Political Knowledge and Party Identification in the Netherlands: Analyzing the Different Components of Political Knowledge
Heleen Heijungs Student number: 12283444
Master’s thesis Supervisor: Linda Bos Graduate School of Communication Master’s programme Communication Science
28 June 2019
Abstract: While the link between political knowledge and party identification has been studied in the past, no consensus has come forward about this relationship. This study expands on previous research by disentangling political knowledge into several distinct components. Through quantitative survey research (N = 259), this study assesses party identification, political interest, and knowledge of current political facts, political party stances, structural features of the political system, and the functioning of democracy. The relationship between party identification and each component of political knowledge in the Netherlands is examined. Party identification is assessed by taking into account the
complexity of the multi-party system. The findings show a significant positive relationship between party identification and political interest, as well as between party identification and party knowledge. No significant results are found for the other aspects of political knowledge. The findings support the notion that political knowledge is a multifaceted construct that can be split up into various distinct components, which leads to clearer results when examining the association with party identification.
Introduction
Trends of political dealignment are visible in many democracies across Europe (Dassonneville, 2012). Several reasons for this declining party identification have been proposed. Some (e.g. Dalton, 1984; Dalton, 2004) link this trend to increasing levels of political knowledge. Dalton proposes that, as exposure to information increases and higher education becomes more prevalent, citizens become more knowledgeable. Knowledgeable citizens might not feel the need to follow ‘partisan cues’ as strongly, and therefore they could be less likely to identify with a party (Dalton, 2004). In addition, individuals with higher levels of political knowledge may also be more likely to grasp more nuances and complexity, which could lead to indecision, as Barker and Hansen (2005) suggest. This indecision could also be associated with a decreased level of party identification.
However, one could also expect an increase in political knowledge to be linked to stronger party identification. Individuals with a high level of knowledge may be better able to identify the differences between parties, and align themselves with the party with which they have the most in common. Dassonneville (2012) acknowledges that there is no consensus on the relationship between political knowledge and ‘electoral volatility’, a concept that can be seen as the flipside of party identification.
The aforementioned pattern of declining party identification has also been observed in the Netherlands (Geers & Strömbäck, 2018). The political landscape of the Netherlands provides an interesting context in which to study party identification. Unlike many other European democracies such as Belgium and Sweden, the Netherlands does not set any minimum vote threshold that parties need to meet before they can gain a seat in the national parliament; the only requirement is that they receive at least 1/150 of the national vote share. This means that smaller political parties are able to enter the parliament with relative ease. With 13 political parties in parliament, Dutch politics is extremely fragmented. Even the
VVD, the party that gained the largest vote share in the 2017 parliamentary elections, only received 21.2% of all votes. Due to this splintered political situation, some parties are ideologically quite similar. As a result, people may be less likely to have a strong party preference.
The present paper intends to further study the relationship between the concepts of political knowledge and party identification in the Netherlands. The emphasis in this study is on the different elements that make up political knowledge. Political knowledge is often treated as a single construct, although at times authors refer to ‘political awareness’ or ‘political sophistication’ instead. This study contributes to the existing research by
acknowledging that political knowledge is a multifaceted construct that can be broken down into several separate components, and studying the association between party identification and each knowledge component. This leads to this paper’s main research question:
To what extent are various components of political knowledge associated with party identification in the Netherlands?
To answer the research question, an online survey was conducted among 259 respondents, in which they were asked questions about the extent to which they identified with a political party, as well as a number of factual questions about politics. This paper will first discuss the concepts of cognitive mobilization and electoral volatility, and review how previous researchers have assessed political knowledge and party identification in multi-party systems. After that, it will set out the main findings from the survey and explain the relevance of these results in the context of the existing theories.
Theoretical background The link between political knowledge and party identification
Besides decreasing levels of citizens’ party identification, patterns of declining consistency in voting behavior are visible across several Western countries (Geers &
Strömbäck, 2018; Dassonneville, 2012). Additionally, voters in many countries, including the Netherlands, decide their vote at a later point in the campaign than they have historically done (Irwin and Van Holsteyn, 2008). The link between these trends and political knowledge relates to two key concepts: electoral volatility and cognitive mobilization. Electoral volatility refers to voters who either do not vote consistently for the same party across elections or change their voting intention in the middle of an election campaign. Volatility is highly negatively related to party identification; a high level of volatility would indicate a low level of party loyalty and vice versa. The Netherlands experiences high electoral volatility (Van der Meer et al., 2015), thus party identification is rather low. While increasing electoral volatility is occurring across many European countries, the trend appears to be particularly strong in the Netherlands (Mair, 2008). Mair argues that this can be partially attributed to patterns of individualization and depillarization that have occurred in the Netherlands over the past decades, and partially to some particular features of the Dutch electoral system. However, as the present paper will argue, political knowledge may also play a role.
Cognitive mobilization, as proposed by Dalton (1984), is linked to the international trends of declining party identification. This concept associates a high level of knowledge with a low level of party identification. Dalton understands cognitive mobilization as a process through which individuals are exposed to an ever-increasing volume of information about politics, due to the rise of mass media. This exposure to information leads to higher levels of political knowledge and gives citizens an improved ability to comprehend the complexities and nuances of the political situation in their country. Uninformed voters are likely to rely on partisan cues when deciding for whom they will vote. As political knowledge increases, these additional skills allow voters to make more sophisticated judgments about politics. Therefore, Dalton argues, this need for partisan cues will decrease when political knowledge is higher. Rather than relying on the party to which they feel closest, voters will
instead consider the stances of candidates on key political issues. This would explain the decreasing number of citizens who identify with a particular political party. This decreasing reliance on partisan cues can also be linked to Mair’s (2008) characterization of contemporary Dutch politics; rather than voting in line with traditional party affiliations, citizens decide their vote in the context of a specific election. Dalton (1984) finds evidence to suggest that highly informed voters are disproportionately likely to be political independents. In
subsequent studies (e.g. 2004), Dalton repeats and expands on these findings. Dalton distinguishes between two groups of people not affiliated with a party: apoliticals and apartisans. Apoliticals are political independents with a low level of political knowledge and political engagement. Apartisans, in contrast, are political independents with a high level of knowledge and involvement. Dalton’s results indicate that cognitive mobilization leads to an increase in the number of apartisans in a country.
Dalton’s findings, suggesting a negative association between political knowledge and party identification, are supported by many other authors. However, while the associations they study are similar, these authors frequently refer to different concepts. For instance, Barker and Hansen (2005) do not focus on political knowledge as such, but rather discuss the effect of ‘systematic cognitive processing’. They build on the work of Linville (1982), who found that people who think complexly about an issue tend evaluate it more ambivalently. Zaller & Feldman (1992) similarly found that people’s attitudes became ‘less predictable’ after more consideration of an issue, particularly for people with high levels of knowledge. The findings of Barker and Hansen support the idea that voters who think deeply about the candidates and parties, are less reliant on heuristic cues and are also less consistent in their voting choices and feel more conflicted about the candidates. Barker and Hansen argue that a more complex thought process is likely to cause a sense of “perplexity, equivocation,
However, no consensus exists regarding the effects of ‘cognitive mobilization’. Albright (2009) contests Dalton’s findings. In a study of a large number of European countries, Albright finds the opposite result: informed voters have a higher likelihood of identifying with a political party.
Others have found significant non-linear relationships between knowledge and party identification. Van der Meer et al (2015) and Kuhn (2009) both found a curvilinear
relationship between political sophistication, a construct similar to political knowledge, and electoral volatility; in the Netherlands (Van der Meer et al, 2015) and in Switzerland (Kuhn, 2009). In these studies, moderately sophisticated voters were found to be the most volatile group, while highly sophisticated and highly unsophisticated voters both voted more consistently. This would leave the moderately sophisticated as the most likely group to be persuaded to change their voting intention and thus the least loyal to their party. According to Zaller (1992) and Van der Meer et al (2015), an explanation for this is that citizens with a low level of sophistication may not be exposed to enough information to change their voting intention - while the extremely sophisticated may not be as easily persuaded by additional information. Moderately knowledgeable voters are therefore the most likely to be ‘picky’. Van der Meer et al. (2015) suggest that “[p]icky voters are loyal to their own ideas, not to a single political party” (p. 110).
Dassonneville (2012) suggests that a distinction should be made between voters with an inconsistent voting intention across elections (inter-election volatility) and voters who switch voting intentions during an election campaign (intra-election volatility). Dassonneville found that inter-election volatility is more common for highly sophisticated voters, but that these individuals do decide for whom they will vote early on in an election campaign. Meanwhile, intra-election volatility is more common among less sophisticated voters, who tend to wait until the later stages of a campaign to decide for whom they will vote. However,
in a study of undecided voters in the Netherlands, Irwin and van Holstein (2008) did not find that late deciders were significantly more or less knowledgeable than early deciders. Irwin and van Holstein determined that, more so than early deciders, late deciders were focused on strategic information regarding coalitions. Voters who decided their vote late in the campaign were likely to be less loyal to a particular political party than voters who decided early. Assessing political knowledge
Many studies in the field of political communication attempt to assess citizens’ levels of political knowledge. It is almost always assessed by asking respondents a short series of fairly straightforward factual questions that can be answered objectively. However, while most researchers use some form of factual questions, political knowledge is not measured consistently across studies (Eveland & Hively, 2009; Zaller, 1992). For instance, some (e.g. Andersen, Tilley & Heath, 2005; Barker and Hansen, 2005; Chaffee, Zhao & Lesher, 1994) ask respondents to correctly identify the stances of political parties on a number of major policy issues. This component of knowledge will in the present study be referred to as ‘party knowledge’. In contrast, others (e.g. Greene, 2004) did not ask for the stances of political parties, but instead about current facts, such as the party currently in the majority in the parliament or Congress, or structural facts, such as the functions of certain branches of
government. The former will be referred to as ‘current knowledge’ and the latter as ‘structural knowledge’. Moeller and De Vreese (2015) identify these differences, and argue that political knowledge can take multiple forms. They distinguish between three types of knowledge: factual, structural, and civic.
Factual knowledge includes both “knowledge of textbook facts about politics” and “surveillance knowledge such as correctly identifying political key players” (Moeller & De Vreese, 2015, p. 4). This is the most commonly assessed type of knowledge.
Structural knowledge refers to an understanding of the “political process and structures that form the political system”. This would include knowing “the function of a political party” and “the consequences of certain policy decisions” (Moeller & De Vreese, 2015, p. 4). Eveland and Hively (2009), for instance, assess structural knowledge by asking respondents to identify the extent to which certain issues are interconnected.
Civic knowledge is a more complex type of knowledge which includes a combination of both structural and factual knowledge (Moeller & De Vreese, 2015; Delli Carpini &
Keeter, 1993). Civic knowledge “encompasses both factual knowledge of key institutions in a political system as well as the ability to relate them to each other.” (Moeller & De Vreese, 2015, p. 4). Moeller and De Vreese argue that this type of knowledge is the most relevant for citizens’ ability to participate in the democratic process.
In addition to these three types of knowledge, many authors study ‘political
sophistication’. This is defined more broadly than the civic knowledge that Moeller and de Vreese identify. In some definitions, in addition to a person’s level of knowledge, political sophistication includes their level of interest in politics and their exposure to political information (Luskin, 1990). Kuhn (2009) argues that the ideal way to assess political
sophistication includes a set of knowledge questions, but education and political interest can be used as a proxy. Furthermore, those studying cognitive mobilization seem to assess it in a similar manner. Dalton (2004) writes that “The cognitively mobilized are those who possess both the skills and motivation to grapple with the complexities of politics on their own” (p. 5). Dalton measures cognitive mobilization using a combination of factual political questions, education level, and political interest.
The different possible measures of political knowledge may be partially responsible for the contrasting findings of different studies on knowledge and its link to party
election cycle, as Andersen, Tilley and Heath (2005) suggest; their findings indicate that the average level of knowledge is highest shortly after an election.
Party identification
Party identification in the narrowest sense could refer to a person’s actual membership of a political party. However, Greene (2004) argues that understanding party identification as a ‘psychological group’ helps explain the emotional aspects of belonging to a party. Party identification is not purely based on the party’s policy positions. According to Adams, Ezrow and Somer-Topcu (2011), voters who identify with a political party will largely remain loyal to it even when the party changes its position on major issues. In fact, they find that voters often remain unaware of such changes. Greene (2004) used the Identification with a
Psychological Group scale (IDPG), developed by Mael and Tetrick (1992) to measure party identification as a social identity. Dassonneville (2012) argues that American studies on party identification may not always be replicable in Europe, due to the structural differences in the nature of the political systems. To measure the strength of party identification in multi-party systems, Rosema and Huddy (2012) developed a 4-item scale, largely based on Mael and Tetrick’s scale and specifically aimed at political systems with a larger number of political parties. Bankert, Huddy and Rosema (2017), in an article based on the 2012 questionnaire, suggest that, in multi-party systems, people may not always feel close to one particular party but may instead strongly identify as opposed to a specific party. Alternatively, they might align themselves with a coalition or group of parties. Overall, the dynamics of partisanship tend to be more complex in multi-party systems than in two-party systems (Bankert, Huddy & Rosema, 2017). It would then be plausible that the feeling of indecision and overwhelmedness related to thoughtful consideration (Barker & Hansen, 2005), is exceptionally prevalent in a political system with a large number of political parties, such as the Netherlands.
Overall, no clear consensus arises from previous research on party identification and political knowledge. Some studies suggest a negative relationship, others suggest a positive relationship, and yet other studies find non-linear relationships. In light of the various measures of knowledge and the divergent findings, the present study will attempt to disentangle the many varieties of political knowledge. The types of knowledge are further explained in the method section of this paper. Formulating clear hypotheses is complicated by the lack of consistent measures across previous research. Therefore, the following
sub-questions will be examined:
1. What is the relationship between current knowledge and party identification? 2. What is the relationship between party knowledge and party identification? 3. What is the relationship between structural knowledge and party identification? 4. What is the relationship between civic knowledge and party identification? 5. What is the relationship between political interest and party identification?
Methods
To answer the research question, a cross-sectional survey design was selected. Convenience sampling was used; participants were recruited through a number of online platforms, including social media and internet forums. To ensure diversity in the sample, a variety of forums with different audiences were selected. Since this study is about electoral behavior in the Netherlands, participants had to be eligible to vote in Dutch parliamentary elections. This includes all Dutch nationals who were at least 18 years old at the time of taking the survey. Survey data were collected between 26 April 2019 and 9 May 2019. 259 respondents fully completed the questionnaire. Ages within the sample ranged from 18 to 88, with a mean of 48.14 (SD = 18.50). 37.8% of the respondents were male, 60.6% were female, and 1.5% identified as neither male nor female. The sample was very highly educated: 49.4% of respondents has obtained or is pursuing a master’s degree, and 40.5% has obtained or is
pursuing a bachelor’s or HBO degree. The remaining 10.1% has a lower education level than HBO. This is considerably higher than the national education level. Data from the Dutch government statistics institution CBS indicates that 69% of Dutch residents between 15 and 75 years old has an education level below HBO (Centraal Bureau voor de Statistiek, 2018). Political knowledge
As in most previous studies, political knowledge was measured by asking participants a number of questions with objectively true and false answers. This approach has been advocated by many, including Zaller (1992) and Delli Carpini and Keeter (1993). Some generally accepted questions to assess political knowledge were developed in an extensive study by Delli Carpini and Keeter (1993). These questions have since been used in many studies on political knowledge, such as Greene (2004). However, these questions measure several different varieties of knowledge as a single construct. In the present study, the broad construct of political knowledge was split up into four components to assess different varieties of knowledge. These were identified based on the existing literature and are explained in more detail later. The four components are: current knowledge, party knowledge, structural
knowledge, and civic knowledge. Each was assessed through a series of three factual multiple-choice questions (see appendix 1 for a complete list of questions and all answer options). For the four components of knowledge, sum scores were calculated for each
respondent. Additionally, respondents were asked to indicate their level of political interest on a scale ranging from 0 to 7. The results of these are shown in table 1. The highest achievable score for each component was 3 (all questions answered correctly) and the lowest achievable score was 0 (no questions answered correctly). An ‘I don’t know’ option was available for each question, to prevent random guessing. To ensure that people would not guess or look up the answers, a line was added at the start of the section on political knowledge, asking them to not look up the answers as well as informing respondents that “many people do not know the
answers to these questions”, which Delli Carpini and Keeter (1993) advised in their research on political knowledge (p. 1183).
Table 1
Composite knowledge scores
Mean (Standard deviation) Sum current knowledge 2.41 (.84)
Sum party knowledge 2.51 (.72) Sum structural knowledge 2.41 (.75) Sum civic knowledge 2.70 (.54) Political interest 5.01 (1.34)
Respondents who selected that they did not know the answer did not receive any points for that question, and neither did respondents who selected an incorrect response; incorrect and ‘don’t know’ were treated the same for the purpose of calculating knowledge scores. There were substantial differences between the questions on how well people performed; the proportion of correct answers ranged from 64.9% to 97.7%.
Current knowledge
Current knowledge in this study refers to knowledge of political topics and persons that are variable over time. To obtain and retain a high level of current knowledge, a person needs to pay attention to news about politics. In the present study, the questions on current knowledge included: identifying the chairperson of the parliament; identifying the position of Dutch foreign minister Stef Blok; and identifying a party in the current government. The mean of the composite variable is 2.41 (SD = .84). Scores for individual questions can be found in appendix 2. Delli Carpinin and Keeter (1993) warn against considering item-total correlations in scales that assess knowledge, because such correlations are largely unable to account for the varying level of difficulty of the questions. However, the Mokken scale is
designed to account for such variation. According to this scale, the three items assessing current knowledge form a weak scale (H = .39).
Party knowledge
Some (e.g. Andersen, Tilley & Heath, 2005; Barker & Hansen, 2005) assess
knowledge by asking respondents to identify the position of political parties or candidates for political office on a number of major issues. However, these authors conducted their research in the United Kingdom and the United States respectively, which have only a small number of major political parties. The political landscape in the Netherlands stands in stark contrast to the United States, with more than a dozen major parties. This means that a far higher level of knowledge is necessary to accurately know the positions of all parties on key issues. To prevent confusion about the many parties and their policy positions, the present study limited the questions on party knowledge to issues with very clear proponents and opponents. The issues on which respondents were asked to determine the parties’ stances included one right-wing, one left-right-wing, and one centrist position; namely: position on the European Union, position on tax rates for the highest income groups, and position on the concept of a binding referendum. Questions were taken from the voting advice application ‘Stemwijzer’ of the 2017 elections, and up-to-date issue stances were retrieved from each political party’s official website. The mean score for party knowledge is 2.51 (SD = .72). The items form a strong scale (H = .61).
Structural knowledge
Unlike current knowledge, paying attention to the news is not required to gain a high level of structural knowledge. Instead, this type of knowledge can be acquired by studying the basic foundations and structures of the political system. For this type of knowledge,
participants were asked how members of the Dutch Senate (Eerste Kamer) are selected; to identify the process of changing the constitution; and to know the frequency of local elections.
The mean score of structural knowledge is 2.41 (SD = .75). The first and the third question form a weak scale (H = .30), and do not form a scale together with the second question -- however, for the sake of keeping a consistent three-item scale for all knowledge components, all three questions were used. That said, any potential significant results relating to structural knowledge should be regarded with some skepticism.
Civic knowledge
The questions assessing civic knowledge were taken from a study on civic knowledge by Moeller and de Vreese (2015). The questions in this section are generic questions about the functioning of politics and democracy; unlike the other questions, the civic knowledge
questions do not pertain specifically to Dutch politics. Participants were asked what the function of having multiple political parties is in a democratic country, and asked to identify the political implications of two hypothetical scenarios. Participants tended to score higher on civic knowledge than on current, party and structural knowledge; the mean score is 2.70 (SD = .54). The three questions form a weak scale (H = .39).
Party identification
Measurements of party identification were based on research by Rosema and Huddy (2012) intended to assess partisanship in multi-party systems. First, participants were asked whether they considered themselves a supporter of a particular political party. 38.6% of respondents self-identified as supporters. In a large (N = 5195) sample of Dutch residents over the age of 16 in 2012, Rosema and Huddy (2012) found that a considerably lower 21.6% of respondents considered themselves supporters of a party. Those who did not consider
themselves a supporter, were asked whether they felt closer to one political party than to other parties. 48.6% of the total respondents in the present study considered themselves close to a particular party. Lastly, respondents who did not consider themselves close to a party were asked for which party they would vote if national elections were held today. 10.0% of the
total sample gave a party preference here. Only 2.7% of the total respondents still did not name a particular party in the final section. The specific parties that respondents mentioned can be found in appendix 3. Overall, respondents were considerably more left-leaning than the general Dutch population. After this initial measure, the strength of party identification was assessed using a slightly adapted version of the questionnaire developed by Rosema and Huddy (2012), which is based on the Identification with a Psychological Group scale designed by Mael and Tetrick (1992). The scale by Rosema and Huddy includes eight self-reported statements to assess how strongly a person feels connected to their political party. All eight items of this scale were used in the present study. Respondents were asked to rate their level of agreement on each statement on a Likert scale. In the original study by Rosema and Huddy, a Likert scale of 1 to 4 was used, but the present study included a scale ranging from 1 to 8 to capture more variation. A composite variable was calculated as the mean score of the 8 items, with equal weight assigned to each item. The mean of this composite score is 2.87 (SD = 1.18), which is fairly low considering the highest possible score is 8. However, Rosema and Huddy also found a rather low mean score of 1.66 on a scale of 1 to 4. A principal factor analysis in the present study indicates that all 8 items are a single factor with an eigenvalue above 1 (eigenvalue 4.51). This explains 56.42% of the total variance. This scale is highly reliable; Cronbach’s alpha = .89. The question battery was asked in reference to the person’s party of choice. All respondents who either considered themselves supporters or close to a party, or who indicated they would vote for a particular political party, were asked this battery of questions. The 7 respondents who did not mention any party were left out of this part and coded as 1, the lowest possible score on party identification. Table 2 shows that, as can be expected, individuals who considered themselves supporters of a particular political party tended to score highest on the battery of eight party identification questions. A one-way analysis of variance (ANOVA) indicates that these differences are statistically significant, F
(2, 249) = 35.51, p < .001, η2 = .22. A Bonferroni post-hoc test suggests that the mean party
identification score for supporters was significantly higher than for people who felt close to a political party (Mdifference = 1.07, p < .001). However, the score for people who felt close to a political party was not significantly higher than for those who did not feel close to any particular party but would vote for a party if the elections were held today.
Table 2
Mean party identification scores per group
Mean party identification score (standard deviation) N
Supporter of political party 3.59 (1.20) 100
Close to a political party 2.52 (.87) 126
Would vote for a political party 2.31 (1.01) 26
Political interest
The original intent of this study was to consider political sophistication as one of the independent variables. Consistent with Kuhn (2009), the intent was to calculate a score for sophistication based on a combination of education level and political interest. However, the sample of the present study had a disproportionately high education level compared to the general population. Education was left out of the measure of sophistication due to the lack of variance and the extreme skew towards higher levels. Instead, political interest alone was used as a proxy for political sophistication. Political interest was self-reported on a scale ranging from 0 (least interested) to 7 (most interested). The mean of political interest in the sample is 5.01 (SD = 1.34).
Results
Before the main research question is addressed, the correlations between the different types of knowledge were determined. As can be seen in table 3, most components of
Pearson correlation coefficients are weak to moderate. This suggests that the types of knowledge can be viewed as separate but related concepts.
Table 3
Political knowledge correlation matrix Current knowledge Party knowledge Structural knowledge Civic knowledge Political interest Current knowledge -Party knowledge .407** -Structural knowledge .375** .326** -Civic knowledge .169** .255** .031 -Political interest .508** .411** .346** .041 -*p < .05 *-*p < .01
To address the main research question, all relevant relationships were examined through linear regression models. The models all control for the effects of age, gender, and education level. Several different regression models were conducted to examine both the combined and the separate effects of the various components of knowledge. In all these models party identification is the dependent variable. The reported coefficients are unstandardized.
The first set of regression models can be found in table 4. Model 6, which includes all variables under examination, only indicates that political interest is a statistically significant predictor of party identification. Higher levels of interest in Dutch politics are associated with stronger feelings of party identification. Political interest remains a significant positive predictor of party identification in model 5, in which its predictive value is examined excluding the political knowledge variables.
Table 4
The relationship between political knowledge and party identification
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
Age -.002 (.004) -.002 (.004) .000 (.004) .000 (.004) -.055 (.004) -.003 (.004) Female -.317* (.151) -.279 (.151) -.356* (.150) -.357* (.150) -.172 (.141) -.165 (.142) Education .128 (.111) .071 (.113) .141 (.112) .137 (.112) .066 (.102) .058 (.107) Current knowledge .166 (.093) -.145 (.102) Party knowledge .299** (.109) .103 (.113) Structural knowledge .073 (.101) -.112 (.100) Civic knowledge .134 (.139) .127 (.133) Political interest 358** (.053) .399** (.062) Constant 2.448** (.359) 2.215** (.372) 2.587** (.367) 2.366** (.478) 1.248** (.373) .989* (.495) F (p) 2.83 (.026) 3.94 (.004) 2.14 (.077) 2.25 (.065) 13.86 (.000) 7.60 (.000) R2 .04 .06 .03 .04 .18 .20 *p < .05 **p < .01
Note: table shows unstandardized coefficients and standard errors in parentheses. N = 255 in all models
Out of the four knowledge components assessed through factual questions, only party knowledge was a significant predictor of the strength of a person’s party identification. This relationship is visible in model 2, which excludes political interest and the other components of political knowledge. However, it does not hold up in model 6, in which the other
independent variables are included. This indicates that the relationship between party identification and party knowledge is not as solid as the relationship between party identification and political interest.
The other three components of knowledge did not show any significant relationship in any of the models tested. No evidence was found to suggest that current, structural, and civic
knowledge were either significant positive or significant negative predictors of party identification.
Some models show statistically significant effects for the control variable of gender; in these cases, women are significantly less likely than men to express a strong sense of party identification. In contrast, Van der Meer et al. (2015) found that men were somewhat more electorally volatile than women. Greene (2004) did not find significant differences between men and women on party identification. The relationship between party identification and gender may be interesting to examine further in future research, but it is not the focus of the present study and the statistical significance does not hold up across all models.
In addition to the models assessing the linear relationships, some models were added to examine whether any curvilinear association exists (see table 5). Kuhn (2009) and van der Meer et al. (2015) found significant non-linear relationships between political sophistication and electoral volatility, with moderately politically sophisticated citizens being the most likely to be volatile voters. To test for such curvilinearity, the relationships were assessed through OLS regression, and dummy variables were created for three levels of knowledge. Knowledge is separated into three groups: low, medium, and high. Low knowledge means that
respondents answered 0 or 1 questions correctly, medium means 2 correct responses, and high means 3 correct responses. Medium knowledge is treated as the reference category in these models. No curvilinear relationships were found for any type of knowledge using this method.
Table 5
Low and high political knowledge compared to moderate political knowledge
Model 7 Model 8 Model 9 Model 10 Model 11
Age -.002 (.004) -.002 (.004) -.001 (.004) .000 (.004) -.002 (.004) Female -.319* (.153) -.283 (.151) -.357* (.151) -.362* (.150) -.282 (.144) Education .131 (.111) .070 (.114) .138 (.113) .126 (.113) .09 (.106) Low current knowledge -.278 (.247) High current knowledge .135 (.174) Low party knowledge -.336 (.281) High party knowledge .320 (.170) Low structural knowledge -.082 (.244) High structural knowledge .087 (.167) Low civic knowledge -.470 (.511) High civic knowledge .103 (.174) Low political interest -.778 (.209)** High political interest .582 (.208)** Constant 2.798** (.354) 2.800** (.338) 2.737** (.359) 2.706** (.374) 2.931** (.329) F (p) 2.23 (.052) 3.08 (.010) 1.73 (.129) 1.92 (.091) 6.80 (.000) R2 .04 .06 .03 .04 .12 *p < .05 **p < .01
Note: table shows unstandardized coefficients and standard errors in parentheses. N = 255 in all models
Robustness check
One could argue that the limited results found in this study are in part due to the skewed nature of the sample, which is disproportionately left-leaning when compared to national polls. As a robustness check, an additional regression model was therefore added (table 6) in which cases were weighted based on the political party that respondents supported. The party choices of actual respondents in this sample were compared to the Peilingwijzer, a website that aggregates polls about the Dutch parliamentary elections to calculate the polling average. Peilingwijzer data from May 1st, 2019 were used to create this model. Due to the rather skewed nature of the sample this model must be viewed with some skepticism. For instance, the sample only contains a single PVV-voter (representing .4% of all 259 respondents), but nationally the PVV polls at 8%. On the opposite end, 13.3% of
respondents support or vote for the Partij voor de Dieren, while it polls at 4% nationwide. Two parties (DENK and SGP), which both poll at 2% nationally, were not represented at all in the sample.
In the model with weighted responses, the positive relationship between political interest and party identification still holds up. In addition, a significant negative relationship is now found between current knowledge and party identification; those who know less about the current state of politics, are more likely to strongly identify with a political party.
Table 6
The relationship between political knowledge and party identification, with cases weighted by political party Model 12 Age .003 (.004) Female -.549** (.142) Education .055 (.097) Current knowledge -.233* (.097) Party knowledge .006 (.113) Structural knowledge -.191 (.099) Civic knowledge .079 (.116) Political interest .343** (.059) Constant 1.899** (.457) F (p) 11.19 (.000) R2 .28 *p < .05 **p < .01
Note: table shows unstandardized coefficients and standard errors in parentheses. N = 245
Discussion and conclusion
All in all, strong evidence was found to suggest that the more interested an individual is in politics, the more likely they are to feel a strong sense of identification with a particular political party. This association holds up across all tested models. Some weaker evidence was found to suggest that the more knowledge a person has about the positions of political parties on various issues, the more likely they are to report a strong party identification. However, the overall results of this study indicate no significant relationships between party identification and the current, structural and civic components of political knowledge.
Taken together, these findings suggest that political knowledge and sophistication are, if at all, positively associated with party identification. No component of knowledge was found to have a significant negative relationship with party identification except in the
weighted model, and no curvilinear effects of any kind were found. These findings are not entirely in line with existing literature on the topic; some authors researching political knowledge and party identification have found negative or curvilinear associations. As addressed in the theory section of this paper, political interest is often regarded as a proxy for political sophistication (e.g. Kuhn, 2009; Van der Meer et al., 2015). Dalton, meanwhile, assessed sophistication using a combination of factual questions, interest, and education, as did Albright (2009). Dalton found a negative relationship between political sophistication and party identification; Albright found a positive relationship. Kuhn (2009) and Van der Meer et al. (2015) found curvilinear effects. The findings of the present study therefore come closest to the findings of Albright, although this study did not include education level as part of political sophistication, as explained in the method section.
However, the dependent variable of the present paper did not exactly match that of previous research. While the present paper has examined party identification, Kuhn (2009) and van der Meer et al (2015) studied electoral volatility. Although these concepts are highly related - volatility can be regarded as roughly the inverse of party identification - they are somewhat distinct. Studies on volatility generally examine voting intention as a binary variable; people either intend to vote for a specific party, or not. In contrast, the present study considers the extent to which people identify with a particular party on an eight-point scale. These differences between the exact constructs under examination make it harder to
contextualize the findings of the present study.
Regarding the positive relationship between party identification and party knowledge found in model 2, it appears that individuals who are better able to distinguish between the various political parties are more likely to identify with one. Rather than the “perplexity, equivocation, vacillation, and/or intellectual paralysis” (Barker and Hansen, 2005, p. 322) that one might expect to originate from high levels of political knowledge, it seems that those who
understand the differences between the parties are better able to select with which party they align, and therefore they feel closer to this party. Particularly in the Dutch multi-party system, citizens might need more knowledge than in two-party systems before they feel informed enough to select the party with which they identify the most. Future cross-country
comparative research would be needed to determine the specific relevance of party knowledge in the Netherlands. That said, this study is not able to determine any causality. It is also
possible that, for instance, feeling strongly affiliated with a political party motivates individuals to learn about the different parties and their positions. Further research is necessary to determine the direction of such a relationship.
This study has several limitations. Convenience sampling was used to recruit
participants for this study, as it is the cheapest way to reach a large number of respondents in a short time period. However, convenience sampling leads to biased samples that are not representative of the actual population. That is certainly the case here. Age, gender, party preference, and education level are all different from the distribution in the entire population of the Netherlands. On average, respondents in this study scored surprisingly high on the political knowledge questions. Based on the research of Zaller (1992), one would expect that “[t]here is high variance in political awareness around a generally low mean.” (p. 18). That said, Zaller examined knowledge in the United States in the early 1990s while the present study examines knowledge in the Netherlands in 2019, which might explain some of this disparity. In addition, the sample of respondents in the present study is not representative of the Dutch population as a whole; education levels in the sample level are considerably higher than average education levels in the Netherlands. In fact, education level is the most obvious indicator that the sample is unrepresentative of the overall population of the Netherlands; 49.4% of respondents have or are currently pursuing a master’s degree, and only 10.1% has an education level below HBO. As explained previously, this deviates considerably from
national education levels. This has likely had a substantial impact on the results, especially considering that some authors (e.g. Dalton, 2004) have explicitly acknowledged that
education level and political knowledge are highly interrelated. As such, the results may not be generalizable to the larger Dutch population.
Unlike Dalton (2004), this study did not make a clear distinction between what Dalton refers to as ‘apartisans’ - highly knowledgeable individuals who make the conscious choice not to identify with any particular party - and ‘apoliticals’ - less knowledgeable individuals who do not identify with a party because they are not politically engaged. The questionnaire of the present study asked those who did not consider themselves a supporter of and did not feel close to any particular party who they would vote for if elections were held today. This way, it encouraged respondents to pick a certain party. Only 7 respondents did not name any party that they identified with or would vote for. However, those who neither considered themselves a supporter of a particular political party nor felt more attracted to one party than to any others, can be considered apartisans or apoliticals. In total, this group consists of only 33 out of the 259 respondents. As a result, making a distinction between apoliticals and apartisans within this group would leave very small groups that are likely not representative due to the small group size. In future studies with larger sample sizes, such a distinction could be made to obtain a clearer picture of the variation amongst people who do not identify with a political party.
Additionally, there was very little variation in the level of civic knowledge in the sample. This knowledge component had both the highest mean (2.70 out of 3) and the lowest standard deviation (.54). The civic knowledge questions were directly adapted from a study by Moeller and de Vreese (2015), in which respondents scored considerably lower. That said, Moeller and de Vreese specifically studied adolescent respondents between 15 and 18 years old, while the present study was focused on adults, so the results cannot truly be compared. It
must be noted that the questions on civic knowledge did not relate specifically to Dutch politics. Instead, they were based more on logic than knowledge. An example of such a question is: “Imagine the following scenario: a country has a declining birth-rate and an increasing life span. Which of the following problems will have to be solved as a result? (a) Schools need to be built, (b) Pensions for the elderly have to be financed, (c) Low income housings have to be built, (d) Crime and violence have to be fought, (e) Don’t know” (p. 15). It is likely that the disproportionately high education level of the sample in the present study contributed to the unexpectedly high scores on civic knowledge, since this type of knowledge is more dependent on logical reasoning than on information about current affairs.
Another limitation to this study is that only three questions were asked to assess each type of political knowledge. This was done to ensure a high rate of survey completion - a substantially higher number of questions would likely have led to a drop-off in respondents. However, Zaller (1992) advocates for assessing knowledge with a larger battery of questions. In future studies, researchers could decide to expand the questionnaire to capture more variation in the different components of knowledge.
There is also some selection bias: those already interested in politics or with self-perceived high knowledge levels might be more likely to complete a survey on political knowledge. Furthermore, while respondents were asked not to consult others or look up answers to any of the political knowledge questions, it cannot be verified that all respondents complied with this request.
The timing of the study is also important to keep in mind. The questionnaire in the present study was fielded from late April to early May. This period was fairly close to two elections: the provincial elections on March 20th, and the elections for the European Parliament on May 23th. Andersen, Tilley and Heath (2005) determined that political
to be high in the period shortly preceding an election. The period in which the questionnaire was fielded may therefore be a period in which individuals were unusually politically knowledgeable, due to the proximity to two elections. On the other hand, both the elections for the European Parliament and the provincial elections had relatively low turnout rates compared to national elections, so the effects are likely to be weaker. That said, it is notable that the level of political knowledge in this study’s sample was very high. The same may be true for party identification; citizens might report that they feel closer to their party in the lead-up to the election as a form of support. While Geers and Strömbäck (2018) focused more on voting intention than on party identification, they did find evidence that ‘crystallization’ - a process in which citizens acquire a voting intention for a specific party when they previously did not have one - was most common in the late phases of an election campaign. Lastly, it is also plausible that political interest is affected by the proximity to an election; individuals might become more politically attentive in the lead up to an election and therefore consider themselves temporarily more politically interested. All in all, there is no ‘neutral’ time to conduct public opinion research, as the distance from elections will always have some sort of impact. For future research, it would be interesting to conduct a longitudinal study to further address the impact of the proximity to an election on party identification, political interest, and the various types of political knowledge.
As has been discussed at length in this paper, assessments of political knowledge are rarely consistent, which leads to deviating results. Some, such as Delli Carpini and Keeter (1993), have attempted to create a more unified measurement for political knowledge.
However, their scale does not apply to non-American politics, as it includes questions such as “How much of a majority is required for the U.S. Senate and House to override a presidential veto?” (p. 1204). In addition, their scale includes a combination of questions about current, structural and party knowledge. The present study has demonstrated that not all components
of the broad concept of political knowledge can be viewed as one and the same. In the context of this paper’s main research question (To what extent are various components of political knowledge associated with party identification in the Netherlands?), while significant results were found for political interest and for party knowledge in some contexts, other knowledge aspects were not found to be significantly correlated with party identification. While the different types of knowledge were largely found to be correlated with each other, these correlations were only weak to moderate. Therefore, a general recommendation is that those who study political knowledge ought to be clear about the type of knowledge that they intend to measure.
References
Adams, J., Ezrow, L., & Somer-Topcu, Z. (2011). Is anybody listening? Evidence that voters do not respond to European parties’ policy statements during elections. American Journal of Political Science, 55(2), 370-382.
Albright, J. J. (2009). Does political knowledge erode party attachments?: A review of the cognitive mobilization thesis. Electoral Studies, 28(2), 248-260.
Andersen, R., Tilley, J. & Heath, A. F. (2005). Political knowledge and enlightened preferences: Party choice through the electoral cycle. British Journal of Political Science, 35, 285-302.
Bankert, A., Huddy, L., & Rosema, M. (2017). Measuring partisanship as a social identity in multi-party systems. Political Behavior, 39, 103-132.
Barker, D. C. & Hansen, S. B. (2005). All things considered: Systematic cognitive processing and electoral decision-making. The Journal of Politics, 67(2), 319-344.
Centraal Bureau voor de Statistiek (2018). Cijfers - Onderwijs. Retrieved from longreads.cbs.nl
Chaffee, S. H., Zhao, X., & Lesner, G. (1994). Political knowledge and the campaign of 1992. Communication Research, 21(3), 305-324.
Dalton, R. J. (1984). Cognitive mobilization and partisan dealignment in advanced industrial democracies. The Journal of Politics, 46(1), 264-284.
Dalton, R. J. (2004). Partisan mobilization, cognitive mobilization and the changing American electorate. UC Irvine: Center for the Study of Democracy.
Delli Carpini, M. X., & Keeter, S. (1993). Measuring political knowledge: Putting first things first. American Journal of Political Science, 37(4), 1179-1206.
Dassonneville, R. (2012). Electoral volatility, political sophistication, trust and efficacy: A study on changes in voter preferences during the Belgium regional elections of 2009. Acta Politica, 46(1), 18-41.
Eveland, W. P. & Hively, M. H. (2009). Political discussion frequency, network size, and ‘‘heterogeneity’’ of discussion as predictors of political knowledge and participation. Journal of Communication, 59, 205-224.
Geers, S., & Strömbäck (2018). Patterns of intra-election volatility: The impact of political knowledge. Journal of Elections, Public Opinion and Parties.
Greene, S. (2004). Social identity theory and party identification. Social Science Quarterly, 85(1), 136-153.
Irwin, G. A., & Van Holsteyn, J. J. M. (2008). What are they waiting for? Strategic
information for late deciding voters. International Journal of Public Opinion Research, 20(4), 483-493.
Kuhn, U. (2009). Stability and change in party preference. Swiss Political Science Review, 15(3), 463-94.
Linville, P. W. (1982). The complexity-extremity effect and age-based stereotyping. Journal of Personality and Social Psychology, 42(2), 193-211.
Luskin, R. C. (1990). Explaining political sophistication. Political Behavior, 12(4), 331-361. Mael, F., & Tetrick, L. (1992). Identifying organizational identification. Educational and
Psychological Measurement, 54, 813-824.
Mair, P. (2008). Electoral volatility and the Dutch party system: A comparative perspective. Acta Politica, 43(2-3), 235-253.
Moeller, J., & De Vreese, C. (2015). Spiral of political learning: The reciprocal relationship of news, media use and political knowledge among adolescents. Communication Research, 1-17.
Rosema, M., & Huddy, L. (2012). Attachment to Political Parties. LISS panel data archive. Van der Meer, T. W. G., Elsas, E., Lubbe, R, & Van der Brug, W. (2015). Are voters erratic,
whimsical or seriously picky? A panel study of 58 waves into the nature of electoral volatility (The Netherlands, 2006-2010). Party Politics, 21(1), 100-114.
Zaller, J. (1992). The nature and origins of mass opinion. Cambridge: Cambridge University Press.
Zaller, J., & Feldman, S. (1992). A simple theory of the survey response: Answering questions versus revealing preferences. American Journal of Political Science, 36(3), 579-616.
Appendix 1: Survey questions
We beginnen met enkele vragen over uw achtergrond. Hierna volgen er een aantal vragen over uw politieke voorkeur.
Wat is uw leeftijd?
Heeft u de Nederlandse nationaliteit? a. Ja b. Nee Wat is uw geslacht? A. Man b. Vrouw c. Overig
Wat is uw hoogst genoten opleiding? Dit is de hoogste opleiding die u heeft afgerond of waar u op dit moment mee bezig bent.
a. Basisonderwijs
b. Lager Beroeps Onderwijs, bijvoorbeeld VMBO (kader- en beroepsgerichte leerweg) of praktijkonderwijs
c. Lager Beroeps Onderwijs, bijvoorbeeld VMBO (kader- en beroepsgerichte leerweg) of praktijkonderwijs
d. Middelbaar beroepsonderwijs (MBO)
e. Hoger Algemeen en Voorbereidend Wetenschappelijk Onderwijs, bijvoorbeeld HAVO, VWO of HBS
f. Hoger Beroepsonderwijs en Wetenschappelijk Onderwijs, bijvoorbeeld HBO, Bachelor of Kandidaats
g. Master of doctoraal h. Geen onderwijs i. Anders, namelijk...
Hoe geïnteresseerd bent u over het algemeen in Nederlandse politiek? (schaal: 0 (helemaal niet geïnteresseerd) tot 8 (zeer geïnteresseerd)) Vindt u uzelf aanhanger van een bepaalde politieke partij? a. Ja
Van welke politieke partij bent u een aanhanger? a. VVD b. PVV c. CDA d. D66 e. GroenLinks f. SP g. PvdA h. ChristenUnie
i. Partij voor de Dieren j. 50PLUS
k. SGP l. Denk
m. Forum voor Democratie n. Anders, namelijk…
U heeft aangegeven geen aanhanger te zijn van een bepaalde politieke partij. Veel mensen die geen aanhanger zijn van een bepaalde partij, hebben toch een duidelijke voorkeur voor één partij boven anderen. Voelt u zich tot een bepaalde politieke partij meer aangetrokken dan tot andere politieke partijen?
a. Ja b. Nee
Tot welke politieke partij voelt u zich aangetrokken? a. VVD b. PVV c. CDA d. D66 e. GroenLinks f. SP g. PvdA h. ChristenUnie
i. Partij voor de Dieren j. 50PLUS
k. SGP l. Denk
m. Forum voor Democratie n. Anders, namelijk…
Als er vandaag verkiezingen voor de Tweede Kamer zouden zijn, op welke partij zou u dan stemmen? a. VVD b. PVV c. CDA d. D66 e. GroenLinks f. SP g. PvdA h. ChristenUnie
i. Partij voor de Dieren j. 50PLUS
k. SGP l. Denk
m. Forum voor Democratie n. Anders, namelijk… o. Ik zou niet stemmen p. Ik zou blanco stemmen
Nu volgen enkele uitspraken over de politieke partij die u zojuist heeft genoemd. Geef voor elke uitspraak aan in hoeverre deze op u van toepassing is.
(schaal: 1 (zelden of nooit) tot 8 (bijna altijd))
1. Als ik over deze politieke partij praat, dan zeg ik ‘wij’ in plaats van ‘zij’ 2. Ik ben geïnteresseerd in wat andere mensen van deze partij vinden.
3. Als mensen kritiek leveren op deze partij, dan voel ik mij persoonlijk beledigd. 4. Ik heb veel gemeen met andere mensen die deze partij steunen.
5. Als deze partij het slecht doet in een opiniepeiling, dan word ik daar chagrijnig van. 6. Als ik iemand ontmoet die deze partij steunt, dan voel ik me met die persoon verbonden. 7. Als ik over deze partij praat, dan verwijs ik ernaar als ‘mijn partij’.
8. Als mensen positief praten over deze partij, dan geeft me dat een goed gevoel.
Het volgende onderdeel bestaat uit een aantal kennisvragen over politiek. Veel mensen weten de antwoorden op deze vragen niet. Het is belangrijk dat u bij deze vragen niet overlegt met anderen en geen antwoorden opzoekt op internet. Als u het antwoord niet weet, selecteer dan ‘weet ik niet’. In totaal zijn er 12 vragen.
Welke van deze partijen zit in de huidige regering? a. GroenLinks
b. ChristenUnie c. PvdA
d. Forum voor Democratie e. Weet ik niet
Wie is op dit moment de voorzitter van de Tweede Kamer? a. Sigrid Kaag
b. Gerdi Verbeet c. Khadija Arib d. Marianne Thieme e. Weet ik niet
Welke functie heeft Stef Blok? a. Minister van Justitie en Veiligheid b. Minister van Defensie
c. Minister van Sociale Zaken en Werkgelegenheid d. Minister van Buitenlandse Zaken
e. Weet ik niet
Welke van deze partijen is over het algemeen het meest negatief over de Europese Unie? a. VVD
b. D66 c. PVV
d. GroenLinks e. Weet ik niet
Welke van deze partijen vindt dat de belasting op de hoogste inkomensgroepen omhoog moet? a. CDA b. SGP c. VVD d. SP e. Weet ik niet
Welke van deze partijen vindt dat er een bindend referendum moet komen? a. VVD
b. D66 c. CDA d. SGP
e. Weet ik niet
Hoe worden leden van de Eerste Kamer verkozen? a. Ze worden direct verkozen door Nederlandse kiezers b. Ze worden gekozen door leden van de Tweede Kamer c. Ze worden gekozen door leden van de Provinciale Staten d. Ze worden gekozen door leden van de gemeenteraden e. Weet ik niet
Hoe vaak worden er normaal gesproken gemeenteraadsverkiezingen gehouden? a. Elke 3 jaar
b. Elke 4 jaar c. Elke 5 jaar d. Elke 6 jaar e. Weet ik niet
Hoe kan de grondwet gewijzigd worden?
a. Alle ministers moeten unaniem de wijziging goedkeuren
b. Minstens 3/4 van alle ministers, Tweede Kamerleden en Eerste Kamerleden moeten de wijziging goedkeuren
c. De wijziging moet eerst aangenomen worden door een meerderheid van de Eerste en Tweede Kamer, en na de volgende Tweede Kamerverkiezingen moet de wijziging door minstens 2/3 van beide Kamers worden aangenomen
Waarom is het belangrijk dat er meer dan één politieke partij bestaat in een democratisch land?
a. Om verschillende meningen en belangen te vertegenwoordigen in het parlement b. Om politieke corruptie te beperken
c. Om politieke demonstraties te voorkomen d. Om economische concurrentie te stimuleren e. Weet ik niet
Stelt u zich voor: de overheid heeft de belasting verlaagd voor inkomens uit rente en investeringen en heeft de belasting verhoogd op salarissen. Een grote groep
demonstranten protesteert voor de overheidsgebouwen. De demonstranten zijn waarschijnlijk:
a. Mensen met veel geld op hun spaarrekening b. Mensen die aandelen van bedrijven bezitten
c. Mensen die uitkeringen ontvangen van de overheid d. Mensen die in fabrieken werken
e. Weet ik niet
Stelt u zich voor: een land heeft een dalend geboortecijfer en een toenemende levensduur. Welk probleem zal opgelost moeten worden?
a. Er moeten meer scholen worden gebouwd
b. Pensioenen voor ouderen moeten worden gefinancierd
c. Er moeten huizen gebouwd worden voor mensen met lage inkomens d. Misdaad en geweld moeten worden bestreden
Appendix 2: Political knowledge descriptives
Knowledge type
Question Incorrect Correct Don’t
know Current Welke van deze partijen zit in de huidige regering? 14.7% 81.1% 4.2% Current Wie is op dit moment de voorzitter van de Tweede
Kamer?
3.5% 88.8% 7.7%
Current Welke functie heeft Stef Blok? 17.0% 71.0% 12.0%
Party Welke van deze partijen is over het algemeen het meest negatief over de Europese Unie?
1.2% 95.8% 3.1%
Party Welke van deze partijen vindt dat de belasting op de hoogste inkomensgroepen omhoog moet?
3.9% 90.3% 5.8%
Party Welke van deze partijen vindt dat er een bindend referendum moet komen?
6.9% 64.9% 28.2%
Structural Hoe worden leden van de Eerste Kamer verkozen? 7.7% 85.7% 6.6% Structural Hoe vaak worden er normaal gesproken
gemeenteraadsverkiezingen gehouden?
7.7% 85.3% 6.9%
Structural Hoe kan de grondwet gewijzigd worden? 12.4% 69.5% 18.1% Civic Waarom is het belangrijk dat er meer dan één
politieke partij bestaat in een democratisch land?
2.3% 96.9% 0.8%
Civic Stelt u zich voor: de overheid heeft de belasting verlaagd voor inkomens uit rente en investeringen en heeft de belasting verhoogd op salarissen. Een grote groep demonstranten protesteert voor de
overheidsgebouwen. De demonstranten zijn waarschijnlijk:
19.3% 75.7% 5.0%
Civic Stelt u zich voor: een land heeft een dalend
geboortecijfer en een toenemende levensduur. Welk probleem zal opgelost moeten worden?
Appendix 3: Political party descriptives
Party N (of all groups) % (of all groups)
VVD 25 9.7% PVV 1 .4% CDA 7 2.7% D66 41 15.8% GroenLinks 89 34.4% SP 9 3.5% PvdA 28 10.8% ChristenUnie 9 3.5%
Partij voor de Dieren 34 13.1%
50Plus 2 .8%
SGP 0 0%
Denk 0 0%
Forum voor Democratie 4 1.5%
Other 3 1.2%
