Media coverage versus individual opinions : does media coverage of partisan polarization affect issue positions or the perception of public opinion polarization?

(1)

Media coverage versus

individual opinions

Does media coverage of partisan polarization

affect issue positions or the perception of public

opinion polarization?

5/2/2016

Author: P.A.L. (Pascalle) van Eerden BSc Student number: 10615857 Supervisor: dr. J.M. (Jonas) Lefevere Master’s Thesis

Graduate School of Communication Master’s programme Communication Science

(2)

(3)

2 ABSTRACT

This study tested whether the finding of Levendusky and Malhotra (2015) that reading about partisan polarization in newspapers caused voters to moderate their own opinions, would hold in a multiparty setting. Overall, the findings of Levendusky and Malhotra (2015) are not supported by this thesis. Participants were randomly assigned to read a news article in one of three experimental conditions: polarized condition, moderate condition and a control group. Using an online survey (N=130), participants placed exemplars and themselves on a political landscape ranging from left to right and from conservative to progressive. Findings showed that political knowledge is an important variable in self-placement on the left/right scale. Furthermore this thesis shows that reading about extreme exemplars in the polarized condition results into disliking the exemplar more and taking greater personal distance to the exemplar, compared to the moderate condition.

INTRODUCTION

A 2009 study indicated that the Dutch public discourse in media coverage has polarized and hardened (IVAbeleidsonderzoek en advies, 2009:11). Nowadays, according to the

newspapers, the Dutch society is becoming even more polarized (e.g. “[…] het toenemend

gepolariseerd klimaat in Nederland”, Algemeen Dagblad, 2015; NRC Handelsblad, 2015).

Such newspapers portray the public as deeply divided. When reading those newspapers one can easily think Dutch society indeed is strongly polarized, but researchers found no proof for such a society (Raad voor Maatschappelijke Ontwikkeling, 2009). This leaves a certain

paradox. How come newspapers portray a polarized society, but when surveying those readers such a deeply divided society does not show up? An American study of Levendusky and Malhotra (2015) might have found the explanation for this paradox, as they found that reading about partisan polarization in newspapers caused voters to moderate their own opinions.

(4)

3 This thesis focuses on the finding of Levendusky and Malhotra that voters moderate their own opinions when reading about polarization: the findings of Levendusky and Malhotra are however not undisputed. A year before their study Ahler (2014) presented findings that would suggest that citizens “consider peers’ positions within public debate when forming their own opinions and adopt slightly more extreme positions as a consequence” (p.607). This is in direct contradiction with the findings from Levendusky and Malhotra. The research approaches of both studies differ, among other things, in used respondents and experimental setting. Therefore it is important to be very precise in the design and execution of this study, since it is proven to have major effects on the outcome.

Moreover, both studies focused on the US. As the American political system is a two party system, and the findings may be affected by American norms and ideals, this thesis investigates the effects of polarization coverage in a European multi-party system. By

applying their propositions to a multiparty system – The Netherlands - we can assess whether their results also hold in different contexts.

Additionally, if citizens’ resources of political information portray the existence of a deeply divided society, then their perceptions of political positions and their beliefs about mass-level polarization may be different from the reality (Ahler, 2014:608). “These

perceptions, in turn, have important consequences for the political behavior of mass publics” (Mutz, 1998:5). Meaning that citizens might not vote during elections or referenda on what is best for themselves, but what they think is best for society as a whole. That would defy the economic idea of rational voting, where it is usually assumed that individuals vote in their own best interest (Akerlof, 1989). In turn, this may affect the way politicians and spin doctors profile their election programs (like the PvdA party resolution (2013): “Van Waarde”, about the party moving from individualism to community, solidarity and togetherness). As such, the

(5)

4 impact of coverage on partisan polarization may have long standing effects on the functioning of democracy.

Therefore the research question of this thesis is: Does media coverage of partisan polarization affect issue positions of Dutch voters and their perception of public opinion polarization?

The remainder of this thesis proceeds as follows. First I expand on the concept of polarization, and how it relates to media use and political attitudes. Then I present the

hypotheses on the effects of perceived polarization on political attitudes and affection. This is followed by an elaboration on the methods of this research and results. Finally I conclude with a discussion.

POLARIZATION, MEDIA AND ATTITUDES

Most scholars hold a notion of polarization “as a bimodal distribution of observations” (Fiorina & Abrams, 2008:566). Movement away from the center toward the extremes seems to be a noncontroversial definition of polarizing. In other words: “Suppose that a population of individuals may be grouped according to some vector of characteristics into "clusters,"

Figure 1

Polarized and nonpolarized distributions

(6)

5 such that each cluster is very "similar" in terms of the attributes of its members, but different clusters have members with very "dissimilar" attributes. In that case we say that the society is polarized” (Esteban & Ray, 1994: 819). For example, suppose that the vector of

characteristics is political orientation, ranging from liberal to conservative, and the attribute of a cluster is the degree of liberalism or conservativism. If society is grouped in two clusters, one liberal and one conservative, and the clusters have very dissimilar attitudes (the upper distribution showed in figure 1), than society is highly polarized. If the clusters would have less dissimilar attitudes (the lower distribution showed in figure 1), then society is less polarized.

As for a multi-party system one can arrange the parties along ideological dimensions (e.g. left-right (Downs, 1957), liberal-conservative (Hotelling, 1929), socio-economic (Karl Marx), GAL-TAN1 (Bakker et al., 2015)). If aligned on one vector, parties near the center of the continuum (see figure 2a) would reflect a society with low polarization (centrist party system). Conversely, a system with all parties at the political extremes is a highly polarized society (see figure 2b) (Dalton, 2008). This corresponds with the models of Fiorina and

1

Green/alternative/libertarian (GAL) to traditional/authoritarian/nationalist (TAN), (Bakker et al., 2015:144)

Figure 2

(7)

6 Abrahms (2008).

The model of a single vector in a multi-party system is difficult to reconcile with reality. Support for parties in multi-party systems is often better explained by the presence of several dimensions (Stokes, 1963). If aligned along two vectors, parties near the center of the axes (see figure 2c) would reflect a society with low polarization. Conversely, a system with all parties at the political extremes is a highly polarized society (see figure 2d). One can imagine that there might also be a combination of models 2a ,b, c and d. Like, if aligned on the vectors, when parties highly oppose on one vector, but are relatively similar to another party (see figure 2e and 2f). Most of the communication studies concerning party polarization are conducted in the US, like Fiorina and Abrahms (2008), whereas the US political system consists of only two parties. In those models of polarization only extreme polarized systems and extreme centrist systems (Dalton, 2008; Fiorina & Abrams, 2008), have been

conceptualized by scientists. Since this study is conducted in a multi-party system it argues for a middle between these two kinds of systems. Therefore it argues for the possibility for the existence of a ‘clustered polarization’ in multi-party systems (see figure 2e and 2f). In that model parties oppose on one vector, but might be similar to another party on the other vector. For example in the Netherlands both the PvdA and GroenLinks are typical left-wing parties and quite similar on the left-right placement, but on the GAL-TAN division they differ (see figure 3 on page 13).

This thesis has no intention of measuring the actual level of polarization in the Dutch society, but focuses on the effects of media coverage suggesting that society is polarized on those vectors. In recent decades the role of media in politics has been often discussed in scholarly debates. A central topic in this ongoing debate is media logic (Mazzoleni, 1987; Strömbäck, 2008; Takens et al., 2013). Media logic refers to “the news values and the storytelling techniques the media make use of to take advantage of their own medium and its

(8)

7 format, and to be competitive in the ongoing struggle to capture people’s attention”

(Strömbäck, 2008: 233). Amongst these storytelling techniques are polarization,

intensification and stereotypization (Strömbäck, 2008:233). The economic market forces news organizations to “view their product as just another means of obtaining advertising revenue, require that the news be presented in a format that has significant entertainment and interest value” (Iyengar, Norpoth & Hahn, 2004: 159). “The human interest stories […] will attract a segment of viewers in search of entertaining stories. But satisfying this demand will lead news outlets to substitute horse race coverage of who is ahead and who is behind for policy

discussions, and will shift the focus to candidates’ personal lives rather than their policy pronouncements.”(Hamilton, 2004: 12). This means that newspapers increasingly use these narratives of polarization to be able to sell their news on the market. With a declining number of newspaper subscribers nationally (Volkskrant, 2012) it is a struggle for newspaper

organizations to remain profitable (RTL Nieuws, 2014). A 2006 study found signs of this media logic in the Dutch media system (Brants & van Praag, 2006), later to be confirmed by Takens et al (2013).

So what does reading about a deeply divided society do to a person? It is known that opinions of others can lead to modifications of one’s judgement (e.g. Sears & Freedman,1965; Moscovici & Zavalloni, 1969). And as Mutz stated: “One might say that mass media may not be particularly influential in telling people what to think, or perhaps even what to think about, but media are tremendously influential in telling people what others are thinking about ...” (1998:5). Levendusky and Malhotra claim that “when citizens read media coverage claiming that the electorate is polarized using exemplars, they will think that voters are more divided on the issues.” (2015: 4). From this, the following hypotheses follows:

(9)

8 H1: Press coverage suggesting the electorate is polarized will increase perceptions of partisan issue polarization compared to press coverage suggesting that the electorate is moderate.

As such the polarization narrative present in the media can affect citizen’s political attitudes. Media “depict polarization as widespread, occurring across many issues, and accompanied by incivility and dislike of the opposition, not simply issue-based disagreement […] lack of compromise and consensus in the contemporary political sphere” and “not via abstract statistics, but through the experiences of particular people” (Levendusky & Malhotra 2015:4). This notion of how media discuss polarization is widely supported (e.g. Ahler, 2014; Mason, 2013; Iyengar, Sood, & Lelkes, 2012; Zillmann, Gibson, Sundar, & Perkins, 1996). As for the Dutch news the same trends are spotted by researchers (Brants & Van Praag, 2006) for exemplars versus the abstract statistics (IVAbeleidsonderzoek en advies, 2009:46) and growing coverage of polarization (Raad voor Maatschappelijke Ontwikkeling, 2009:39).

Reading about individuals (exemplars) who are “divided and take extreme positions, who eschew compromise, and display incivility toward one another” (Levendusky &

Malhotra, 2015:4) contrasts with one’s ideals, norms and values, consequently, disliking these individuals (Harbridge & Malhotra, 2011). The extremism of the exemplars is depicted as “the error of turning a virtue into a vice through dogmatic, single-minded devotion. The preferred alternative is some kind of reasonable balance that tries to satisfy the multiplicity of competing values” (Nelson, Gwiasda & Lyons, 2011:815). Our social values are an important foundation of political attitudes (Nelson, Gwiasda & Lyons, 2011), therefore these exemplars become ‘anti-cues’ for what not to believe (Ledgerwood & Chaiken, 2007; Levendusky & Malhotra, 2015) leading to the expectation that the reader will moderate his or her opinion (Levendusky & Malhotra, 2015). In other words, if a participant reads a news article about supporters of opposing parties who are heavily discussing, without respect and will to

(10)

9 compromise, the participant will dislike both exemplars. Furthermore the participant will try to compensate the gap in moral polarization and moderate his own issue position to the center. Following, this thesis tests attitude extremity to support the second claim, that media coverage of polarization moderates respondents’ issue positions.

H2: Press coverage suggesting the electorate is polarized will make voter’s issue positions less extreme compared to press coverage suggesting that the electorate is moderate.

Hence, this thesis investigates the claim of Levendusky and Malhotra that “because depictions of polarization violate norms of compromise, consensus, and civility, the polarized exemplars will be seen in a more negative light and be taken as examples of their parties more generally” (2015: 5). In other words, while the respondents moderate their issue positions, they tend to dislike the more extreme character they read about (Iyengar & Westwood, 2015).

H3: Press coverage suggesting the electorate is polarized will increase negative affective polarization compared to press coverage suggesting that the electorate is moderate.

The study of Levendusky and Malhotra (2015) was conducted in the US. This study tests whether the finding of Levendusky and Malhotra (2015) would hold in a multiparty setting, like The Netherlands. A comparative study by Hallin and Mancini (2004) shows that party system may affect political communication. The party system of the US is a typical example of the liberal model (Hallin and Mancini, 2004: 10) and the party system of The Netherlands is a typical example of the democratic corporatist model (Brants & Van Praag, 2006: 27). Strömbäck and Dimitrova (2006) find that nations fitting in a democratic

corporatist model frame news items of politics different than nations fitting in a liberal model. Therefore we might expect a difference in the effect of the news article on the public opinion, compared to the study of Levendusky and Malhotra (2015). One of the specific differences between the countries is the number of political parties. The U.S. has only two major political

(11)

10 parties (Democrats (left) and Republicans (right)), whereas twelve parties are represented in Dutch parliament. On the left-right scale four of the Dutch parties, are seen as left (Partij voor de Dieren, SP, GroenLinks and PvdA), four as right (SGP, VVD, CDA and PVV) even though some of these might have left-ish viewpoints on some issues. The rest is not directly classifiable on this left/right scale2. Besides, The Netherlands are depillarized and party membership numbers are low (Lechner, 1996). In the Netherlands only 2% of all eligible voters are members of a political party (NRC, 2015). If polarized press coverage makes voter’s issue positions less extreme and party loyalty runs low, the odds are that after reading polarized press coverage voters may switch to more center-based political parties, which is possible in The Netherlands due to the number of parties. The next hypothesis follows: H4: Press coverage suggesting the electorate is polarized will increase the likelihood that voters switch to a more ‘center’-based party compared to press coverage suggesting that the electorate is moderate.

A last variable must be kept in mind. When exposed to press coverage individuals with weak political attitudes moderate their opinions, while individuals with strong attitudes are less likely to (Leeper, 2014). This might be a possible explanation between the different outcomes of Levendusky and Malhotra (2015) and Ahler (2014), since Ahler’s respondents were sampled from a political survey and are therefore more likely to be extreme and interested in politics and behave differently (Malhotra & Krosnick, 2007; Ansolabehere & Schaffner, 2014). Therefore this research tests for the moderating role of political knowledge among the respondents.

H5: Subjects which have less political knowledge and are exposed to press coverage

suggesting the electorate is polarized will make voter’s issue positions less extreme compared to subjects which have more political knowledge.

2

(12)

11 METHOD

To answer the research question, an online experimental study was conducted to test whether voters moderate their own opinions, using three different experimental conditions.

Participants were invited to participate through Facebook and e-mail and included colleagues, friends and friends of friends, thus collecting 130 responses. The first condition was the polarized condition, the second was the moderated condition and the third was the control condition. Subjects were randomly assigned to read a newspaper article in one of these three conditions followed by answering questions to measure the variables of interest. After completing the survey, participants were informed of the true purpose of the study.

From the original data all non-Dutch citizens, non-eligible voters and unfinished surveys were removed, leaving 126 valid responses. Almost two third of the subjects were female (61.9%), the average age of the participants was 27 years old (SD=11 years and 8 months) and the population was highly educated (71.4% went to university). The distribution of education, age and gender was evenly divided amongst the three experimental groups.

The online study (via Qualtrics) was spread to all who were willing to cooperate. Comparative research on web-based studies showed that report questionnaires from self-selected samples, are relatively diverse with respect to gender, socio-economic status, geographic region, and age. Although Internet samples are “not representative of the population at large, they are generally more diverse than samples published in a highly selective psychology journal” (Gosling et al., 2004:99). The sample is not representative of the Dutch voting aged population, most of all because this thesis came with limitations on time and funding. Therefor it will be impossible to draw conclusions for society as a whole. This thesis can however generalize more about the Dutch higher educated.

Depending on the condition, respondents were shown different versions of a fictitious news article. The polarized news article included the five themes Levendusky and Malhotra

(13)

12 encountered in their content analysis: depict polarization as widespread, occurring on more than one issue, dislike of the opposition, perception of lack of consensus in the community and used exemplars. The moderate news article showed polarization as not existing over more than one issues, showed no dislike of the opposition, no lack of consensus in the community, but does use exemplars. So the two articles have the same structure. The news articles

discussing Dutch political polarization were written by the researcher to be similar to those in real national newspapers. The news articles were presented as being extracted from a national newspaper for ecological validity. The subjects in the control condition read an article about the premiere of the latest James Bond movie, being a real article copied the ‘Algemeen Dagblad’. To check the ecological validity, subjects were finally asked how often they saw these kind of news articles in real life on a 5-pointscale, ranging from never (1) to

continuously (5). The polarization and moderate stimuli groups pointed out they saw these kind of articles ‘sometimes’ (35.1%) and ‘regularly’ (21.6%) as for the control group (respectively 55.3% and 31.6%).

The news article contains four natural opposing parties, chosen on their place in the Dutch political landscape. To make the polarized news article as realistic as possible ideally one natural left-progressive, one natural right-progressive, one right-conservative and one left-conservative party would have been chosen (as in figure 2d). To be able to appoint four natural opposing parties, the Chapel Hill Expert Survey 2014 (published in 2015) was consulted. The Chapel Hill Expert Survey questions subjects about ideological positions of political parties throughout Europe, including questions about the GAL-TAN and left-right division. The question used to measure the left-right division was: “Please tick the box that best describes each party's overall ideology on a scale ranging from 0 (extreme left) to 10 (extreme right).” The question used to measure the GAL-TAN division was: “Parties can be classified in terms of their views on democratic freedoms and rights. “Libertarian” or

(14)

13 “postmaterialist” parties favor expanded personal freedoms, for example, access to abortion, active euthanasia, same-sex marriage, or greater democratic participation. “Traditional” or “authoritarian” parties often reject these ideas; they value order, tradition, and stability, and believe that the government should be a firm moral authority on social and cultural issues.”

Subjects were asked to place the parties again on an 11-point scale ranging from 0 (Libertarian/ Postmaterialist) to 10 (Traditional/ Authoritarian). Extracting the Dutch parties (N=120) the following placements were calculated (see figure 3). Since the Dutch political landscape has no left-conservative party, two different right-conservative parties have been chosen, being SGP (very conservative) and VVD (very right) (see figure 3). Leaving the SP to be the most natural ‘extreme’ left-progressive party, D66 the only natural right progressive party, VVD the most natural right-winged party and the SGP the most natural conservative party. The news article presents exemplars out of these four parties.

The researcher chose the ‘Algemeen Dagblad’ because it is the second largest Dutch newspaper and has no political or religious binding, whereas the ‘Telegraaf’ (the largest

Figure 3

(15)

14 Dutch newspaper) is locally known for its right wing populism (Akkerman, 2011) and

therefore less suitable for this research. After reading the news article, to test the first

hypothesis, the subjects were asked to place the four exemplars out of the news article on two single vectors: one left/right, one progressive/conservative on 11-point scales ranging from totally left (0) to totally right (10), and progressive (0) to conservative (10). These vectors were chosen because they are probably the most known under the Dutch population and used by scholars investigating the Dutch party system (Bakker et al., 2015). Voting Advice

Applications like Kieskompas (which was used 1,3 million times in the Dutch elections of 2012 (AD, 2012)), uses the left/right and progressive/conservative division, thereby

constructing the progressive/conservative using the GAL-TAN division (Bakker et al., 2015). The operationalization of perceived polarization in this thesis is the distance between the placed parties and the axes as shown in figure 4. While answering the survey questions the researcher provided the subjects with the operationalization of left, right, progressive and conservative to ensure internal validity. Next the subjects were asked to place themselves on

Figure 4

(16)

15 the same two single vectors to test H2. After that the subjects were asked about their attitudes and feelings towards the exemplars to test H3. To test H3, subjects rated their feelings

towards the exemplar on a scale from zero to hundred, whereas zero was total negativism and hundred was a total positivism. Scoring characters fifty meant having no feelings towards the exemplar at all. Additionally subjects were asked to what extent they recognized themselves in the news article exemplars on a four-point scale ranging from (1) totally no recognition to (4) very recognizable.

Afterwards the subjects were asked “If there were national elections today, what party would you vote for?” Answer options were: (1) VVD, (2) PvdA, (3) PVV, (4) SP, (5), CDA, (6) D66, (7) ChristenUnie, (8) GroenLinks, (9) SGP, (10) PvdD, (11) 50Plus, (12) Other, namely …, (13) I did not vote, (14) I voted blank or (15) I don’t remember. According to the researcher’s earlier calculations (see figure 3) and based on the fourth hypothesis, subjects with the polarized stimuli would be more likely to choose a center-based party like CDA, 50Plus of PvdA, than the subjects with the moderate stimuli.

Furthermore the subjects were tested on their political knowledge, previous voting behavior, party commitment and attitudes towards policy issues. To be able to control the data for issue saliency, subjects were asked to arrange ten issues in an order from one to ten on their importance, since this experiment would likely have less effect if the subjects have strong prior opinions about the issues.

RESULTS

During the survey participants answered four questions testing their political

knowledge. These questions were recoded into wrong or right leaving an average of 2.89 out of four correct answers (SD=1.16). As a manipulation check subjects were asked whether the individuals in the news article were agreeing or disagreeing with one and another. Subjects in

(17)

16 the polarized condition (91.1%) and the moderate condition (68.4%) answered overall correct. Participants were also asked whether they could remember in which newspaper the article was printed. The polarized group 62.2% were correct the most, compared to the moderate group (57.9%) and the control group (36.8%). Why the control group scored that low remains unknown.

To test whether polarizing coverage indeed led to greater perceived polarization compared to the moderate coverage (H1), subjects rated the exemplars they read about on a eleven point scale (varying from 0 being extreme left to 10 extreme right and from 0 being extreme conservative to 10 extreme progressive). Rating them 5 meant neutral. The results are shown in table 1 and 2. It shows that there is very little difference between conditions in terms of the perception of the position of the typical political exemplars. The largest differences are found at the typical SGP voter and the typical SP voter. The first hypothesis expected that the polarized condition would give more polarized placements compared to the moderate

condition. An one-way ANOVA (Field, 2009:375) was conducted to compare the effect of the stimuli on the perceived polarization. When testing for H1 within the perceived left-right and conservative-progressive polarization only one significant outcome showed. There was a very weak significant effect of the stimuli on the perceived left-right polarization with the SGP party. Post hoc comparisons using the Bonferroni test (p<0.1) indicated that the mean score for the control condition on the perceived left-right polarization with the SGP was

significantly different from the polarized condition (Mdifference=1.13, p=.052) and the moderate

condition (Mdifference=1.06, p=.078). However, the polarized condition did not significantly

differ from the moderate condition (Mdifference=-0.07, p=1.000).

Because political knowledge may affect the placement (Leeper, 2014), political knowledge was added as covariate. Using an univariate analysis of variance, the covariate, political knowledge, was weakly significant related to the perceived left-right polarization

(18)

17 Table 1

Perceived Left-Right placement

N M SD SP [F (2, 116)= 0.20, p=.822, 2 = .00] Polarization stimuli 40 2.78 1.68 Moderate stimuli 39 2.59 1.68 Control group 40 2.83 1.89 VVD [F (2, 116)= 0.25, p=.778, 2 = .00] Polarization stimuli 40 8.70 1.47 Moderate stimuli 39 8.97 1.91 Control group 40 8.93 2.08 SGP* [F (2, 116)= 3.65, p=.029, 2 = .06] Polarization stimuli 40 6.60 2.16 Moderate stimuli 39 6.67 2.02 Control group 40 7.73 2.08 D66 [F (2, 116)= 0.04, p=.958, 2 = .00] Polarization stimuli 40 6.40 1.46 Moderate stimuli 39 6.49 1.50 Control group 40 6.40 1.58

*Note: the mean difference is significant at the 0.05 level.

Table 2

Perceived Conservative-progressive placement

N M SD SP [F (2, 116)= 1.06, p=.351, 2 = .02] Polarization stimuli 40 6.50 2.36 Moderate stimuli 39 6.77 2.39 Control group 40 7.25 2.26 VVD [F (2, 116)= 0.54, p=.584, 2 = .01] Polarization stimuli 40 6.05 1.89 Moderate stimuli 39 5.82 1.88 Control group 40 5.60 2.04 SGP [F (2, 116)= 0.20, p=.818, 2 = .00] Polarization stimuli 40 2.33 1.83 Moderate stimuli 39 2.44 1.70 Control group 40 2.20 1.40 D66 [F (2, 116)= 0.25, p=.782, 2 = .00] Polarization stimuli 40 8.13 1.84 Moderate stimuli 39 8.08 2.06 Control group 40 8.35 1.64

(19)

18 with the SGP [F (1, 119) = 23.39, p = .000, 2 = .22] and the significant effect of the stimuli after controlling for the effect of political knowledge [F (2, 115)= 4.38, p=.0.015, 2 = .22] grew stronger on the perceived polarization. The results do not support H1, therefore we can reject H1.

H2 expected that in the polarized condition subjects would rate themselves less polarized than subjects out of the moderate condition. Subjects rated themselves after exposure to the stimuli on a eleven point scale, varying from 0 being extreme left to 10 extreme right and from 0 being extreme conservative to 10 extreme progressive. Rating themselves 5 meant neutral. The outcome is shown in table 3. A one-way ANOVA was conducted to compare the effect of the stimuli on the placement. There was no significant effect of the stimuli on the self-placement on either the left-right scale or the conservative- progressive scale. To test whether

the results were obscured by lower scores reflecting more extreme positions (on both the left and right hand side), the variable was recoded. The eleven point scale was recoded in a 0 to 5 scale, where 5 remained neutral and where 10 and 0 became extreme positions. 9 was recoded in 1, 8 was recoded in 2, 7 was recoded in 3 and 6 was recoded in 4 (see table 4 for results). A one-way ANOVA was conducted to compare the effect of the stimuli on this self-placement. There was a weak significant effect of the stimuli on the self-placement on the conservative- progressive scale. Post hoc comparisons using the Bonferroni test (p<0.1) indicated that the

Table 3

Self placement

N M SD

Left-Right

[F (2, 111)= 1.30, p=.277, 2 = .02] Polarization stimuli _{Moderate stimuli} 38 ₃₈ 6.50 _6.77 2.36 _2.39

Control group 38 7.25 2.26 Conservative-Progressive [F (2, 111)= 1.01, p=.367, 2= .02] Polarization stimuli 38 6.05 1.89 Moderate stimuli 38 5.82 1.88 Control group 38 5.60 2.04

(20)

19 mean score for the moderate condition on the perceived conservative- progressive was

significantly different from the polarized condition (Mdifference=1.11, p=.059) and the control

condition (Mdifference=1.05, p=.079). However, the polarized condition did not significantly

differ from the control condition (Mdifference=-0.05, p=1.000). Therefore H2 may be rejected as

well.

To test H3, which expected that subjects tend to dislike the more extreme character, subjects were asked how they felt towards the article characters. They rated their feelings towards the exemplar on a scale from 0 to 100, whereas 0 was total negativism and 100 was a total positivism. Scoring characters 50 meant having no feelings towards the exemplar at all. The results are shown in table 5. Striking is that the feelings from the polarization stimuli on three out of the four characters are negative (below 50), compared to only one from the moderate stimuli. A one-way ANOVA was conducted to compare the effects of the stimuli on the (dis-)liking of the exemplars. On three out of four of the exemplars do subjects out of the polarization stimuli dislike them significantly more than the subjects out of the moderate stimuli, although the relation is weak ton not existing in all three significant cases. Out of the 85 participants (in the experimental conditions moderate and polarized) 21 voted VVD and 36 voted D66 in the last elections. Therefore it is not unlikely that that respondents rate their own preferred party systematically higher. Whether they voted for this party last elections was

Table 4

Extreme placement

N M SD

Left-Right extremism

[F (2, 111)= 1.01, p=.369, 2= .02] Polarization stimuli _{Moderate stimuli} 38 ₃₈ 2.84 _3.26 1.22 _1.78

Control group 38 3.26 1.43 Conservative-Progressive extremism* [F (2, 111)= 3.56, p=.032, 2= .06] Polarization stimuli 38 2.34 1.32 Moderate stimuli 38 3.45 2.96 Control group 38 2.39 1.41

(21)

20 added as the control variable: party preferences. Performing a multivariate analysis, the covariate party preferences, was not significantly related to liking or disliking the SP voter [F (1, 44) = 0.34, p=.562, 2=.10], the SGP voter [F (1, 44) = 0.24, p=.625, 2=.28] and the VVD voter [F (1, 44) = 0.254, p=.617, 2=.08], but was weakly significantly related to liking or disliking the D66 voter [F (1, 44) = 6.265, p=.016, 2=.14]. After controlling for the effect of previous party preferences there was still a significant (p<.05) effect of the stimuli on the liking or disliking the SP, SGP and VVD exemplars, but not on the D66 exemplar.

Additionally subjects were asked to what extent they recognized themselves in the news article exemplars on a four-point scale ranging from (1) totally no recognition to (4) very recognizable. Results (see table 6) show that overall subjects out of the polarization stimuli tend to take a greater personal distance to the news characters than the subjects out of the moderate stimuli. Again to control, party preferences was added as control variable, since it is possible that respondents recognize themselves more in their preferred party representative. Performing a multivariate analysis, the covariate, party preferences, was not significantly

Table 5

(Dis-)liking the news article characters

N M SD

Mariken (the SP voter)* [F (1, 55) = 5.16, p =.027, 2= .09]

Polarization stimuli 32 40.65 28.42 Moderate stimuli 25 56.72 23.78 Daniël (the SGP voter)*

[F (1, 61) = 14.26, p=.000, 2= .19]

Polarization stimuli 37 22.49 19.34 Moderate stimuli 26 41.85 20.99 Geert (the VVD voter)*

[F (1, 62 = 6.61, p=.013, 2= .01]

Polarization stimuli 36 46.61 2049 Moderate stimuli 28 59.36 18.47 Emma (the D66 voter)

[F (1, 56) = 0.72, p =.399, 2= .10]

Polarization stimuli 32 60.34 24.80 Moderate stimuli 26 65.15 16.29 *Note: the mean difference is significant at the 0.05

(22)

21 Table 6 Self-recognition N M SD

Mariken (the SP voter) [F (1, 77)= 2.64, p=.109, 2= .03]

Polarization stimuli 40 2.22 0.89

Moderate stimuli 39 2.54 0.82

Daniël (the SGP voter)* [F (1, 77)= 12.78, p=.001, 2= .14]

Geert (the VVD voter)** [F (1, 77)= 3.55, p=.063, 2= .04]

Emma (the D66 voter) [F (1, 77)= 0.98, p=.324, 2= .01]

Moderate stimuli ₃₉ _2.92 _0.62

*the mean difference is significant at the 0.05 level. **the mean difference is significant at the 0.1 level.

Table 7

Party voting by participants (frequencies)

Last elections If election were ‘today’ Polarization stimuli Moderate stimuli Polarization stimuli Moderate stimuli VVD 12 9 10 8 PvdA 4 4 3 1 PVV 1 0 1 0 SP 1 0 0 0 CDA 0 3 1 2 D66 18 18 7 12 ChristenUnie 3 2 2 1 GroenLinks 2 2 5 4 SGP 0 1 0 0 PvdD 0 1 0 2 I don’t know ₂ 2 9 8 N 43 42 38 38

(23)

22 related to liking or disliking the SP voter [F (1, 44)= 0.63, p=.429, 2= .04], the SGP voter [F (1, 44)= 1.03, p=.312, 2= .15], the VVD voter [F (1, 44)= 0.53, p=.468, 2= .05], but was weakly significantly related liking or disliking the D66 voter [F (1, 44)= 6.266, p=.014, 2= .09]. There was no change in the significance (p<.05) on the effect of the stimuli after controlling for the effect of previous party preferences on relating to the exemplars. Thereby partially confirming H3.

To be able to test for party shifting (H4), the subjects were asked what they voted for in the last national elections and to what extent they are loyal to that party. 62.4% of the subjects stated to vote more often for the same party, compared to always voting for the same party (9.6%) and or no regular party voting (19.2%). After the stimuli the subjects were asked again what they would vote, but now for if there were elections ‘today’ (see table 3). 45.6% of the subjects switched party favor, almost evenly divided over the three experimental groups. Therefore the stimuli itself were no trigger for party shifting, but it may effect to which party the subject might shift. Table 3 shows that the VVD, SP, D66 and the SGP, all parties in the stimuli, are less likely to obtain the subjects votes in new elections. Striking feature of this table is the even more rapid decrease of votes after exposure to the polarization stimuli. Since a lot of the respondents stated not to know what they would vote for in next elections, there is very little data to confirm the fourth hypothesis. A variable voting behavior was created coding for either (0) being loyal to the same party or switching to another party to vote for (1). Subjects in the polarized group (M=0.42, SD=0.50) were the most loyal compared to the moderate group (M=0.47, SD=0.51) and the control group (M=0.47, SD=0.51). A one-way ANOVA was conducted to compare the effect of the stimuli on voting behavior. There was no significant effect found of the stimuli on voting behavior [F (2, 111)= 0.14, p=.871, 2= .00]. Therefore H4 is rejected.

(24)

23 This thesis expected (H5) that subjects with low political knowledge within the

polarized stimuli group, would rate themselves as less polarized than subjects with high political knowledge (see table 8). A one-way ANOVA was conducted to compare the effect of the political knowledge on the self-placement. There was a weak significant effect of the political knowledge on the self-placement on the left-right scale. Post hoc comparisons using the Bonferroni test indicated that the mean ‘self-placement on the left-right scale’ for the highest score on political knowledge was significantly different from the second highest score for political knowledge (Mdifference=1.66, p=.003). Since a mean of 6 means neutral we cannot

speak of difference in polarization, because the average self-placement of 5.26 means slightly left and 6.91 means slightly right. Therefore all that is proven here is a difference in left-right self-placement, although it is a weak relation. The lower scores 0 out of 4 correct

(Mdifference=1.63, p=1.000), 1 out of 4 correct (Mdifference=0.85, p=1.000) and 2 out of 4 correct

(Mdifference=0.41, p=1.000) did not significantly differ in the self-placement. Noticeably, when

the stimuli groups were entered as covariates to political knowledge in a multivariate analysis, they were almost significantly (p<.10) related to the self-placement left-right [F (1, 108)=2.73

Table 8

Political knowledge on self-placement

N M SD Left-Right [F (4, 109)= 3.67, p=.008, 2 = .12] 0 out of 4 correct 4 5.75 3.78 1 out of 4 correct 9 6.11 2.09 2 out of 4 correct 20 6.50 1.64 3 out of 4 correct 35 5.26 2.12 4 out of 4 correct 46 6.91 1.79 Conservative-Progressive [F (2, 111)= 1.01, p=.367, 2= .03] 0 out of 4 correct 4 8.00 2.45 1 out of 4 correct 9 7.89 1.05 2 out of 4 correct 20 7.37 1.67 3 out of 4 correct 35 7.76 1.70 4 out of 4 correct 46 7.52 1.92

(25)

24 , p=.101, 2= .10]. There was a stronger significant effect of political knowledge after

controlling for the effect of the stimuli on self-placement on the left-right scale [F (4, 108)=3.82, p=.006, 2= .10]. But based on these finding, this study cannot confirm H5.

CONCLUSION AND DISCUSSION

This study is not the first to illustrate the political consequences of media coverage of partisan polarization (Levendusky & Malhotra, 2015; Ahler , 2014). But where there has been a debate about the levels and changes of polarization in the US (Abramowitz & Saunders, 2008;

Fiorina & Abrams, 2008), no research has yet tried to explore the media’s role in describing a country as deeply divided along partisan lines in a multiparty system. The aim of the thesis was to see whether the finding of Levendusky and Malhotra (2015) would hold in a European multi-party system. Using an online survey (N=126), this research shows very little support for the finding of Levendusky and Malhotra (2015). It finds that depictions of a highly divided population transmitted through the mass media, like newspapers do not increase perceived polarization. Therefore polarized media coverage does not moderates issue positions or alter party preferences, but it does increase affective polarization.

These findings seem to conflict with the article of Levendusky and Malhotra (2015), but also with a recent article by Ahler (2014). Levendusky and Malhotra (2015), show that polarized media coverage transmitted through the mass media can increase perceived polarization, moderates issue positions and increases affective polarization. Which in turn conflicts with Ahler (2014) who finds that providing people with information about a divided electorate causes people to become more extreme and more polarized.

There are several differences between their studies and this thesis that could account for these differences. First, the sample population is different in all three studies. Ahler (2014) studied samples of people who opted into completely political surveys. It is possible that these

(26)

25 people are more interested in politics and therefore behave differently from people who are not as interested in politics (Malhotra & Krosnick, 2007: 293). Besides this thesis shows that political knowledge affects self-placement on the left-right division. Therefore this thesis might support Levendusky and Malhotra (2015) in their claim that their findings differ from Ahler (2014) because of the sample used (2015:15). Levendusky and Malhotra conducted an online experimental study with a representative probability sample of the US population. Hence, it is possible that the difference in outcome is due to the highly skewed sample population presented in this thesis, with high scores on political knowledge and education.

Another possible explanation for the lack of significant outcomes in this thesis is the excessive presence of D66 and VVD voters amongst the sample population. The D66 and VVD parties were used in the stimuli alongside the SP and SGP. D66 and VVD voters might have known where the party ‘officially’ belongs in the Dutch political landscape, therefore ignoring the stimuli. The SP and SGP voters were not excessively presence amongst the participants, hence the greater difference in means (see table 1 and 2).

This study shows that polarized media coverage increases affective polarization. Reading about extreme characters in news articles results into disliking them and distancing yourself from the exemplar. Even after controlling for party preferences three out of the four cases showed significant differences in disliking. The only case not being significantly different was the (dis)liking of the D66 exemplar. But this might be related to the excessive presence of D66 voters (35,7% of which 75,0% are loyal voters).

An interesting case for further future in-depth research would be replacing the exemplars for politicians, since political debates can be “polarized, with an added emotional and moral aspect”(MacGregor, 2013:225). A recent Dutch study argues that “due to the mediatization of politics […] journalists use politicians’ quotes and sound bites” (Scholaus, 2013:502). These quotes can be used to support conflict frames of news items (Scholaus,

(27)

26 2013:509). These one-liners in conflict frames might be comparable with the news articles used in this thesis. Reading one-liners of politicians in news articles might also result into disliking those politicians and distancing yourself from them. For example, the Dutch politician Geert Wilders, even though he heavily discusses, occasionally without the necessary respect and will to compromise, his support grows, which contradicts to the findings of this thesis.

The big question whether polarized media coverage could alter party preferences, could not be confirmed. This might also be due to the fact that the last Dutch national elections were in 2012. It has been a few years since the last elections, therefore the party preference might also have changed due to natural causes outside of the experiment.

Despite all the limitations of this thesis and differences in the method and sample population compared to Levendusky and Malhotra (2015) and Ahler (2014), these findings on perceived polarization differ from the more conventional findings. That said, these results add a new wrinkle to that larger debate. What if this effect does not occur in multiparty systems? Although it is not proven that this effect does not occur in Europe at all, this thesis shows that the findings of Levendusky and Malhotra (2015) and Ahler (2014) do not hold amongst Dutch educated citizens.

(28)

27 REFERENCES

Aalberg, T., Van Aelst, P., & Curran, J. (2010). Media systems and the political information environment: A cross-national comparison. The International Journal of Press/Politics, 15(3), 255-271.

Abramowitz, A., & Saunders, K. (2008). Is polarization a myth? Journal of Politics, 70(2), 542–555.

Ahler, D. J. (2014). Self-Fulfilling Misperceptions of Public Polarization. The Journal of Politics, 76(03), 607-620.

Akerlof, G. A. (1989). The Economics of Illusion, Economics & Politics, 1(1), 1-15.

Akkerman, T. (2011). Friend or foe? Right-wing populism and the popular press in Britain and the Netherlands. Journalism, 1464884911415972.

Algemeen Dagblad (12-09-2012), “Kiezers surfen massaal naar stemhulpen.” Algemeen Dagblad (18-02-2015), “Koenders trekt visa van omstreden imams in.”

Ansolabehere, S., & Schaffner, B. F. (2014). Does survey mode still matter? Findings from a 2010 multi-mode comparison. Political Analysis, 22(3), 285-303.

Bakker, R., De Vries, C., Edwards, E., Hooghe, L., Jolly, S., Marks, G., ... & Vachudova, M. A. (2015). Measuring party positions in Europe The Chapel Hill expert survey trend file, 1999–2010. Party Politics, 21(1), 143-152.

Brants, K., & Van Praag, P. (2006). Signs of media logic half a century of political communication in the Netherlands. Javnost-the public, 13(1), 25-40.

Dalton, R. J. (2008). The quantity and the quality of party systems party system polarization, its measurement, and its consequences. Comparative Political Studies, 41(7), 899-920.

Downs, A. (1957). An economic theory of political action in a democracy. The journal of p Harold Hotelling, "Stability in Competition,"

Esteban, J. M., & Ray, D. (1994). On the measurement of polarization. Econometrica: Journal of the Econometric Society, 819-851.

(29)

28 Field, A. P. (2009). Discovering statistics using SPSS (and sex and drugs and rock ‘n’roll). Los Angeles [ie Thousand Oaks, Calif.

Fiorina, M. P., & Abrams, S. J. (2008). Political polarization in the American public, Annual Review of Political Science, 11, 563-588.

Gosling, S. D., Vazire, S., Srivastava, S., & John, O. P. (2004). Should we trust web-based studies? A comparative analysis of six preconceptions about internet questionnaires. American Psychologist, 59(2), 93.

Hallin, D. C., & Mancini, P. (2004). Comparing media systems: Three models of media and politics. Cambridge University Press.

Hamilton, J. (2004). All the news that's fit to sell: How the market transforms information into news. Princeton University Press.

Harbridge, L., & Malhotra, N. (2011). Electoral incentives and partisan conflict in congress: Evidence from survey experiments. American Journal of Political Science, 55(3), 494-510. Hotelling, H., (1929) Stability in Competition, Economic Journal, 39, 41-57.

IVA beleidsonderzoek en advies, “Polarisatie en radicalisering in Nederland: Een verkenning van de stand van zaken in 2009.”, 2009.

Iyengar, S., Norpoth, H., & Hahn, K. S. (2004). Consumer demand for election news: The horserace sells. Journal of Politics, 66(1), 157-175.

Iyengar, S., Sood, G., & Lelkes, Y. (2012). Affect, not ideology a social identity perspective on polarization. Public Opinion Quarterly, 76(3), 405-431.

Iyengar, S., & Westwood, S. J. (2015). Fear and loathing across party lines: New evidence on group polarization. American Journal of Political Science, 59(3), 690-707.

Lechner, F. J. (1996). Secularization in the Netherlands?. Journal for the scientific study of religion, 252-264.

Ledgerwood, A., & Chaiken, S. (2007). Priming us and them: automatic assimilation and contrast in group attitudes. Journal of personality and social psychology, 93(6), 940.

(30)

29 Leeper, T. J. (2014). The informational basis for mass polarization. Public opinion quarterly, 78(1), 27-46.

Levendusky, M., & Malhotra, N. (2015). “Does Media Coverage of Partisan Polarization Affect Political Attitudes?”, Political Communication, (ahead-of-print), 1-19.

Malhotra, N., & Krosnick, J. A. (2007). The effect of survey mode and sampling on inferences about political attitudes and behavior: Comparing the 2000 and 2004 ANES to Internet surveys with nonprobability samples. Political Analysis, 15(3), 286-323.

Mason, L. (2013). The Rise of Uncivil Agreement Issue Versus Behavioral Polarization in the American Electorate. American Behavioral Scientist, 57(1), 140-159.

MacGregor, S. (2013). Barriers to the influence of evidence on policy: Are politicians the problem?. Drugs: education, prevention and policy, 20(3), 225-233.

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.

Moscovici, S., & Zavalloni, M. (1969). The group as a polarizer of attitudes. Journal of personality and social psychology, 12(2), 125.

Mutz, D.C., (1998) Impersonal Influence: How Perceptions of Mass Collectives Affect Political Attitudes. Cambridge: Cambridge University Press.

Nelson, T. E., Gwiasda, G., & Lyons, J. (2011). Vilification and values. Political Psychology, 32(5), 813-835.

NRC (14-03-2015), “De onderliggende partij”.

NRC Handelsblad (08-01-2015), “'Dit stelt hoge eisen aan politiek'; Opdracht aan de politiek zelf: deëscaleren.”

Parlement & Politiek, http://www.parlement.com/id/vh8lnhrp8wsy/links_en_rechts, bezocht op 25-09-2015.

Partij van de Arbeid (PvdA)(2013). “VAN WAARDE, sociaal-democratie voor de 21ste

eeuw” Prior, M. (2013). Media and political polarization. Annual Review of Political Science, 16, 101-127.

(31)

30 Raad voor Maatschappelijke Ontwikkeling, “Polarisatie. Bedreigend en verrijkend” , 2009. Ryan Bakker, Erica Edwards, Liesbet Hooghe, Seth Jolly, Gary Marks, Jonathan Polk, Jan Rovny, Marco Steenbergen, and Milada Vachudova (2015), 2014 Chapel Hill Expert Survey, Version 2015.1. Available on chesdata.eu. Chapel Hill, NC: University of North Carolina, Chapel Hill.

RTL Nieuws (13-10-2014), Reorganisatie Sanoma na krimp bladenmarkt

Sears, D. W., & Freedman, J. L. (1965). Effects of expected familiarity with arguments upon opinion change and selective exposure. Journal of Personality and Social Psychology, 2(3), 420.

Schohaus, B. (2013). THE PERFECT CUT: The use of quotes in interpretation strategies on Dutch television news. Journalism Practice, 7(4), 502-516.

Stokes, D. E. (1963). Spatial models of party competition. American Political Science Review, 57(02), 368-377.

Strömbäck, J. (2008). Four phases of mediatization: An analysis of the mediatization of politics. The International Journal of Press/Politics, 13(3), 228-246.

Strömbäck, J., & Dimitrova, D. V. (2006). Political and Media Systems Matter A Comparison of Election News Coverage in Sweden and the United States. The Harvard International Journal of Press/Politics, 11(4), 131-147.

Takens, J., Atteveldt, W. van, Hoof, A. van, & Kleinnijenhuis, J. (2013). Media logic in election campaign coverage. European Journal of Communication, 28(3), 277–293.

Takens, J., van Atteveldt, W., van Hoof, A., & Kleinnijenhuis, J. (2013). Media logic in election campaign coverage. European Journal of Communication, 0267323113478522. Vergeer, M., & Hermans, L. (2013). Campaigning on Twitter: microblogging and online social networking as campaign tools in the 2010 general elections in the Netherlands. Journal of Computer‐Mediated Communication, 18(4), 399-419.

(32)

31 Zillmann, D., Gibson, R., Sundar, S. S., & Perkins, J. W. (1996). Effects of exemplification in news reports on the perception of social issues. Journalism & Mass Communication Quarterly, 73(2), 427-444.

FIGURE SOURCES

Figure 1: “Polarized and nonpolarized distributions” Source: Fiorina & Abrams (2008:566)

LIST OF ABBREVIATIONS

SP, Socialistische Partij D66, Democraten 66 PvdA, Partij van de Arbeid PvdD, Partij voor de Dieren

CDA, Christen-Democratisch Appèl

VVD, Volkspartij voor Vrijheid en Democratie SGP, Staatkundig Gereformeerde Partij

PVV, Partij Voor de Vrijheid CU, ChristenUnie

GL, GroenLinks 50Plus, 50Plus Partij

(33)

32 QUESTIONAIRE

INTRODUCTION AND INFORMED CONSENT

Bedankt om aan deze enquête te willen deelnemen. De vragenlijst gaat over de Nederlandse politiek en beeldvorming.

Er worden onder andere vragen gesteld over uw politieke voorkeur. Alle informatie die u verstrekt in dit onderzoek zal worden geanalyseerd met de antwoorden van de overige deelnemers en wordt uitsluitend voor dit scriptieonderzoek gebruikt. Uw bijdrage blijft anoniem. Dit onderzoek zal ongeveer 10 minuten in beslag nemen.

Alvast hartelijk dank voor uw medewerking!

Lees voordat u verder gaat eerst de onderstaande punten goed door. Als u doorgaat met dit onderzoek dan erkent u het volgende te begrijpen:

• Uw deelname is vrijblijvend.

• U kunt er op elk moment voor kiezen om uw deelname aan het onderzoek te beëindigen, zonder verdere gevolgen.

• Er wordt vertrouwelijk omgegaan met uw gegevens.

Als u meer informatie wilt, kunt u dat opvragen bij Pascalle van Eerden via

(pascallevaneerden@gmail.com).

---[PAGE BREAK]---

SOCIO-DEMOCRAPHICS

Beste deelnemer,

Eerst wil ik u een paar algemene vragen stellen over uw achtergrond. 1. Wat is uw geboortejaar?

[Dropdown met jaartallen 1935 – 2015] 2. Wat is uw geslacht?

[1] Man [2] Vrouw

3. Wat is uw hoogst behaalde opleidingsniveau? [1] Basisonderwijs/Lagere school [2] VMBO/MAVO [3] HAVO/VWO [4] MBO [5] HBO [6] WO [7] Geen onderwijs

(34)

33 [8] Anders, namelijk: ...

4. Wat is uw nationaliteit? [1] Nederlands

[2] Anders, namelijk: ...

Let op: u kunt niet teruggaan naar eerdere pagina’s wanneer u verder gaat in deze vragenlijst. ---[PAGE BREAK]---

POLICITAL PREFERENCE

Hierna volgen enkele vragen over uw politieke voorkeur en uw mediagebruik. 1. Hoe geïnteresseerd bent u in de Nederlandse politiek?

[1] Helemaal niet geïnteresseerd - [7] Heel erg geïnteresseerd 2. Op welke partij heeft u afgelopen Tweede Kamer verkiezingen gestemd?

1. VVD 2. PvdA 3. PVV 4. SP 5. CDA 6. D66 7. ChristenUnie 8. GroenLinks 9. SGP 10. PvdD 11. 50PLUS

12. Andere partij, namelijk:… 13. Ik heb niet gestemd 14. Ik heb blanco gestemd

[99] Weet ik niet

3. Indien u al vaker gestemd heeft, hoe vaak heeft u op deze partij gestemd in het verleden? [1] Ik heb pas één keer op deze partij gestemd

[2] Ik stem vaker op deze partij [3] Ik stem altijd op deze partij [4] Niet van toepassing

4. Voor het onderzoek is het van belang te weten hoeveel tijd u besteedt aan mediagebruik. Hoe vaak per week doet u gemiddeld één van de volgende activiteiten?

Stelling Meerdere keren per dag Één keer per dag Enkele keren per week Één keer per week Minder dan één keer per Nooit

(35)

34 week

Kijken naar het nieuws op televisie

[1] [2] [3] [4] [5] [6]

Een krant lezen (ook online of op tablet / mobiele telefoon) Naar het nieuws op de radio luisteren

Een nieuwswebsite raadplegen

POLICITAL KNOWLEDGE

Hierna volgen enkele vragen over de politiek. Zoek het antwoord niet op of vraag het niet aan anderen. Probeer zelf te zien hoe ver u komt.

5. Hoeveel zetels zijn er in de Eerste Kamer? [1] 50 [2] 75 [3] 85 [4] 100 [5] 150 [99] Weet ik niet

6. Wie is er momenteel minister van Onderwijs, Cultuur en Wetenschap? [1] Jeanine Hennis-Plasschaert

[2] Edith Schippers [3] Jet Bussemaker

[4] Melanie Schultz van Haegen [99] Weet ik niet

7. Wie is er recent afgetreden? [1] Staatssecretaris Mansveld [2] Minister van der Steur [3] Staatssecretaris van Rijn [4] Minister Lianne Ploumen [99] Weet ik niet

8. Welke partij is van mening dat voor vrouwen geen plaats is in de politiek? [1] CDA [2] Christenunie [3] 50Plus [4] SGP [5] RPF [99] Weet ik niet ---[PAGE BREAK]---

(36)

35

TREATMENT

Zo meteen krijgt u een bericht te lezen dat onlangs in het nieuws is verschenen. Lees deze tekst alstublieft van begin tot einde. Daarna willen we u een aantal vragen stellen over het nieuwsbericht. Let op: u kunt later niet teruggaan naar het bericht.

---[PAGE BREAK]--- [Respondents are randomly assigned to one of the following 3 conditions]

Stimuli: 1. Wel polarisatie 2. Geen polarisatie 3. Controle situatie

ATTITUDES TOWARDS EXEMPLARS

Wanneer u denk aan de geïnterviewden uit het nieuwsbericht, hoe voelt u zich dan tegenover hen? Dit mag u aangeven op een schaal van 0 tot 100. Nul betekent zeer negatief en 100 betekent zeer positief. Vijftig betekent dat u geen negatieve of positieve gevoelens heeft ten opzichte van de geïnterviewden.

9. Hoe zou u uw gevoelens richting Mariken (de SP stemmer) beschrijven?

[0] zeer negatief - [100] zeer positief [Respondents slide the bar in any number within the [0,100] range, starting position: 50]

10. Hoe zou u uw gevoelens richting Daniël (de SGP stemmer) beschrijven?

11. Hoe zou u uw gevoelens richting Geert (de VVD stemmer) beschrijven?

12. Hoe zou u uw gevoelens richting Emma (de D66 stemmer) beschrijven?

---[PAGE BREAK]--- 13. In hoeverre herkent u zichzelf in de geïnterviewden uit het nieuwsbericht?

(37)

36 geen

herkenning

herkenbaar herkenbaar herkenbaar Mariken [1] [2] [3] [4] Daniël Geert Emma ---[PAGE BREAK]---

PERCIEVED POLARIZATION

Politieke partijen en kiezers worden vaak omschreven als ‘links’ of ‘rechts’. Kunt u voor volgende partijen aangeven waar u denkt dat haar kiezers zich bevinden in het politieke spectrum?

Links = Staatsbemoeienis op sociaal-economisch gebied, bijvoorbeeld het vaststellen van minimumloon om tot een gelijke verdeling van welvaart te komen. De overheid heeft de taak om de burger te verzorgen van de wieg tot het graf.

Rechts = Geen staatsbemoeienis, op sociaal-economisch gebied, maar geestelijke, politieke en economische vrijheid van burger en bedrijfsleven. De overheid hoeft niet voor alles te zorgen. De taak van de overheid ligt in hoofdzaak bij het handhaven van en toezicht houden op recht en orde.

---[PAGE BREAK]--- 5. Waar zou u een typische stemmer van de onderstaande partijen op het politiek spectrum

plaatsen?

Links = Staatsbemoeienis op sociaal-economisch gebied, bijvoorbeeld het vaststellen van

minimumloon om tot een gelijke verdeling van welvaart te komen. De overheid heeft de taak om de burger te verzorgen van de wieg tot het graf.

Rechts = Geen staatsbemoeienis, op sociaal-economisch gebied, maar geestelijke, politieke en economische vrijheid van burger en bedrijfsleven. De overheid hoeft niet voor alles te zorgen. De taak van de overheid ligt in hoofdzaak bij het handhaven van en toezicht houden op recht en orde.

Partij Helemaal links Neutraal Helemaal rechts

SP [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] VVD SGP D66 ---[PAGE BREAK]---

(38)

37 Politieke partijen en kiezers worden ook vaak omschreven als ‘progressief’ of ‘conservatief’. Kunt u voor volgende partijen aangeven waar u denkt dat haar kiezers zich bevinden in het politieke spectrum?

Progressief = Wil veranderingen doorvoeren in de maatschappij. Conservatief = Is tevreden met de huidige situatie.

6. Waar zou u een typische kiezer van de onderstaande partijen op de conservatief-progressief dimensie plaatsen? Geef een score van 0 tot 10, waarbij 0 staat voor ‘geheel conservatief’ en 10 voor ‘geheel progressief’.

Partij Geheel conservatief Neutraal Geheel progressief SP [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] VVD SGP D66 ---[PAGE BREAK]---

7. Wanneer u denkt aan uw eigen politieke opvattingen, waar zou u deze dan in onderstaande schaal van 0 tot 10 plaatsen?

Links = Staatsbemoeienis op sociaal-economisch gebied, bijvoorbeeld het vaststellen van

minimumloon om tot een gelijke verdeling van welvaart te komen. De overheid heeft de taak om de burger te verzorgen van de wieg tot het graf.

Rechts = Geen staatsbemoeienis, op sociaal-economisch gebied, maar geestelijke, politieke en economische vrijheid van burger en bedrijfsleven. De overheid hoeft niet voor alles te zorgen. De taak van de overheid ligt in hoofdzaak bij het handhaven van en toezicht houden op recht en orde.

[0] Helemaal links - [10] Helemaal rechts

8. Wanneer u denkt aan uw eigen politieke opvattingen, waar zou u deze dan in onderstaande schaal van 0 tot 10 plaatsen?

Progressief = Wil veranderingen doorvoeren in de maatschappij. Conservatief = Is tevreden met de huidige situatie.

(39)

38 ---[PAGE BREAK]--- 14. Als er nu verkiezingen gehouden zouden worden, op welke partij zou u dan stemmen?

1. VVD 2. PvdA 3. PVV 4. SP 5. CDA 6. D66 7. ChristenUnie 8. GroenLinks 9. SGP 10. PvdD 11. 50PLUS

12. Andere partij, namelijk:… [99] Weet ik niet

CONTROL FOR SALIENCY

15. Hieronder staan enkele politieke verkiezingsthema’s. Kunt u per thema aangeven in welke mate het thema van belang zou zijn bij uw stemkeuze als er nu verkiezingen zouden zijn? U kunt de thema's omhoog en omlaag verschuiven door erop te klikken en de thema's te verschuiven. Werkgelegenheid Milieu Criminaliteit Asielzoekers en Migranten Economie Sociale zekerheid Privacy Belastingen en overheidsuitgaven Staatshervorming ---[PAGE BREAK]---

MANIPULATION CONTROL

U bent bijna aan het einde van deze vragenlijst. Graag stellen we u nog enkele vragen over het nieuwsbericht dat u heeft gelezen.

16. Als u terugdenkt aan het nieuwsbericht dat u gelezen hebt, kunt u dan aangeven welke partij(en) werd(en) genoemd in het artikel? Er zijn meerdere antwoorden mogelijk.

(40)

39 [1] SP

[2] PvdA [3] VVD [4] CDA

[5] geen van bovenstaande partijen

17. De personen in het nieuwsartikel waren het met elkaar: [1] oneens

[2] eens

[9] niet van toepassing [99] Weet ik niet

18. Dit nieuwsartikel stond gepubliceerd in: [1] Het Algemeen Dagblad

[2] De Telegraaf [3] De Volkskrant [4] De NRC [99] Weet ik niet

19. Hoe vaak ziet u zit soort artikelen in het nieuws?

[1] Nooit - [5] Continu

DEBRIEFING

U bent aan het einde gekomen van deze vragenlijst. Hartelijk bedankt voor uw deelname aan dit onderzoek!

Dit onderzoek is gericht op de invloed van een nieuwsbericht op de perceptie van polarisatie. Van het nieuwsbericht dat u heeft gelezen bestonden meerdere versies welke geheel ten behoeve van dit onderzoek zijn geschreven en niet daadwerkelijk zijn gepubliceerd in een krant. De informatie weergegeven in het bericht is fictief en berust niet op de waarheid. Daarnaast zijn meerdere versies van het nieuwsbericht voorgelegd respondenten.

Mochten er naar aanleiding van uw deelname aan dit onderzoek bij u klachten of opmerkingen zijn, dan kunt u contact opnemen met Pascalle van Eerden via pascallevaneerden@gmail.com.

Heeft u nog een afsluitende opmerking? Zo nee, klik alstublieft op het pijltje rechts onderin zodat uw gegevens worden opgeslagen.

[Open field for remark]

---[PAGE BREAK]--- We thank you for your time spent taking this survey.