weeks before the constitutional assembly election, a poll by Market and Media Research (MMR) showed that 57.4% of the voters had not acquainted themselves with any of the 522 candidates (MMR 2010b).
During the weeks prior to election day, the national media, especially the Icelandic

Introduction
On November 27, 2010, Icelandic voters had the unique opportunity to elect 25 of their own countrymen to a constitutional assembly whose role was to review the current constitution of Iceland and lay the foundation for a new one.Due to the unusual nature of the constitutional assembly election, it is the main objective of the present paper to examine the Icelandic news media's coverage of that particular elect ion.Furthermore, to gain a better understanding of how the coverage of the media might have differed from their regular election coverage, two other elections are studied, i.e. the general election in April of 2009 and the municipal election in May of 2010.
Voter turnout in the constitutional election was low.Only 35.95% of the eligible voters actually voted (Stjórnlagaráð 2010).In comparison, voter turnout was 85.1% for the general election in 2009 (Hagstofa Íslands 2010a).Three weeks before the constitutional assembly election, a poll by Market and Media Research (MMR) showed that 57.4% of the voters had not acquainted themselves with any of the 522 candidates (MMR 2010b).
During the weeks prior to election day, the national media, especially the Icelandic

Introduction
On November 27, 2010, Icelandic voters had the unique opportunity to elect 25 of their own countrymen to a constitutional assembly whose role was to review the current constitution of Iceland and lay the foundation for a new one.Due to the unusual nature of the constitutional assembly election, it is the main objective of the present paper to examine the Icelandic news media's coverage of that particular elect ion.Furthermore, to gain a better understanding of how the coverage of the media might have differed from their regular election coverage, two other elections are studied, i.e. the general election in April of 2009 and the municipal election in May of 2010.
Voter turnout in the constitutional election was low.Only 35.95% of the eligible voters actually voted (Stjórnlagaráð 2010).In comparison, voter turnout was 85.1% for the general election in 2009 (Hagstofa Íslands 2010a).Three weeks before the constitutional assembly election, a poll by Market and Media Research (MMR) showed that 57.4% of the voters had not acquainted themselves with any of the 522 candidates (MMR 2010b).
During the weeks prior to election day, the national media, especially the Icelandic The Icelandic media coverage of the constitutional assembly election Guðbjörg Hildur Kolbeins, Ph.D. in mass communication National Broadcasting Service (RÚV), were harshly criticized by the candidates and others for not fulfilling, what many argued was its legal duty, to allow the candidates on the air, and to cover the election more extensively (Hersveinn 2010;Tryggvason 2010).
In retrospect, and keeping in mind what we know about the agenda-setting effect of the media, it is argued that the lack of news coverage manifested itself in low voter turnout due to such effects.Of course it is possible to say that the news media only reflected the public's attitude towards the election.Rogers and Dearing (2007), for example, have maintained that the relationship between the media agenda and the public agenda is bidirectional.But considering the importance of the issue and the fact that 80% of people rely on the mass media for news and public affairs, according to one foreign study (Yang & Stone 2003), one might have expected the media to play a larger role.

The advent of the constitutional assembly election
At the end of September 2008 and in the beginning of October that same year, the three largest banks in Iceland, i.e.Glitnir, Kaupþing and Landsbankinn, defaulted and crashed (Bragadóttir 2008;Mbl.is 2008;Rögnvaldsson 2008).What followed was a time of economic insecurity, political instability and social unrest.
Shortly afterwards, people took to the streets to protest.The main gathering place was Austurvöllur, the public square in front of Alþingi, the Icelandic parliament.The demonstrations, known for the participants' use of pots and pans to make noise, started out peacefully but escalated into violence by the end of the year (Bergþórsdóttir 2008).In January of 2009 things had gotten out of hand with riots breaking out on regular basis, and on January 26 the coalition government of the Independence Party and the Social Democratic Alliance, led by Geir H. Haarde, resigned (DV.is 2009;Norden 2009).An interim minority government, formed by the Social Democratic Alliance and the Left-Green Movement and supported by the Progressive Party, took over and continued its alliance (in a majority position) after the general election in April 2009.
During the mass demonstrations, the issue of a new constitution had been raised (Torfason n.d.)Also, the left-wing government felt early on that it was a priority to review the constitution.The constitution dates from 1944 and was to a large extent based on the Danish constitution (Tómasson, Thorarensen, Kristinsson & Stefánsson 2005).
On June 16, 2010, Alþingi passed a bill on the constitutional assembly election that determined how the assembly would be formed.According to the law on the constitutional assembly, it should be made up of 25 to 31 individuals who were democratically elected by Icelandic voters.The number would be greater than 25 if either women or men were drastically underrepresented (Lög um stjórnlagaþing 2010).When the deadline to announce one's candidacy for the constitutional assembly had passed, it turned out that there were 522 candidates for those 25 to 31 seats.
The election was held on November 27, 2010, and the majority of those who were

368
Fraeðigreinar elected were either already quite well-known among the population for their work or media personalities.Only one or two of the elected candidates were practically unknown before running for a seat in the constitutional assembly.
The large number of candidates complicated the election and made it difficult to adhere to normal election procedures.Consequently, the Supreme Court of Iceland ruled on January 25, 2011, that the election was invalid on technical grounds.For example, the voting booths were open and did not ensure privacy (Mbl.is 2011).
Two months later, on March 24, 2011, Alþingi decided to appoint those who were elected to a constitutional council instead of repeating the election (Þingsályktun um skipun stjórnlagaráðs 2011).The constitutional council submitted its blueprint of a new constitution at the end of July 2011 (DV.is 2011).
It is safe to say that the constitutional assembly election was already controversial before it took place.Among those who most fiercely opposed the election procedures were people who were either blind or visually impaired.They pointed out that it was impossible for them to vote in complete secrecy as they would need an assistant (Helgason 2010).Each candidate was given a number and voters had to write down the numbers on the ballot in rank-order.In each voting booth there was a poster with the names of all of the candidates and their numbers.
As mentioned earlier, some of the candidates already had a jumpstart on others because of their previous visibility in the media.Other candidates had to find ways to reach people.Quite a few of them used social media like Facebook (Guðmundsson 2010).Direct advertising was discouraged as there had been some controversy a few months earlier due to high contributions by banks and companies to political campaigns (DV.is 2010).
The Ministry of Justice was legally obligated to print and distribute pamphlets with information on the candidates and the election, and did so shortly before the election.
There seemed to be a general consensus among the public that the media had somehow failed the test of providing adequate information on the election and the candidates.In particular, the Icelandic State Broadcasting Service (RÚV) came under fire (Viðskiptablaðið 2010).RÚV is legally bound to ensure that the public receives objective information on Icelandic society.It must be objective in its programming, coverage and interpretation.It must provide extensive, reliable, general and objective news service on domestic and global affairs.Moreover, it must be a forum for differ ent points of view on current topics that concern the public (Lög um Ríkisútvarpið ohf.2007).
Many candidates harshly criticized RÚV for its inability to cover the constitutional assembly election (Tryggvason 2010) and on November 16, 2010, some of them met with the director of RÚV, Páll Magnússon, and demanded air time (Eyjan.is2010).
On November 4, it had been stated by a staff member of RÚV that because of the large number of candidates, RÚV had decided not to speak to any of them in order to be fair to everyone (Guðbrandsdóttir 2010).Nevertheless, RÚV responded to the criticism by allocating five minutes to each candidate on Channel 1 (radio) from November 22 through November 26.

369
The Icelandic media coverage of the constitutional assembly election Guðbjörg Hildur Kolbeins On the eve of the election, RÚV's main newsmagazine show, Kastljós, which is broadcast straight after the 7 o'clock nightly news, didn't mention the election at all but rather discussed five women's allegations of sexual misconduct by a religious leader (RÚV.is2010).

Agenda-setting in elections
As stated above, the Icelandic State Broadcasting Service (RÚV) has a legal obligation to be objective in its news coverage.General work rules for the news room staff emphasize how highly valued objectivity is.The broadcast journalists are required to be fair and objective, and allow opposing viewpoints at all times, if possible.Also, it is clearly stated in the in-house guidelines for the journalists that all issue perspectives should be covered in news analysis and news programs (Jónsson 2011).
These formal and informal requirements of objectivity posed a conundrum for RÚV leading up to the constitutional assembly election.How should RÚV maintain its objectivity in the coverage of 522 candidates?The solution was to ignore them all.However, by so doing RÚV minimized the issue salience among its audience.
Based on Cohen's famous words that the press does not tell people what to think but what to think about (Cohen 1963), McCombs andShaw said in 1972: Readers learn not only about a given issue, but also how much importance to attach to that issue from the amount of information in a news story and its position (176).This is an important point because the lack of coverage of the constitutional assembly might have installed the sense in the public that writing a new constitution was not a top priority and rather unimportant.According to Rogers and Dearing (2007), when the media completely ignore an issue it doesn't register on the public's agenda."Saliences have consequences.The transmission of saliences is the conceptual heart of agenda setting" (McCombs 1992, 821).It's also important to note that the agendasetting theory assumes causality (McCombs, Llamas, Lopez-Escobar & Ray 1997) even though most studies within the field have been cross-sectional.
During the past 40 years, the agenda-setting theory has become one of the dominant paradigms in mass communication research (Dunn 2009;Strömback & Kiousis 2010;Tsfati 2003) or ever since McCombs and Shaw came to the conclusion in the 1968 presidential election in the United States that there was a strong correlation between the media's coverage of issues and the perceived importance of these same issues among undecided voters.A later study by Stone and McCombs (1981) found that it may take the media agenda two to six months to transfer over to the public agenda.

First-level and second-level agenda-setting
In general, one can speak of first-level and second-level agenda-setting effects.Firstlevel agenda-setting effects have simply been examined by looking at the volume of coverage of particular objects (McCombs 2009) and their salience among the public.

Fraeðigreinar
The objects under investigation can be candidates, institutions or specific issues (McCombs, Lopez-Escobar & Llamas 2000).
Second-level agenda-setting studies, on the other hand, have concerned themselves with the substantive and affective attributes of the objects.Substantive attributes can be descriptions of candidates' personalities in the media, for example, while affective attributes relate to whether the object is portrayed in a positive, neutral or negative light in the media (McCombs, Llamas, Lopez-Escobar & Rey 1997).For instance, in the 1996 Spanish general election, there was a strong correlation between how candidates were portrayed by the media and how the public perceived them (McCombs, Lopez-Escobar & Llamas 2000).Put another way: Substantive attributes refer to specific cognitive components of an object, and affective attributes refer to evaluations and assessments of an object.For agenda-setting studies about the transfer of salience for issues, these theoretical concepts distinguish between the "issue itself (i.e., what to think about), specific aspects or frames of an issue (i.e., how to think about) and, finally specific journalistic evaluations (i.e., what to think)" (Chernov, Valenzuela & McCombs 2011, 144;Matthes (2006) as cited in Chernov, Valenzuela & McCombs 2011).
Other studies have found that the attribute agenda of the media and the attribute agenda of the public tend to lean heavily towards the personal qualifications and character of political candidates (Kim & McCombs 2007) and that the public's perception of candidates' personality predicted voting intention better than the candidates' position on issues (Wu & Coleman 2009).
Second-level agenda-setting effects appear to be conditional as a study by Wu and Coleman (2009) found that in the 2004 U.S. presidential race between John Kerry and George W. Bush there was a second-level effect for Kerry but not for Bush.The authors concluded that president Bush was already well known and thus the public had less need for information on him.
An interesting study by Kiousis and McCombs (2004) found strong correlations between the public's recognition of 11 political figures and both print and broadcast media's amount of exposure of the politicians.The more the media covered them, the more people tended to recognize them.The media coverage not only affected the public's sheer recognition of the politicians but also their attitudes toward them.
These findings are relevant for the present study as four of the 25 elected candidates to the constitutional assembly were media personalities, another six were current or previous university employees, and the rest of the elected candidates, except for one or two, were quite well-known for their work.Shortly before the election, 39 out of 365 candidates who responded to a survey by Guðmundsson (2010) reported that they considered themselves to be famous.Thus, one can claim that for the most part voters used past media-generated information as a heuristic device in selecting 25 individuals for the constitutional assembly.

371
The Icelandic media coverage of the constitutional assembly election Guðbjörg Hildur Kolbeins Many of the agenda-setting studies have been based on content analysis and not actually measured people's attention to different media outlets.According to Strömbäck & Kiousis (2010) and Shehata (2010), paying attention to political news increased the perceived salience of issues among the Swedish public during the 2006 Swedish general election.Also, when people paid more attention to political news during the campaign, their perception of the importance of issues changed.
Some have wondered if the agenda-setting effect might become obsolete with increased technological changes, more diverse media and people relying less on traditional media.This does not seem to be the case even though the agenda of high Internet users is certainly more weakly correlated with the media agenda than of those who use the Internet less.Coleman & McCombs (2007) concluded that although young people use traditional media less than older people, their media use "did not seem to influence the agenda-setting effect much at all" (503).
Finally, it should be noted that Vu and Gehrau (2010) have argued that the magnitude of the agenda-setting effect can to a certain extent be attributed to the idea of the two-step flow of communication (originating with Lazarsfeld, Berelson & Gaudet 1944).In their study, readers read an article that raised their interest in an issue.The increased interest prompted them to discuss the issue with others and perceive the issue as being more important than it previously was to them.This chain of events led to others reading the article as well.Vu and Gehrau believe that the direct agendasetting effects of the article was actually insignificant compared to the effects of the diffusion.

Who shapes the media agenda?
There is an ongoing debate whether the media set an independent agenda or whether the media agenda is shaped by others -elite groups in society such as political candidates, for instance.
The media-centered model assumes that the media are autonomous actors in influencing the public's interest.The transaction model, on the other hand, claims that the agenda is formed by the interaction of the media, the public and other players.For example, candidates try to get their messages across to voters but the messages are dependent on what the candidates expect will be welcomed by the voters.The media will not cover issues that neither the candidates nor the public are interested in (Dalton, Beck, Huckfeldt & Koetzle 1998).
Some have proposed models of agenda-building where elites and interest groups affect the media agenda which in turn influences the audience agenda which then leads to different evaluations of the public (Scheufele 2000).
In the early '80s, Weaver and Elliott examined the relationship between the agenda of the Bloomington Indiana City Council and the agenda of the newspaper Herald-Telephone.The newspaper covered the council's agenda for the most part, especially when it came to economic issues.However, the newspaper's ranking of social and recreational issues was different from the ranking of these issues on the council's agenda.Weaver and Elliott concluded: "…it is not quite accurate to speak of the

372
Fraeðigreinar press setting agendas if it is mainly passing on priorities set by other actors and institutions in the society (Weaver & Elliott 1985, 87)."Interestingly, some evidence has been found of intermedia agenda-setting, i.e. that certain media set the agenda for other media.Specifically, studies tend to show that newspapers affect the agenda of television (Roberts & McCombs 1994) and this has been supported in an eight-year longitudinal study in Belgium where the agenda-setting time-lag turned out to be one day.During elections, however, the intermedia agenda-setting effect disappears as the focus of all the media is then on following the candidates (Vliegenthart & Walgrave 2008).
Several studies on gubernatorial and presidential elections in the United States have examined the effects of political advertising, press releases and blogs on the media agenda.In 1996, when Bob Dole challenged Bill Clinton, Dole's advertising influenced news coverage (Boyle 2001).
Eight years later, in the race between George W. Bush and John Kerry, results from studies indicated that the websites of the candidates and their advertising were able to influence the media agenda somewhat.Nevertheless, Sweetser, Golan and Wanta (2008) maintained that the media still took the initiative in setting the agenda.Tedesco (2005), however, claimed that during the Bush versus Kerry race, Bush's press releases managed to have an effect on the agenda of three major U.S. newspapers while at the same time the newspapers affected Kerry's issue agenda.
More recently, findings from a seven-country study on the 2008 U.S. presidential election showed that Barack Obama was covered three times more often than John McCain and that there was a relationship between the candidates' press releases and the global media coverage.Furthermore, there was also a connection between the media coverage of Barack Obama and the foreign public's support for him (Kim, Xiang & Kiousis 2011).
In a 2002 Florida gubernatorial race between Jeb Bush and Bill McBride, a positive relationship was found between the candidates' press releases and the media coverage of issues, and between the media coverage and the perceived importance of these issues to the public (Kiousis, Mitrook, Wu, Seltzer 2006).
It is clear that the connection between candidates' public relation messages and the media is not simply a matter of cause and effect.In some cases candidates are able to influence the media coverage while at other times they are responding to issues in the media.In the 2005 Virginia gubernatorial election, the relationship between the agenda of the two candidates and two major newspapers was bidirectional while the relationship between the candidates' agendas and two other smaller newspapers was unidirectional (Dunn 2009).
Finally, Hopmann and his colleagues found in the 2007 national election campaign in Denmark that political parties can influence the evening television news by the sheer amount of press releases on specific subjects.More relevant parties had more success with their press releases (Hopmann, Elmelund-Praestekaer, Albaek, Vliegenthart & de Vreese 2012).
To summarize; in ordinary circumstances, at least some political candidates can

373
The Icelandic media coverage of the constitutional assembly election Guðbjörg Hildur Kolbeins affect the media coverage through their advertising, press releases and websites.Prior to the constitutional assembly election, the majority of the candidates stayed away from advertising as 255 out of 365 candidates said they didn't buy any advertising space at all.The candidates did, however, publish articles in online news outlets and use blog sites (Guðmundsson 2010).It can be argued that if the purpose of these publicity attempts was to raise the awareness of the mainstream news media, it was to no avail as the media appeared to be determined not to cover the individual candidates.

Psychological factors
According to Chernov, Valenzuela and McCombs (2011), the need for orientation is a key concept in agenda-setting research.As McCombs (2005) pointed out, in the 1968 Chapel Hill study on undecided voters the issue (the election) was highly relevant to the voters and the uncertainty was high as they were still undecided.Consequently, the voters need for orientation was high."The media set the agenda only when citizens perceive their news stories as relevant" (McCombs 2009, 8).The greater the need for orientation, the greater the agenda-setting effect of the media (Chernov, Valenzuela & McCombs 2011).
The discussion on the need for orientation is certainly important for the case of the constitutional assembly election in Iceland.Chernov, Valenzuela and McCombs (2011) have maintained that when voters cannot rely on party affiliations and the candidates are unfamiliar to them, the news media are their main source for information.In such circumstances relevance and uncertainty are both high.Thus, one can say that the lack of coverage by the Icelandic media increased uncertainty among voters as the news media failed to provide them with adequate information.Shortly before the election in November of 2010, more than one third of voters were still undecided as to whether they would vote or not (MMR 2010b).
Results from studies on the need for orientation have shown that the agendasetting effect is often conditional.Not everyone is affected to the same degree by media coverage (Weaver 2007).For instance, at least one study has found that trust in the media plays a role in the strength of the effect.When people trust the media, the correlation between the media agenda and the public agenda is .70.However, when people are skeptical of the media, the correlation is a lot lower or .58.In addition, the more the skeptical people are exposed to the media, the closer their response is to those who trust the media (Tsfati 2003).
It is tempting to blame the lack of media coverage for the low voter turnout.

374
Fraeðigreinar Uninformed voters are less likely to vote (Sobbrio & Navarra 2010) and there is a clear relationship between the amount of election coverage on television news and one's voting behavior.More importantly, high visibility of campaigns in the media can dramatically increase the likelihood that those with low interest in political news will actually vote (Banducci & Semetko 2004).Also, lack of election news may give the audience the impression that elections lack legitimacy (De Vreese, Banducci, Semetko & Boomgaarden 2006).This may have been the case in the constitutional assembly election in Iceland, i.e. because the media did not pay much attention to the election, the election lacked legitimacy in the eyes of the public.
Studies on the relationship between media use and voter turnout have repeatedly stressed the importance of the public broadcasting systems.For example, public broadcasting services spend more time on elections than privately owned media (De Vreese, Banducci, Semetko & Boomgaarden 2006).Schmitt-Beck and Mackenrod (2010) even claimed that soft news on commercial channels decreases voter turnout while hard news on public broadcasting channels has the opposite effect.
In Denmark, it was found that the public broadcaster DR had more conflict in its coverage than its competitors and being exposed to conflict news increased people's likelihood of voting.Reading newspapers and discussing politics with others also increased the voter turnout (De Vreese & Tobiasen 2007).The role of newspapers in civic participation has also been found in the United Kingdom (Livingstone & Markham 2008).
A comparison of media systems and voter turnout in 74 countries revealed that voter turnout is highest where there are state/public broadcasting systems or mixed systems.The voter turnout rate is 11-13% lower in countries with only privately owned broadcasting services (Baek 2009).
To summarize, public broadcasting systems like the Icelandic National Broadcasting Service play a crucial role in elections.Their election coverage is more extensive and more in-depth, and the users of public broadcasting channels are more likely to vote than those who rely on commercial media for information.Hence, the salience of campaign issues is higher on public channels.

Methodology
The data for the present study were gathered by 22 students in a mass communication class at the University of Iceland, under the supervision of the author and Þorbjörn Broddason, professor of sociology.Most of the data were made accessible to the coders, free-of-charge, through an online database owned and operated by CreditInfo.However, as CreditInfo did not collect online news until 2010, the data for the general election were collected from each website individually.Unfortunately, data for the online news site Vísir.iswere not available for the general election.This means that figures for the general election are underestimated.
As the main objective of the study was to look at the news media coverage of the elections, only news-related stories were coded that specifically referred to the general election in April, 2009, the municipal election in May, 2010, and the constitutional

375
The Icelandic media coverage of the constitutional assembly election Guðbjörg Hildur Kolbeins assembly election in November, 2010.Each news story on the elections was the unit of analysis.
The time frame for analysis was 14 days, i.e. the election day itself and the last 13 days prior to the elections.
The sample included 13 national media outlets: Fréttablaðið (a free paper that is published six days a week), Morgunblaðið (published six days a week), DV (published three times a week), RÚV radio news at 12.20 p.m., RÚV television news at 7 p.m., Bylgjan radio news at noon, Stöð 2 television news at 6.30 p.m., Mbl.is (the news website of Morgunblaðið and the most visited online news site in Iceland, according to Modernus (2012)), RÚV.is (RÚV's online news), Vísir.is(the online news site of Bylgjan and Stöð 2), DV.is (the online news site of DV), Eyjan.is and Pressan.is.

The total news coverage of the three elections
It was hypothesized that the media had spent less time on the constitutional assembly election than on the general election or the municipal election.The findings clearly confirm this hypothesis.There are 632 news stories on the general election, 590 stories on the municipal election but only 165 stories on the constitutional assembly election.
The media coverage of the constitutional assembly election is only 12% of the total election coverage that was analyzed (figure 1).Thus, for some reason, the Icelandic media did not feel compelled to cover the constitutional assembly election to the same extent as they normally cover elections.As seen in figure 2, there is a huge difference in the news coverage between the constitutional assembly election, on one hand, and the general election and the municipal election, on the other hand.Looking more specifically at the coverage on day-by-day basis, one notices that there are fluctuations.Those fluctuations are caused by the fact that since the fall of 2008 the newspapers ceased publishing seven days a week.Also, before the general election in April 2009, one day was a national holiday,

STJÓRNMÁL
& STJÓRNSÝSLA 376 Fraeðigreinar i.e. the First Day of Summer that is celebrated the third Thursday of April each year.
Not until 11 days before the constitutional assembly election did the media coverage pick up slightly.There wasn't a single news item about the election on November 14 and November 15, 2010 (figure 2).There are some differences in the news coverage by medium (figure 3 and figure 4).In general, the online news sources covered the general and the municipal elections better than the traditional media.It should be kept in mind, of course, that on the election day itself the online media have an advantage over the traditional media as they can provide constant updating as the election results trickle in.

377
The Icelandic media coverage of the constitutional assembly election Guðbjörg Hildur Kolbeins

378
Fraeðigreinar Three of the online media, i.e.RÚV.is, Mbl.is and Vísir.is,covered the constitutional assembly election the most (figure 5).RÚV may have increased its coverage as a response to the criticism.

The attitude of the news media towards the three elections
The present study attempts to evaluate whether the news coverage is positive, negative or neutral (the affective tone).The coverage has a tendency to be rather neutral in all three cases (figure 6).If anything, there are more positive stories on the constitutional assembly election than on the other two elections.
It should be acknowledged at this point that trying to determine whether a specific news story is positive, negative or neutral is in the eye of the beholder and highly subjective.For the purpose of this study, the coders tried to predict whether the news story they were coding would have encouraged or discouraged voter participation.

379
The Icelandic media coverage of the constitutional assembly election Guðbjörg Hildur Kolbeins

Who made it into the news?
Prominence is an important factor in the agenda-setting effect.Consequently, it is important to examine the salience of candidates and others in the news media coverage of the three elections that are under investigation in this paper.Moreover, by looking at whom the news media interviewed, one gets a sense of from what perspective the news media covered the elections.
Not surprisingly, the main interviewees for news stories on the general election are all politicians (figure 7).Bjarni Benediktsson, the leader of the Independence Party, is most frequently the main interviewee or the main source for a story; followed by Prime Minister Jóhanna Sigurðardóttir, the leader of the Social Democratic Alliance.Interestingly, Ástþór Magnússon, the leader of the Democratic Movement, is more frequently the main interviewee than the leader of the Progressive Party, Sigmundur Davíð Gunnlaugsson.The Democratic Movement had no members in Alþingi.
Magnússon's prominence is due to his battle for publicity.He has repeatedly accused RÚV of not paying any attention to his political efforts (Morgunblaðið 2004).The comedian and a candidate for the Best Party, Jón Gnarr who later became the mayor of Reykjavík, received considerable media attention before the municipal election (figure 8).There are 39 news stories on Jón Gnarr and 17 news stories on the frontrunner of the Independence Party, Hanna Birna Kristjánsdóttir, who was the mayor of Reykjavík at the time of the election.The Best Party won quite a victory or 34.7% of all the votes in Reykjavík, even though it promised to break all of its campaign promises (Magnúsdóttir 2011;MMR 2010a).
The prominence of experts is another thing worth mentioning.Three of those, to whom the media spoke most often prior to the municipal election, are political scientists (figure 8).This reflects the horse-race aspect of the election.The media seemed more likely to cover polls predicting the election results than the issues of the political parties.This is consistent with McComb's and Shaw's Chapel Hill study that found that news coverage of campaigns often focus more on analysis of the campaigns rather than the issues (McCombs & Shaw 1972), and a study by Boyle on the 1996 U.S. presidential election that showed that horse-race stories play a dominant role in campaign-related coverage.Horse-race stories tend to focus on the latest poll results, campaign strategies and campaign events (Boyle 2001).These findings are also consistent with Fico and Freedman (2001) who posit that horse-race experts are often university political scientists.In their study of the 1998 gubernatorial election in Michigan, 17% of the stories cited horse-race experts, and nearly all of them were political scientists at Michigan State University.

381
The Icelandic media coverage of the constitutional assembly election Guðbjörg Hildur Kolbeins Figure 8.The number of news stories on the main interviewees in the municipal election When looking at the interviewees for the constitutional assembly election, it is obvious that the media stayed away from talking to the candidates for news stories.Not a single candidate was interviewed (figure 9).
The person, who appeared most often in the news, is an election supervisor for the Reykjavík District Commissioner.Others who made the news were either directly affiliated with the technical side of the election or were raising the issue of blind people not being able to vote without assistance.
The Minister of Justice, Ögmundur Jónasson, was also interviewed quite often on the election procedures and the difficulties of the blind people and the visually impaired (Logason 2010).
To summarize, contrary to other elections, the media did not allow any appearance of the candidates for the constitutional assembly in news stories.It may be assumed that this was due to the media's fear of being accused of being partial by favoring some of the candidates.

STJÓRNMÁL
. The number of news stories on the main interviewees in the constitutional assembly election

Discussion
The present study reiterates that we gain a much deeper understanding of a phenomenon by placing it in a larger context.By examining three elections that were held within a two-year period, we are able to detect a pattern.The emerging picture looks something like this: The media thrive on the kind of tension and polarization, and for-and-against discourse, that occurred in Iceland in the spring of 2009.The media could embrace their impartiality by zigzagging between the views of the government in power and the opposition.The issues were relatively clear cut; The Social Democratic Alliance and the Left-Green Movement emphasized, among other things, that Iceland should be a Nordic welfare-state and join the European Union while the Independence Party and the Progressive Party were playing defense after the onset of the economic depression.
Things got a bit more complicated in the municipal election but Jón Gnarr and his Best Party saved the day with their talk of buying a polar bear for the domestic animal zoo (Halldórsson 2010).As the issues are more diverse in a municipal election, depending on the municipalities, the media tend to cover this election as a race, i.e. interviewing university professors who can tell them who is likely to succeed and who is not.Hence, the media may have looked past some of the critical issues in their news stories.
Once the candidates and issues are even more fragmented, as happened in the STJÓRNMÁL & STJÓRNSÝSLA

383
The Icelandic media coverage of the constitutional assembly election Guðbjörg Hildur Kolbeins constitutional assembly election, the media did not know how to align their own need for impartiality and objectivity with the large number of candidates.In the end they focused on a single issue with two sides, i.e. the visually impaired versus the Ministry of Justice and the Electoral Commission, and completely ignored the candidates and their issues.
Even though the media did not cover the candidates for the constitutional assembly in their news stories, they might have spent more time on the issue itself, i.e. the constitution and the upcoming election.Because of the election, the ground was fertile for discussion on the subject.One could claim that it was important to give voice to the opposing viewpoints and the matter in general.Thus, the need for impartiality does not alone account for the scarcity of news stories on the election.
What were the implications of the failure of the media to cover the constitutional assembly election to the same extent as other elections?The media revealed their inadequacy to fulfill their role as an informant when the issues and perspectives are many and varied.
The lack of information might also have contributed to the public's feeling of not knowing enough about the candidates, or what they stood for, to be able to elect 25 candidates to the constitutional assembly.Hence, the media may have deterred people from carrying out their constitutional right to vote.
The Office for Democratic Institutions and Human Rights, and the European Commission for Democracy through Law (The Venice Commission), have stated that the media play a crucial role in elections by being a forum for public debate and by being a channel for information between the public and the candidates.Furthermore, the media are obligated to provide candidates with an equal opportunity to express their views and the voters have the right to make informed decisions based on accurate media coverage.Finally, public broadcasters are held to higher standards in terms of fairness than privately owned media and they "should provide a complete and impartial picture of the entire political spectrum in their coverage of an election" (OSCE 2009, 7).Even the European Court of Human Rights has stressed the right of the public to be informed by the media (Thorgeirsdottir 2004). " The visibility and identification of potential representatives is a prerequisite for a healthy democratic process," as de Vreese and his colleagues have pointed out (De Vreese, Banducci, Semetko & Boomgaarden 2006, 482).Keeping in mind the responsibility of the media to inform potential voters in elections and the right of the voters to receive impartial and accurate information on candidates, it seems clear that the Icelandic National Broadcasting Service fell short in fulfilling its legal duty to provide voters with adequate coverage of the constitutional assembly election.The election was out of sight, out of mind.RÚV and the other Icelandic media ranked the constitutional assembly election low on their agenda and by so doing decreased the legitimacy of the election in the eyes of the public.As McCombs (2005) has put it: "Setting the agenda is an awesome responsibility" (556).

384
Fraeðigreinar The Icelandic media coverage of the constitutional assembly election Guðbjörg Hildur Kolbeins

Figure 1 .
Figure 1.The total news media coverage of the 2009-2010 elections (percentages)

Figure 2 .
Figure 2. A comparison of the number of stories on the 2009-2010 elections, dayby-day

Figure 3 .Figure 4 .
Figure 3.The number of news stories on the general election, by medium

Figure 5 .
Figure 5.The number of news stories on the constitutional assembly election, by medium

Figure 6 .
Figure 6.The general attitude of the news media towards the 2009-2010 elections (percentages)

FraeðigreinarFigure 7 .
Figure 7.The number of news stories on the main interviewees in the general election