Party leadership over political affiliation?
What indicates whether and how a politician appears on TV during election time Melle de Haan
11421428 Master’s Thesis
Graduate School of Communication
Master’s programme Communication Science: Political Communication Michael Hameleers
Abstract
Why do some politicians appear more on television than others? For many people, television is the major source of information to learn about politics. It is therefore interesting to find out why politicians differ in how and whether they appear on TV. In this study I investigated what indicators influence the appearance of politicians on TV and thus why some politicians appear more than others. For this reason, a content analysis of Dutch politicians (N=102) appearing on political television programs during the 2017 general parliamentary election was conducted. The results show that, despite contrary claims from academia and politics, the appearance of politicians on TV is not influenced by a politician’s political affiliation. Therefore, the findings suggest that Dutch media is secularized. When selecting which politicians to invite as guests on their programs Dutch media make use of a so-called equal access approach. A second important finding of this study is the fact that being the leader of a political party has the largest impact on the likelihood of a politician to appear on TV.
Keywords: commercialization, personalization, mediatization, content analysis, appearance of politicians.
Party leadership over political affiliation?
What indicates whether and how politicians appear on TV during election time Introduction
Appearing on television is of great importance for politicians, especially during election times. For many people television is the major source of information to learn about politics (Schoenbach, De Ridder, & Lauf, 2001; Strömbäck, 2008). The public is dependent on the media for information about politics and politicians depend on the media to provide them with a stage (Bos, Van Der Brug, & De Vreese, 2010; Kepplinger, 2002; Strömbäck, 2008). In this setting or mediatized reality, the focus of this study is to explain why some politicians appear more on TV than others.
According to Pitkin (1967) optimal political representation occurs when politicians make sure that the voices and opinions of all citizens are being considered during policy-making processes. In order to make well-informed voting decisions, people first of all need to know whom they can vote for. Television plays a major role in the provision of this information.
The way politicians make use of media platforms, such as television, influences how they reach the public and therefore whether or not political representation is optimized. In addition, a recent growth in voter volatility has increased the influence of television on the popularity of politicians (Geers & Bos, 2017). For this reason, it is interesting to study how politicians that appear often on television differ from those who do not. Here, it will be studied what indicates whether politicians appear on TV during election time and how they do so. It is possible to look at the appearance of politicians in two ways: contextually and content related.
Contextual indicators, like a politician’s gender or political preference, are important predictors of visibility, but also fixed factors beyond the media’s influence. A politician (usually) does not change their gender in order to gain media attention or appear on TV. That is why for the purpose of this research content related factors will be the main focus of study. Using these content features in their advantage politicians can adapt their strategy to ‘what works’ and so influence the amount and in what way they will appear on television. Content related factors focus on the behaviour, style and rhetoric of politicians while appearing on TV. One could say that politicians and the media perform some sort of symbiotic dance, a tango, where sometimes the media leads and sometimes the politicians (Gans, 1979). In this study I want to find out why certain politicians seem to dance better than others.
The goal of this study is to find out what (content related) factors influence a politician’s chances of (positively) appearing on television during election time. This time frame was chosen because during election time most is at stake for both
politicians, who have the possibility of getting elected, and voters, who can elect. Therefore it is expected that during this time politicians appear most on TV.
My research question is the following: what indicates whether and how a politician appears on TV during election time? In order to answer this research question a quantitative content analysis will be conducted of television appearances by Dutch politicians during election time in the Netherlands. Findings will show how much politicians themselves can influence their appearance on TV or whether this is mostly determined by fixed factors beyond their direct influence or control.
Theoretical framework
An increasing amount of personalization, mediatization and
couple of decades (Blumer & Kavanagh, 1999; Kepplinger, 2002; Mazzoleni, 1987; Strömbäck, 2008). These developments have meant a substantial shift in how the media, politicians and the public communicate with each other. This has influenced the production of political television-shows and thus indirectly which politicians appear on it and in what ways they are depicted. In today’s world, politicians have become more dependent on media and lost much grip on whether and how they appear on TV.
Personalization, mediatization and commercialization in media are
interconnected and cannot always be clearly separated from each other. The following example illustrates this: politicians these days share much more of their private life, which can be seen as a sign of personalization of politics. However, for a large part this is rooted in the commercialization of media because it is thought that a more engaging and emotionally rich approach to politics will attract more viewers and thus result in higher profit (Alencar & Kruikemeier, 2016). Adding to this one could argue that the fact that politicians continue to participate in all sorts of non-political
entertainment focused campaigning is due to an increase in mediatization, which entails a growing dependency of politics on media (Hjarvard, 2008).
To better understand the media’s role in whether and how politicians appear on TV during election time, I will first of all discuss how the commercialization of media may have contributed to the visibility of politicians on TV. Next, the
personalization of politics and its consequences on the production of political television shows will be discussed. At last, processes related to mediatization and media logic are taken into account.
Commercialization of media
According to Blumler and Kavanagh (1999) we currently find ourselves in the third age of political communication, which is marked by the professionalization and commercialization of our media landscape. Commercialization of media means that content is produced principally to earn money and to make profit. A direct
consequence for the appearance of politicians on TV here is that it is expected that the producers of television programs take into consideration which political guests attract large audiences since that means more profit. According to Brants and Neijens there has been “a shift from programs in the public interest to programs the public is interested in” (1998, 150). The effects of commercialization are not possible to strip down into one hypothesis but more so play a part in basically all and therefore are included within the hypotheses covering personalization and mediatization. Personalization of politics
The personalization of politics is twofold because it entails an increasing attention in politics and media for the personalized and individual characteristics of both politicians and their voters (Caprara & Zimbardo, 2004). The focus on the individual actor, a party’s political leader, has increased (Rahat & Shaefer, 2007). Bennett talks in his study about “the rise of an era of personalized politics in which individually expressive personal action frames displace collective action frames” (2012, p. 21). When trying to attract large audiences personalization is one of the news values used by the mass media (Kriesi, 2012). According to Strömbäck personalization is, together with conflict, “the most important storytelling technique that media prefer when choosing what and how to cover politics” (2008, p. 238). The popularity of personalization in politics has for example meant that we know that the Obama’s have a dog called Bo and that Dutch green party leader Jesse Klaver used to
be very close to grandfather before he passed away. In other words, voters get to know things about politicians that have nothing to do with the policies they promote or even the ideology they represent.
Therefore, it is interesting to study whether the personalization of politics has made its way into political television programs. As mentioned before, personalized politics is interconnected with commercialization of media. TV programs hope to reach a bigger audience by focusing not only on hard-news policy topics when covering politics but also to shine light on the private life of a politician (Alencar & Kruikemeier, 2016). Personalized coverage of politics focuses on politicians both as public and private citizens (Takens, Van Atteveldt, Van Hoof, & Kleinnijenhuis, 2013). Personalization of politics leads to politicians using easier to understand language, to talk more about their private life and to discuss non-political content when appearing on TV. This is why I introduce the following hypotheses on the effects of personalization:
H1a: talking about private matters has a positive effect on the likelihood of a politician to appear on TV.
H1b: using colloquial language has a positive effect on the likelihood of a politician to appear on TV.
H1c: discussing non-political content has a positive effect on the likelihood of a politician to appear on TV.
Mediatization
Mediatization occurs when the impact of media on society increases (Asp & Esaiasson, 1996; Blumler & Kavanagh, 1999; Falasca, 2014). A large part of this impact is on politics and mediatization is often used to explain the dependency of
politics on media (Ekström, 2001; Hjarvard, 2008). The dependency of politics on media can be explained by the existence of a ‘media logic’, where media outlets, independently from political players, set the media agenda and thus lead the conversation both in the TV studio and on the street.
Because mediatization is not as a straight-forward term as commercialization or personalization, the effects of it on the appeareance of politicians on TV will subdivided it into different factors. The following factors of mediatization will be tested: the existence of media secularization, the popularity of conflict frames and populist rhetoric when selecting politicians to appear on TV and the influence of horse race journalism in this process. Before discussing these factors individually below, I will first clarify more in depth how academia has interpreted the umbrella term mediatization before.
Blumler and Kavanagh speak of a “media-constructed public sphere” when describing what happens when mediatization occurs (1999, p. 211). Western
democracies like the Netherlands transitioned during the last centuries from a political party logic dominated media landscape to a media logic one (Mazzoleni, 1987;
Strömbäck, 2008). Magin describes best the way mediatization of politics works in her article: “the more important the media are as information sources, the more independent they are of political institutions, and the more media coverage and political actors are governed by media logic, the more mediatized politics is” (Magin, 2015, p. 417). To assess the degree of which politics is mediatized argues Strömbäck (2008) in his study of the existence of different phases of mediatization. He identifies four to be precise.
In the Netherlands we currently find ourselves in the fourth phase of
communication, the commercialization of media and the existence of permanent political campaigns (Strömbäck, 2008). These developments have lead to the existence of a 24/7 news cycle and popular use of click bait journalism. For politicians this final phase of mediatization has meant that endless hours of media trainings and working closely together with spin-doctors has become an important part of their job.
Media secularization. When you look at the selection of politicians appearing as guests on political programs one can assume that journalists in a mediatized world have much more autonomy on deciding whom they invite. During ‘the party logic era’ it was easy to affiliate a television channel or media outlet with a political party (Mazzoleni, 1987). Media would cover political parties and politicians of their preference much more frequently and more favorably than others. During the pillarization of the Netherlands journalists were said to treat politicians belonging to the same ideological pillar better than those who were not (Kleinnijenhuis, Oegema, & De Ridder, 1995).
In a mediatized media landscape there is expected to be less party-press
parallelism and more media secularization (Magin, 2015). According to Boomgaarden and Vliegenthart is “the Dutch media system characterized by a rather autonomously acting press, without partisan alignment” (2009, p.407). Therefore, I expect that politicians appearing on television programs get treated similarly regardless of their political background, despite allegations of a leftist media bias (Schoenbach et al., 2001). I expect to see this media secularization to be visible in both the access to appearance on TV (meaning the amount they appear) as in the treatment while appearing. The following hypotheses will be tested:
H2a: a politician’s political preference does not have effect on the likelihood of a politician to appear on TV.
H2b: a politician’s political preference does not have effect on the way he or she gets treated by the interviewer while appearing on TV.
Conflict frames. In the current media logic it is expected that political
television programs choose to invite their guests based more on their newsworthiness rather than on ideological similarities between the guest and the program (Mazzoleni, 1987). Blumler and Kavanagh argue that media coverage of politics happens “on the basis of its news value or likely audience appeal” (1999, p. 218). Harcup and O’Neill (2017) have written extensively about news values and the selection of news. In their most recent study Harcup and O’Neill (2017) present a contemporary set of news values in which bad news is noted as the most prominent one. Bad news can be linked to controversy and unexpectedness, which according to other research contributes to the newsworthiness of a story in media (Helfer & Van Aelst, 2015). Bad news and controversy both are part of conflict, which is a frequently used frame for media when reporting about politics (Bartholomé, Lecheler, & De Vreese, 2017; Strömbäck, 2008). A conflict frame focuses on disagreement between (political) actors. Mass media rather focus on conflicts between politicians than on agreements (Kepplinger, 2002). Conflict and disagreement on television is thought to increase ratings and can therefore also be linked to the commercialization of our media landscape. It is popular amongst journalists since it usually highlights both sides of an argument and thus is considered to be (at least on paper) objective, which is a popular striving within journalism and media (Bartholomé et al, 2017). Therefore I introduce the following hypothesis:
H3: using a conflict frame has a positive effect on the likelihood of a politician to appear on TV.
Populist rhetoric. All over Europe have populist political parties enjoyed electorate success during the last century and in many Western countries they make up for a significant part of parliament. A strong anti-establishment characterizes populist parties (Bos, Van Der Brug, & De Vreese, 2011). Often, but not always, are populist parties critical of mass-immigration and the Islam (Bos & Brants, 2014).
Because of the recent popularity of populist parties more politicians, also those not traditionally seen as populist, use characteristics of populist rhetoric when campaigning during election time (Braouezec, 2016). Populist rhetoric amongst mainstream politicians “is said to have spread like the flu” (Bos & Brants, 2014, p. 750).
The fact that media gives populist rhetoric a platform can be linked to all three aforementioned developments in political communication; with personalization because populist rhetoric often is ‘easy to understand’ and ‘speaks directly to the public’ (Bos et al., 2010), with commercialization because it attracts large audiences and with mediatization because populist rhetoric fits in perfectly with the fourth phase of mediatization characterized by the professionalization of political communication and the need to fulfill a 24/7 news-cycle media landscape. Therefore I introduce the following hypothesis:
H4a: using populist rhetoric has a positive effect on the likeliness of a politician to appear on TV.
The relationship between the media and those using populist rhetoric seems rather bad. US president Donald Trump, who often uses populist rhetoric, repeatedly has called the media the enemy of the American people (Grynbaum, 2017). In the
Netherlands refuses populist Geert Wilders to appear in most TV programs because they, according to him, do not objectively report about him. According to Bos and Brants (2014), who did research on populist rhetoric in media, are elite media often critical of those using populist rhetoric and are they usually aligned with the political status quo. Other studies have made allegations of the existence of a leftist media bias, which entails that ‘conservative’ populists would get treated less favorably (Schoenbach et al., 2001). Likewise is the media said to act hostile towards extreme rightwing politicians (Bos et al., 2011), which are often those using populist rhetoric. Therefore I introduce the following hypothesis:
H4b: using populist rhetoric has a negative influence on the way a politician gets treated by the interviewer while being on air.
Horse-race journalism. The last factor of mediatization I will discuss here is the existence of horse-race journalism. Mediatization and media logic are
characterised by the existence of spectacularization within political coverage,
(Mazzoleni, 1987) which entails that the media presents political campaigns as ‘horse races’ (Kepplinger, 2002; Mazzoleni, 1987; Strömbäck, 2008). Horse race journalism is political journalism of elections, which comes close to the reporting of horse races due to a focus on polling data instead of candidate policy (Boukes & Boomgaarden, 2015; Broh, 1980). Political horse race coverage is characterized by portraying politicians as either winners or losers while using language from sports and wars (Aalberg, Strömbäck, & De Vreese, 2011). The focus on horse race and strategic aspects of politics is believed to help achieve higher ratings and therefore can be connected to commercialization of political coverage (Boukes & Boomgaarden, 2015).
et al., 2013) or the game frame (Aalberg et al., 2011), continues to dominate media coverage of major political events (Iyengar, Norporth, & Hahn, 2004). After losing the US presidential election of 2016 Hillary Clinton criticized the media for turning political coverage into a horse race. She writes in her book; “I understand that
political coverage has to be about the horse race, but it’s become almost entirely about
that” (Clinton, 2017, p. 99).
I expect that horse race journalism also has had impact on the way Dutch media covered politicians during the election in the Netherlands and thus on which political guests were invited to appear. When media coverage of politics is portrayed as a horse race the focus of the coverage lies with which actors are winning and which one’s are not (Aalberg et al., 2011, Broh, 1980; Iyengar et al., 2004). Therefore I introduce the following hypothesis:
H5a: performing unexpectedly (either bad or good) in the polls has a positive effect on the likelihood of a politician to appear on TV.
Besides this, I expect that the way politicians ‘are doing’ in the polls will influence the way they will be treated while appearing on political TV shows. Media are inclined to focus more on negative stories (Takens et al., 2013). Therefore, when a party is doing badly, mediatization in the form of horse race journalism will be
expected to emphasize this. Also, being critical towards politicians’ helps journalists to look independent and professional (Lenguaer, Esser, & Berganza, 2011). My expectation is that a politician that is doing badly in the polls will be treated more critically on TV than a politician who is doing well. This because of the emphasis on negative stories in which media question what is going wrong and why voters don’t intend to vote for politicians doing badly in the polls. Stemming from this theory, I introduce the following hypothesis:
H6b: performing badly in the polls has a negative influence on the way a politician gets treated by the interviewer on air.
Method
The hypotheses are tested by means of a content analysis of Dutch political television shows broadcasted during the general parliamentary election of 2017. In this study, a census sample was used. This means that all politicians appearing on the television programs included were coded and reported as such.
Data collection
A codebook was developed that contained questions regarding contextual and content related characteristics of the politicians appearing. Here contextual
characteristics are for example to which party politicians belonged and whether or not they were the leader of their party. Content related characteristics focused on how politicians appeared when being on air. An example of such question is Q9: does the politician appearing on television visible disagree with the person interviewing him or her? Please see appendix A of this paper for the entire codebook that includes the full list of questions and corresponding explanations provided to make them clearer. Sample
In this study, both hard news and more soft news infotainment type of political television shows were chosen to be part of the sample. The latter type of television shows provides a combination of information and entertainment (Alencar & Kruikemeijer, 2016). Please see table 1 for the full list of television programs and whether they can be seen as more hard news or infotainment programs. The data collection of television programs was available through the public network website npo.nl and for the commercial TV show part of this research through RTLXL.nl.
Excluded from the analysis were Dutch entertainment TV programs, in which politicians during election time do appear such as Koffietijd and RTL Boulevard. Even though such shows do contain politicians appearing as guests the focus is more
towards the politician as a personality rather than his or hers policy ideas (Göran, 2010). This study solely focuses on politicians that appear on television fully as such. My argumentation is here is that I want to study whether the three pillars central in this study - personalization, mediatization and commercialization - have made an impact in only hard news and infotainment television programs.
In total, eight Dutch television programs were selected for this study. De Wereld Draait Door (NPO1 at 6.30 pm), Pauw & Jinek: De Verkiezingen (NPO1 at 11:30 pm) and JINEK (NPO1 at 11.00 pm) are all three serious news-talk shows broadcasted on weeknights by the public network. WNL op Zondag (NPO1 at 09:30 am) and Buitenhof (NPO1 at 12:10 am) are two Sunday morning political talk shows as well part of the public network. The first one can be seen as more leftwing and the other as more rightwing. Breakfast-show Goedemorgen Nederland (NPO1 at 7:10 am) discusses news, soft news and politics during weekdays. Nieuwsuur (NPO2 at 10:00 pm) airs on weeknights and is best described as being a serious current affairs news program. The only commercial television program included is RTL Late Night (RTL4 at 10:30 pm), which is a late night show focused on a mix of infotainment, current affairs and politics.
A total of 62 television program broadcasts, which contained of over 54 hours of television, were coded for this study. Here 102 times a politician appeared as a guest and 33 different politicians were coded. On average appeared politicians a little less than six times and only two political actors appeared ten or more times on TV
during the studied period. Please see table 1 for the complete list of television shows included and number of politicians appearing per show.
Table 1. Television programs included
TV shows # of shows # of politics type of show
DWDD 6 11 infotainment
Jinek 1 1 hard news
Jinek & Pauw 12 36 hard news
Buitenhof 4 9 hard news
Goedemorgen NL 13 13 infotainment
WNL op Zondag 4 8 hard news
Nieuwsuur 12 12 hard news
RTL Late Night 10 12 infotainment
Time frame. Within academia there is a lack of consensus for what time frame is best when studying election campaigns. Some researchers analyzed only 20 days of an election (Hopmann, Elmelund-Praestekaer, Albaek, Vliegenthart, & De Vreese, 2012) and others up to seven weeks (Schoenbach et al., 2001). In this study, politicians appearing on television were coded during the last four weeks of the election. Like this, the 28 days between the 15th of February and the 15th of March are covered. This time frame was chosen because during the last four weeks of an
election the majority of people decide whom they are going to vote for (Schmitt-Beck & Partheymüller, 2012) and therefore has the appearance of politicians during this time most impact.
The politicians appearing on the aforementioned television programs are the units of analysis for this study. The politicians that were coded are all the men and women that were part of the official candidate lists of the political parties included on
the ballot paper for the Dutch general election of 2017. Former politicians not eligible were excluded from the study.
Dependent variables
This study contains of two dependent variables; appearing on television and hostile treatment on television. The variable appearing on television was created manually by counting the amount of times each politician appeared on television. Like this a continuous variable was constructed ranging from one to eleven, which was the most times a single politician appeared throughout the studied time frame. The variable hostile treatment on television was created by adding up four already existing variables constructed out of questions from the codebook into one index variable. This variable was made out of the following questions; Q15: does the interviewer criticize the politician? (yes = 1, no = 0), Q16: does the interviewer criticize the politician’s proposed plans? (yes = 1, no = 0), Q17: does the politician appearing on TV receive equal airtime? (yes = 0, no = 1) and Q18: is the host more critical towards this politician than to others? (yes = 1, no = 0).
Questions 17 and 18 imply that there was more than one politician present on the TV show and when this was not the case coders were asked to code for Q17 yes and for Q18 no. This was done because the variable hostile treatment is intended to measure in what way politicians are treated unfavorably or negatively by the media and when there is only one politician appearing on a show he or she cannot be treated worse than somebody who is not there. There cannot be drawn any comparison and therefore also no negative one. A 5-item likert scale was designed by combining the aforementioned questions into one index variable. Zero implies treated best way possible and four implies treated most hostile to this measurement.
Independent variables
The independent variables created for this study were the indicators in which politicians could differ from one another. Contextual variables included here have in common that they could all be answered before actually watching the coded content. These contextual variables were the following: gender (male = 1, female = 0), whether or not somebody was the leader of their party (yes = 1, no = 0), whether the politician was either leftwing or rightwing (leftwing = 1, rightwing = 0), whether or not the politician appearing belongs to a party in power (yes = 1, no = 0) and whether or not the politician appearing belongs to a party doing unexpectedly in the polls (yes = 1, no = 0).
This last variable was created by using polling data from Kantar Public and comparing this to results of the 2012 general election. When the amount of seats in parliament predicted for political parties in the Kantar Public poll of the 28th of February 2017, exactly halfway in my sample time frame, would differ more than 25% with the amount of seats in parliament they obtained during the 2012 election I coded a party as ‘performing unexpectedly’. Please see appendix B of this paper for the entire list of political parties and whether or not they were coded as performing unexpectedly. Out of the variable doing unexpectedly was created the final contextual independent variable; doing badly in the polls. With data from Kantar Public this variable was constructed by labeling all parties doing here at least 25% worse than during the 2012 election as doing badly in the polls.
Content related variables are the other type of independent variables and only possible to code when watching the actual television programs. The variables using colloquial language (yes = 1, no = 0), whether or not a politician talks about his or her private life (yes = 1, no = 0) and whether or not a politician discusses non-political
content (yes = 1, no = 0) are all single item straightforward indicators taken from the codebook. The independent variable being part of a conflict frame was constructed by adding up the following three questions from the codebook; Q9: does the politician appearing on TV visible disagree with the person interviewing him or her? (yes = 1, no = 0), Q10: does the politician appearing on TV discuss disagreeing with a person not present in the show? (yes = 1, no = 0) And Q11: does the politician appearing on TV visible disagree with other guests appearing on the show? (yes = 1, no = 0). The added variable has four different answer options ranging from zero to three and shows like this to what degree a conflict frame was present and thus the extend to which a politician was part of one.
A second independent variable that was constructed out of multiple questions from the codebook is the variable using populist rhetoric. Here it is intended to measure to what extend a politician appearing on TV uses populist rhetoric. The following three questions from the codebook have been added to construct this index variable: Q19: does the politician appearing make anti-establishment comments? Q20: does the politician refer to “the man on the street” or “the common man”? And Q21: does the politician appearing on TV critiques immigration? The added variable has four different answer options ranging from zero to three and like this illustrates to what extend the politician appearing on TV uses populist rhetoric.
Intercoder reliability. For this study a little over ten percent of the entire sample, which was 14 politicians appearing on TV, was coded by a second coder. It was made sure that all television programs included in this study were at least double coded once. All variables included in this study were measured according to
Krippendorff’s alpha to make sure the intercoder reliability was sufficient. Please see appendix B for an overview of how each variable scored on this measurement. The
average Krippendorff’s alpha score for all coding decisions combined is .85, which is considered a good score (Hayes & Krippendorff, 2007).
Results
Hierarchical multiple regression models were used to asses the relationships between dependent variables the appearance of politicians on TV and hostile treatment of politicians while being on air and the contextually and content related independent variables. Three different tables are constructed in order to answer the hypotheses of this study. The results of these hierarchical multiple regression analyses are presented in tables 2,3 and 4.
Personalization of politics
Table 2 presents results regarding the hypotheses formulated about personalization (H1a, H1b, and H1c). Appearing on television is the dependent
variable. In the first model, consisting of contextual variables, the regression model as a whole is statistically significant and accounts for 60% of the variance in appearance on television, F(4, 97) = 35.99, p<.001, R2 = .60, Adjusted R2 = .60. The second model as a whole is significant as well and accounted for 61% of the variance in the appearance of politicians on TV, F(7, 94) = 20.72, p<.001, R2 = 61, Adjusted R2 = .58, after controlling for three variables (discussing non political content, talking about private life, and using colloquial language) based on literature review.
When looking at the results of model 2 we need to reject hypotheses H1a, H1b and H1c, which state that using colloquial language, talking about your private life and discussing non-political content has a positive effect on the likelihood of a politician to appear on TV. Because the results for the three aforementioned variables are not significant the presumed positive effects cannot be proven.
statistically significant predictor variable with B = 4.90 (p<.001) for model 1 and B = 4.86 (p<.01) for model 2. Being the leader of a political party has a strong positive effect on the likelihood of a politician to appear on TV.
Table 2. Hierarchical multiple regression: personalization.
Model 1 (n=102) Model 2 (n=102) B SE Beta B SE Beta Constant 1.54 0.67 1.25 0.70 Party in power - 0.92 0.54 - 0.12 - 0.78 0.55 - 0.10 Party leader 4.90* 0.69 0.63 4.86* 0.67 0.62 Political spectrum 0.08 0.49 0.01 0.31 0.53 0.04 Gender 1.30 0.70 0.15 1.24 0.72 0.15 Colloquial - 0.21 0.49 - 0.03 Non political 0.19 0.52 0.03 Private life 0.65 0.53 0.09 Adjusted R2 0.58 0.58 F 35.99* 0.75 *p<.05
Note: dependent variable = appearance on television Mediatization on appearance
Table 3 presents the results regarding the hypotheses formulated about mediatization and its effect on the appearing of politicians on TV (H2a, H3, H4a and H5a). In the first model, consisting of only contextual variables, the regression model as a whole is statistically significant and accounts for 61% of the variance in
appearance on television, F(4, 97) = 37.73, p<.001, R2 = .61, Adjusted R2 = .59. The second model is again statistically significant as a whole and accounts for 72% of the variance in the appearance of politicians on TV, F(7, 94) = 34.33, p<.001, R2 = 72, Adjusted R2 = .70, after controlling for three variables (using populist rhetoric, doing unexpectedly in the polls, and being part of a conflict frame) based on literature review.
Hypothesis 2a, which states that a politician’s preference does not have an effect on the likelihood of a politician to appear on TV, needs to be accepted. The independent variable political spectrum has (in both models) a non-significant effect
on the dependent variable appearance on television meaning that in this study there is no proven effect found for this relationship to exist.
When looking at the results from the hierarchical multiple regression models in table 3 H3, which states that using a conflict frame has a positive effect on the likelihood of a politician to appear on TV, needs to be accepted. A significant effect was found between the two variables. A unit increase in the extent to which a politician is part of a conflict frame increases the amount a politician appears on TV with .59 times (p<.05).
H4a, which argues that using populist rhetoric has a populist effect on the likelihood of a politician to appear on TV, needs to be rejected when looking at the results in table 3. A non-significant relationship was found between using populist rhetoric and appearing on television. Therefore a relationship between the two variables cannot be proven.
When looking at the results of our regression analysis H5a, which states that performing unexpectedly (either bad or good) in the polls has a positive effect on the likelihood of a politician to appear on TV, needs to be accepted. A significant result between the two variables was found. When a politician belongs to a political party which is doing unexpectedly in the polls the amount a politician appears on TV increases with 2.38 times (p<.01).
The contextual variables gender and party were not part of any hypotheses but do show significant results. Model 2 shows that being male significantly increases a the amount of times a politician appears on TV (B = 1.77, p<.001). Just like table 2 shows table 3 that being the leader of your political party positively influences the amount of times a politician appears on TV. In table 3 is for model 1 this effect stronger (B = 5.08, p<.001) than seen in model 2 (B = 3.89, p<.001).
Table 3. Hierarchical multiple regression: mediatization on appearing Model 1 (n=102) Model 2 (n=102) B SE Beta B SE Beta Constant 1.57 0.60 - 0.28 0.71 Gender 1.08 0.72 0.13 1.77* 0.63 0.21 Party leader 5.08* 0.67 0.68 3.89* 0.61 0.50 Party in power - 0.90 0.54 - 0.11 - 2.21* 0.51 - 0.28 Political spectrum 0.12 0.48 0.02 - 0.03 0.45 - 0.00 Conflict frame 0.59* 0.21 0.16 Populist rhetoric - 0.40 0.38 - 0.07 Unexpectedly 2.42* 0.51 0.30 Adjusted R2 0.59 0.70 F 37.73* 34.33* *p<.05
Note: dependent variable = appearance on television Mediatization on hostile treatment
Table 4 presents the results of a third hierarchical multiple regression analysis regarding the hypotheses about the effect of mediatization on the treatment of
politicians while appearing on TV (H2b, H4b and H6b). This time hostile treatment on television acted as the dependent variable. In the first model, consisting of the contextual variables, the regression model as a whole was statistically non-significant and accounted for just 4% of the variance in hostile treatment on television, F(4, 97) = 1.08, p>.05 R2 = .,04, Adjusted R2 = .00. The second model was again non-significant and accounted for 6% of the variance in the hostile treatment of politicians on TV, F(7, 94) = .84, p<.001, R2 = .06, Adjusted R2 = -.01, after controlling for three variables (using populist rhetoric, performing badly in the polls, and being part of a conflict frame) based on literature review.
When looking at the results from table 3 we need to accept H2b, which states that a politician’s political preference does not have an effect on the way he or she gets treated by the interviewer while appearing on TV. A non-significant relationship
was found between a politician’s political preferences and hostile treatment while being on TV.
When looking at the results in table 4 H4b, which states that using populist rhetoric has a negative influence on the way a politician gets treated by the
interviewer while being on air, needs to be rejected. A non-significant relationship between the two variables was found and therefore it cannot be proven that using populist rhetoric has any effect on how a politician gets treated while being on air.
At last we have to reject H6b, which states that performing badly in the polls has a negative influence on the way a politician gets treated while being on air because a non-significant relationship was found between both variables.
Table 4. Hierarchical multiple regression: mediatization on hostile treatment
Model 1 (n=102) Model 1 (n=102) B SE Beta B SE Beta Constant 1.04 0.24 1.03 0.30 Gender 0.38 0.30 0.17 0.43 0.30 0.19 Party leader - 0.15 0.27 - 0.07 - 0.26 0.29 - 0.13 Party in power 0.17 0.22 0.08 0.40 0.45 - 0.19 Political spectrum - 0.21 0.19 - 0.11 - 0.26 0.21 - 0.14 Conflict frame 0.09 0.1 0.09 Populist rhetoric - 0.02 0.20 - 0.01 Performing badly - 0.31 0.41 - 0.16 Adjusted R2 0.00 - 0.01 F 1.08 0.52 *p<.05
Note: dependent variable = hostile treatment on TV Discussion
The goal of this study was to find out what contextual and content related indicators influence whether and how a politician appears on TV during election time. Findings indicate that the aforementioned personalization, commercialization and mediatization of the media landscape in the Netherlands have influenced the appearance of politicians on Dutch television programs to varying degrees.
No statistical significant influence of personalization on the appearance of politicians on TV was found. Data from a content analysis showed that three
hypotheses regarding personalization of politics had to be rejected. Presumable is that findings were influenced due to the fact that entertainment television programs were excluded from this study. It is known that in such kind of TV shows much more attention is paid to the private lives of politicians (Göran, 2010) and thus, are more likely to contain personalized content.
Significant influence of mediatization on the appearance of politicians was found. Previous literature argues that in a mediatized world there is less party-press parallelism and more media secularization (Magin, 2015). The findings of this study support this claim by showing that a politician’s political preference has no influence on both a politician’s likelihood of appearance and treatment while being on TV. My findings are in line with what Boomgaarden and Vliegenthart argued before, that the Dutch media system is “characterized by a rather autonomously acting press, without partisan alignment” (2009, p. 407). Results show that the studied media do not favour one political affiliation over another, which is an effect of mediatization because this entails the absence of party-press parallelism. One could argue that this equal
treatment regardless of political affiliation is a positive thing for democracy since it indicates that media is secularized and therefore politicians of both left and right do not differ in access to television platforms and that they get treated similarly by the media while being on TV.
The equal treatment of politicians goes beyond the traditional left versus right distinction because findings also show that using populist rhetoric does not influence a politician’s likelihood to appear on TV or the way a politician gets treated while appearing. So even though politicians that are known for using populist rhetoric, like
Trump and Wilders, often accuse the media of portraying them unfairly, in this study no evidence that supports such claim for the situation in the Netherlands was found.
Another effect of mediatization is the fact that being part of a conflict frame has a positive effect on the likelihood of a politician to appear on TV. In line with previous research is proven here that disagreement and conflict are popular features when covering politics on TV (Kepplinger, 2002; Strömbäck, 2008) and that meeting such criteria when appearing increases the likelihood of a politician to get invited and thus appear on TV. The popularity of a conflict frame can be linked to
commercialization as well; since it is expected conflict increases viewership ratings and thus profit.
A third and final clear effect of mediatization on the appearance of politicians on TV can be explained because of horse race journalism (Aalberg et al., 2011). The findings of this study support the claim of its existence in Dutch election coverage because it shows that doing unexpectedly in the polls increases the likelihood of a politician to appear on TV. Spectacularization of political coverage is connected to commercialization since it again it is expected to increase profit.
The biggest influence on the appearance of politicians on TV was found with two non-hypothesized contextual variables; gender and being the party leader. Not surprising, being male significantly increases the likelihood of a politician to appear on TV. This can be linked to the fact that being the leader of a political party does so as well because from the 15 political parties included in this study woman only lead two.
Findings show that leadership is the most important indicator on the likelihood a politician to appear on television during election time. Previous literature argued
that Dutch media favor something best described as an equal access approach when selecting which politicians to cover during election time (Schoenbach, et al., 2001). Like that, the fact that being the leader of a political party proves to be the most important indicator of appearance on TV makes sense. Dutch media provide viewers with a complete overview of the most prominent representatives of all (relevant) political parties and therefore tend to invite political party leaders in almost equal amounts. This equal access approach is most visible when watching Nieuwsuur, which invited party leaders of the 12 biggest political parties to appear on the program all just once in the two weeks leading up to the election.
In the beginning of this paper the importance of optimal political
representation as explained by Pitkin (1967) was mentioned. The fact that being the leader of a political party has the biggest found influence of all variables measured on the likelihood of a politician to appear on TV tells us that Dutch media
(unconsciously) helps create such an optimal political representation in which the voices and opinions of ‘all’ citizens, or at least all political parties, are being heard. Despite claims in literature of increased personalization of our political media
landscape, in the Netherlands only parts of mediatization and commercialization seem to be significantly integrated in the production of political television programs.
Shortcomings
This study has several shortcomings. First, the assumption made in this paper that appearing on television is equivalent to getting invited to appear is often but not always correct. It can be assumed that, especially during election time, politicians particularly want to appear on TV and therefore appear when invited but there are exceptions. In the Netherlands it is known of politicians from both DENK and the PVV that they frequently reject an invitation to appear on TV. Therefore, the
aforementioned assumption should be nuanced when speaking about politicians from these parties.
Another limitation of this study is the fact that to find out whether or not the media treat some politicians worse than others only negative treatment was measured. However, for a politician not getting any positive treatment in relation to a politician who does get a favorable treatment, this is similar to getting treated badly. Not criticizing someone and telling another person his or her policy plans are great does not mean that both politicians have received the same treatment by the media even though this now does seems like it in this study. Unequal treatment between
politicians could indeed have taken place without this coming forward in my findings. Future research
Future research should be testing whether or not my findings can be extended to other countries, contexts and media systems. It would be interesting to compare these results with those in other countries who according to Hallin and Mancini are as well democratic corporatist media systems (2004). Future studies could also include politicians appearing in entertainment TV programs or make a comparison between public state-owned broadcasters and private owned TV channels to see whether the effects of mediatization, personalization and commercialization differ between those, something that I would expect.
One should take away two key findings from this study. First, that the
appearance of politicians on TV is not influenced by a politician’s political preference and second, that leadership was found to be the most important indicator for the likelihood of a politician to appear on TV.
References
Aalberg, T., Strömbäck, J., & De Vreese, C. (2011). The framing of politics as strategy and game: a review of concepts, operationalizations and key findings. Journalism, 13(2), 162-178. DOI: 10.1177/1464884911427799 Alencar, A., & Kruikemeier, S. (2016). Audiovisual infotainment in European news:
A comparative content analysis of Dutch, Spanish, and Irish television. Journalism, 1-18. DOI: 10.1177/1464884916671332
Asp, K., & Esaiasson, P. (1996). The modernization of Swedish campaigns:
individualization, professionalization, and mediatization. In Swanson, D., & Mancini, P., Politics, Media, and Modern Democracy: an international study of innovations in electioral campaigning and their consequences. Westport: Praeger.
Bartholomé, G., Lecheler, S., & De Vreese, C. (2017). Towards a typology of conflict frames: substantiveness and interventionism in political conflict news.
Journalism Studies, 1-23. DOI: 10.1080/1461670X.2017.1299033
Bennett, L. (2012). The personalization of politics: political identity, social media, and changing patterns of participation. The ANNALS of the American Academy of Political and Social Science, 644(1), 20-39. DOI:
10.1177/0002716212451428
Blumler, J. G., & Kavanagh, D. (1999). The third age of political communication: Influences and features. Political Communication, 16(3), 209-230. DOI: 10.1080/105846099198596
Boomgaarden, H., & Vliegenthart, R. (2007). Explaining the rise of anti-immigrant parties: the role of news media content. Electoral Studies, 26(2), 404-417. DOI: 10.1016/j.electstud.2006.10.018
Bos, L., & Brants, K. (2014). Populist rhetoric in politics and media: A longitudinal study of the Netherlands. European Journal of Communication, 29(6), 703-719. DOI: 10.1177/0267323114545709
Bos, L., Van Der Brug, W., & De Vreese, C. (2011). How the media shape
perceptions of right-wing populist leaders. Political Communication, 28(2), 182-206. DOI: 10.1080/10584609.2011.564605
Bos, L., Van Der Brug, W., & De Vreese, C. (2010). Media coverage of right-wing populist leaders. Communications, 35(2), 141-163. DOI: DOI:
10.1515/COMM.2010.008
Boukes, M., & Boomgaarden, H. (2015). Soft news with hard consequences? Introducing a nuanced measure of soft versus hard news exposure and its relationship with political cynicism. Communication Research, 42(5), 701-703. DOI: 10.1177/0093650214537520
Brants K., & Neijens, P. (1998). The infotainment of politics. Political Communication, 15(2), 149-164. DOI: 10.1080/10584609809342363
Braouezec, K. (2016). Identifying common patterns of discourse and strategy among the new extremist movements in Europe: the case of the English Defence League and the Block Identitaire. Journal of Intercultural Studies, 37(6), 637-648. DOI: 10.1080/07256868.2016.1235023
Broh, A. (1980). Horse-race journalism: reporting the polls in the 1976 presidential election. Public Opinion Quarterly, 44(4), 514-529.
Caprara, G. V., & Zimbardo, P. G. (2004). Personalizing politics: a congruency model of political preference. American Psychologist, 59(7), 581-594. DOI:
10.1037/0003-066X.59.7.581
Clinton. H. R. (2017). What Happened. New York: Simon & Schuster.
Ekström, M. (2001). Politicians interviewed on television news. Discourse & Society, 12(5), 563-584. DOI: 10.1177/0957926501012005001
Falasca, K. (2014). Political news journalism: mediatization across three news reporting contexts. European Journal of Communication, 29(5), 583-597. DOI: 10.1177/0267323114538853
Gans, H. (1979). Deciding What’s News. New York: Pantheon Books.
Geers, S., & Bos, L. (2017). Priming issues, party visibility, and party evaluations: the impact on vote switching. Political Communication, 34(3), 344-366.
Göran, E. (2010). Politicians in celebrity talk shows interviews: the narrativization of personal experiences. Text & Talk, 30(5), 529-551. DOI:
10.1515/text.2010.026
Grynbaum, M. (2017, February 17). Trump calls the media the ‘enemy of the American people’. The New York Times. Retrieved from
https://www.nytimes.com/2017/02/17/business/trump-calls-the-news-media-the-enemy-of-the-people.html
Hallin, D. C., & Mancini, P. (2004). Comparing media systems: three models of media and politics. New York, NY: Cambridge University Press.
Harcup, T., & O’Neill, D. (2017). What is news? Journalism Studies, 18(12), 1470-1488. DOI: 10.1080/1461670X.2016.1150193
Hayes, A. F., & Krippendorff, K. (2007). Answering the call for a standard reliability measure for coding data. Communication Methods and Measures, 1(1), 77-89. DOI: 10.1080/19312450709336664
Helfer, L., & Van Aelst, P. (2015). What makes party messages fit for reporting? An experimental study of journalistic news selection. Political Communication, 33(1), 55-79. DOI: 10.1080/10584609.2014.969464
Hjarvard, S. (2008). The mediatization of society: a theory of the media as agents of social and cultural change. Nordicom Review, 29(2), 105-134. DOI:
10.1515/nor-2017-0181
Hopmann, D., Elmelund-Praestekaer, C., Albaek, E., Vliegenthart, R., & De Vreese, C. (2012). Party media agenda-setting: How parties influence election news coverage. Party Politics, 18(2), 173-191. DOI: 10.1177/1354068810380097 Iyengar, S., Norporth, H., & Hahn, K. (2004). Consumer demand for election news:
the horserace sells. The Journal of Politics, 66(1), 157-175. DOI: 10.1046/j.1468-2508.2004.00146.x
Kepplinger, H. (2002). Mediatization of politics: theory and data. Journal of
Communication. 52(4), 972-986. DOI: 10.1111/j.1460-2466.2002.tb02584.x
Kleinnijenhuis, J., Oegema, D., De Ridder, J., & Bos, H. (1995). De democratie op drift. Amsterdam: VU uitgeverij.
Kriesi, H. (2012). Personalization of national election campaigns. Party Politics, 18(6), 825-844. DOI: 10.1177/1354068810389643
Lenguaer, G., Esser, F., & Berganza, R. (2011). Negativity in political news: a review of concepts, operationalizations and key findings. Journalism, 13(2), 179-202. DOI: 10.1177/1464884911427800
Mazzoleni, G. (1987). Media logic and party logic in campaign coverage: the Italian general election of 1983. European Journal of Communication, 2(1), 81-103. DOI: 10.1177/0267323187002001005
Magin, M. (2015). Shades of mediatization: components of media logic in German and Austrian elite newspapers (1949-2009). The International Journal of Press/Politics, 20(4), 415-437. DOI: 10.1177/1940161215595944
Pitkin, H. (1967). The Concept of Representation, Berkeley: University of California. Rahat, G., & Shaefer, T. (2007). The personalization(s) of politics: Israel, 1949–2003. Political Communication, 24(1), 65–80. DOI: 10.1080/10584600601128739
Riffe, D., Lacy, S., & Fico, F. (2014). Chapter 5: Sampling. In Riffe (Eds).), Analyzing Media Messages (96-122). NJ: Erlbaum Publishers.
Schoenbach, K., De Ridder, J., & Lauf, E. (2001). Politicians on TV news: getting attention in Dutch and German election campaings. European Journal of Political Research, 39(4), 519-531. DOI: 10.1111/1475-6765.00586
Schmitt-Beck, R., & Partheymüller, J. (2012). Why voters decide late: a simultaneous test of old and new hypotheses at the 2005 and 2009 German federal elections. German Politics, 21(3), 299-316. DOI: 10.1080/09644008.2012.716042
Strömbäck, J. (2008). Four phases of mediatization: an analysis of the mediatization of politics. The International Journal of Press/Politics, 13(3), 228-246. DOI:
10.1177/1940161208319097
Takens, J., Van Atteveldt, W., Van Hoof, A., & Kleinnijenhuis, J. (2013). Media logic in election campaign coverage. European Journal of Communication, 28(3), 277-293. DOI: 10.1177/0267323113478522
Appendix A Codebook content analysis
Thank you for participating in this research. This codebook is intended to provide an extensive guide regarding the coding process of Dutch politicians in television programs during the 2017 general parliament election campaign. If you have any questions please contact melle.dehaan@student.uva.nl
Eight Dutch television programs have been selected to be part of the coding process. All eight programs contain political guests and discuss politics in a serious and objectively intended way. When looking at the content please try to focus first and foremost on the politician appearing as a guest on the television show. His or hers features and behavior are central to this research.
General information Coder ID:
Text ID: … Date …
Part of which television program is the item you are about to code from? A. De Wereld Draait Door
B. Pauw & Jinek: De Verkiezingen C. Jinek D. Goedemorgen Nederland E. RTL Late Night F. Buitenhof G. WNL Op Zondag H. Nieuwsuur
Contextual variables
1. Gender
Please write down whether the politician appearing in the television program you are watching is a female or a male.
1. Female 2. Male
2. Politician
Please write down the name of the politician appearing in the television show you are watching: ………
3. Party affiliation
Please write down to which political party the politician that is appearing in the program you are watching is a member of.
1. VVD 2. PvdA 3. PVV 4. SP 5. CDA 6. D66 7. Christenunie 8. GroenLinks 9. SGP
11. 50Plus 12. Denk 13. Artikel 1 14. FvD 15. VNL 4. Party leader
Is the politician appearing the leader of his or her political party? A. Yes
B. No
5. Political spectrum
Please write down whether the politician appearing can be identified as more leftwing or rightwing.
A. Leftwing B. Rightwing
6. Polling
Is the politician appearing on TV performing either bad or good on the polls? See way of measuring in table 5 in appendix B.
A. Bad B. Good
Is the politician appearing on TV performing unexpectedly in the polls? See way of measuring in table 5 in appendix B.
A. Yes B. No
8. Party in power
Does the politician appearing belong to a party that is currently in power? Meaning a political party part of the current government.
A. Yes B. No
Content related variables
Conflict frame
9. Does the politician appearing on television visible disagree with the person interviewing him or her?
The clearest way to code this would be when of the two actors states this directly by saying “I disagree with you”, however you can also code this when two people not directly spoken disagree with each other like by visibly reacting irritated for example.
A. Yes B. No
10. Does the politician appearing on television visible disagree with the other guests appearing in the show?
The clearest way to code disagreement between actors would be when of the two actors directly states this by saying “I disagree with you”, however you can also code this when two people unspoken disagree.
A. Yes B. No
C. The politician is the only guest appearing on the TV show.
11. Does the politician appearing on TV discuss disagreeing with a person not present in the show?
A. Yes B. No
Personalization of politics
12. Does the politician appearing on television use colloquial language while appearing on the show?
In other words does he or she use words that are easy to understand? Also it is
important that the politician tries to avoid using jargon in order for you to code yes as your answer.
A. Yes B. No
13. Does the politician appearing on television talk about his or her private life? An example of this is Jesse Klaver (GL) talking about his grandfather during an interview with Humberto Tan on RTL Late Night. In order to code yes it is important that the politician appearing talks about the roles besides the one of politician he or
she fulfills when not being at work (son/daughter, husband/wife, son/daughter, friend, etc.).
A. Yes B. No
14. Does the politician appearing on television discuss non-political content? An example of discussing non-political content is when a politician talks about entertainment news or is asked about his/her opinion on the something not related to proposed policies or political matters.
A. Yes B. No
Hostile treatment
15. Does the interviewer or host criticize the politician or his party? A. Yes
B. No
16. Does the interviewer or host critique the feasibility of the policy plans presented by the politician appearing?
For example by saying that according to the CBS (Dutch statistics bureau) proposed plans and outcomes of a political party are not are not feasible or harmful.
A. Yes B. No
17. Does the politician receives an equal amount of ‘air time’ compared to the other political guests appearing on the TV show
It is possible that the other politician on the TV happens to be interviewed longer and gets more time to answer questions. If this is the case than code yes.
A. Yes B. No
C. There appears only one politician on the show
18. Is the interviewer or host clearly more critical towards this particular guest than towards other politicians appearing on the show?
For example the host keeps asking tough questions with one politician but is not this critical towards others.
A. Yes B. No
C. There appears only one politician on the show
Populist rhetoric
19. Does the politician refer to the existence of ‘a corrupt elite’ or does the politician makes in any way anti-establishment comments?
A. Yes B. No
20. Does the politician refer during the interview to ‘the man on the street’ or ‘the common man’?
This happens for example when the politician appearing states that his or her party is there for ‘the normal man on the street’ or ‘the normal citizen’.
A. Yes B. No
21. Does the politician appearing on TV critiques immigration?
For example by saying that immigration levels are ‘too high’ or ‘out of control’. A. Yes
Appendix B Table. 5 Political parties doing unexpectedly
KP 28.2 TK2012 +/- 25% Doing
VVD 27 41 yes badly
PVDA 12 38 yes badly
PVV 28 15 yes + SP 13 15 no / CDA 17 13 yes + D66 19 12 yes + CU 6 5 no / GroenLinks 13 4 yes + SGP 3 3 no / 50Plus 6 2 yes + PvdD 4 2 yes + DENK 2 2* no / FvD 0 0 no badly VNL 0 0 no badly Artikel1 0 0 no badly
* As being part of the PVDA
Tables 5 illustrates whether or not political parties were doing unexpectedly or badly in the Kantar Public poll of the 28th of February, exactly half-way during the sample frame of this study. This table was created to answer the following questions from this study whether or not the politician appearing belongs to a party doing unexpectedly in the polls (yes = 1, no = 0) and whether or not the politician appearing
belongs to a party doing badly in the polls (yes = 1, no = 0). Scoring one on this final variable “doing badly in the polls” here as well are new political parties that
according to the Kantar Public polling of February 28th would receive zero seats in parliament. Here a political party is doing badly but not unexpectedly compared to 2012 since they were not part of parliament and thus did not own a seat.
Table. 6 Intercoder reliability scores Variables measured
Krippendorff’s alpha
Disagreement with interviewer 0.87
Disagreement with someone not present 0.77
Disagreement with others appearing 1.00
Using colloquial language
0.86 Talking about private life
0.82
Discussing non-political content 0.73
Politician being criticized 0.86
Plans being criticized
0.59 Equal amount of air time
0.87
More critical than with others 1.00
Makes anti-establishment comments 0.64
Refers to the common man 1.00
Critiques immigration
1.00