Political tolerance in the Netherlands: in
decline? The role of the
libertarian-authoritarian dimension.
Bachelor’s thesis
Jochem Rietveld
18-06-2012
The price of the democratic way of life is a growing appreciation of people's differences, not merely as tolerable, but as the essence of a rich and rewarding human experience.
-Jerome Nathanson
Introduction
The Netherlands have long enjoyed a reputation for tolerance (Erk 2011, 110). The country has been receiving immigrants and political and religious refugees for centuries (McKay et all. 2008, 599). In the period of the Reformation, French Huguenots fled from religious persecution to the Netherlands (McKay et all. 2008, 599). In the same time Jews found a safe haven in Amsterdam, where they could freely engage in their commercial activities, whereas they were prosecuted in large areas of Europe (McKay et all. 2008, 550). In the 20th century,
after the Second World War, the Netherlands maintained their excellent reputation as a
tolerant country: big groups of immigrants went to the Netherlands in the sixties and seventies to find jobs in the booming postwar economy. However, things changed in the eighties: an economic downturn as a result of the oil crisis of the seventies meant that the job market soured (McKay et all. 2008, 1010-11). Immigrants (as well as Dutch workers) lost their jobs and ended up in the Dutch welfare system. In the nineties, Bolkestein, a VVD-politician questioned the immigration and integration problem for the first time in Dutch politics.1
In the early 2000s, Pim Fortuyn, a very charismatic, elite-challenging politician popularized the immigration and integration issues. With his nonconformist, confrontational style he proved to be hard to deal with for the political leaders of the governing parties. More strikingly, however, was Fortuyn’s message: he declared the end of multicultural Holland and he called Islam a ‘backward religion’.2 Surprisingly enough, Fortuyn attracted much support
in Dutch society: he appealed to a large group of voters that worried about the big influx of immigrants.3 As quick as Fortuyn entered the political scene, so abrupt was the end of his
1 Marc van Dijk, “De methode-Bolkestein en het eind van het integratiedebat,” Trouw, 28 september 2011.
2 Frank Poorthuis en Hans Wansink, “De Islam is een achterlijke cultuur,” De Volkskrant, 9 februari 2002.
political career: on May 6, 2002, just a few days before the national elections, Pim Fortuyn was killed by a animal rights activist.4
Several politicians, such as Rita Verdonk and Geert Wilders fought over Fortuyn’s heritage, with Geert Wilders’ PVV turning out to be the most successful. In 2006 the PVV joined the Dutch parliament and in 2010 the party performed really well in the national elections.5 After a long period of coalition formation, the PVV eventually provided the new
VVD-CDA with parliamentary support. The PVV stayed outside the government, so it did not provide any cabinet ministers. On the other hand, the PVV managed to get many of their election promises in the program of the new government. Strict proposals on immigration and Islam were adopted by the new government of prime-minister Rutte. 6
These PVV-measures are in stark contrast with any notion of tolerance. How is it possible that the PVV attracts so much support in Dutch society, if the Netherlands is still that nice, tolerant country? Although this is a very relevant question, this paper will not be limited to the role of the PVV in explaining levels of tolerance. It is an attempt to test a broader hypothesis about political tolerance. First of all, an overview will be provided of what
political tolerance means, how it has been applied in earlier research and what the findings of these studies were. Consequently, based on the works of Herbert Kitschelt of the 1990s it will be argued that the libertarian-authoritarian dimension that he introduced could help explaining variance in political tolerance. The specific characteristics of libertarian and authoritarian parties give rise to the hypothesis that voters of right-authoritarian parties are less likely to be tolerant than voters of left-libertarian parties.
The hypothesis will be put to the test by using the survey data that accompany this research. Given the limits of the data that were collected, it will be hard to establish clear-cut
4 NRC redacteuren, “Fortuyn vermoord, dader ‘milieuactivist’,” NRC Handelsblad, 7 mei 2002.
5 Arne Hankel, “Feest bij PVV om ‘grandioze’ verkiezingswinst,” Elsevier, 9 juni 2010.
6 NRC redacteuren, “Karakteristieken van het kabinet-Rutte,” NRC Handelsblad, 14 oktober 2010.
relations between the variables used in this research. Therefore, the results of the statistical analyses will need to be interpreted with caution. Further research with larger and more representative data sets is necessary in order to deal with the data problems that have constrained this study.
Literature review
Although tolerance is a popularly used word, there is some conceptual confusion. When one finds the word ‘tolerance’ in the Oxford Dictionary, it is defined as: “the ability or willingness to tolerate the existence of opinions or behaviour that one dislikes or disagrees with” (Oxford Dictionary 2012). On the other hand, the word ‘toleration’ is defined as: “the practice of tolerating something, in particular differences of opinion or behavior” (Oxford Dictionary 2012). It is established that the word toleration preceded the word tolerance, however, nowadays, almost anyone speaks about tolerance (Oxford Dictionary 2012). That is why in this paper, I will speak of ‘tolerance’ as opposed to ‘toleration’. More specifically, this paper will address the issue of ‘political tolerance’, which is already more specific than the broad and vague term ‘tolerance’.
Political scientists have had difficulty in defining ‘political tolerance’: they have used different definitions and different measurement techniques (Stouffer 1954; Sullivan et al. 1979). Sullivan et al. (1979) refer to Crick (1973) in their definition of political tolerance: “Tolerance implies a willingness to "put up with" those things that one rejects (Crick 1973. Politically, it implies a willingness to permit the expression of those ideas or interests that one opposes” (see Crick, 1973, Ch. 3) (Sullivan et al. 1979, 784).
Sullivan et all. expanded on Samuel Stouffer’s foundational work “Communism, Conformity and Civil Liberties: A Cross Section of the Nation Speaks its Mind.” In the
summer of 1954, Stouffer had carried out a national sample opinion survey of attitudes toward communism and civil liberties (Stouffer 1954). He found that, contrary to public belief,
Americans were not as tolerant as assumed: many of them were unwilling to extend civil rights to non-conformists in the US society, to communists, but also to socialists and atheists (Stouffer 1954). Although Stouffer’s study was impressive for the time, it was not without pitfalls: Stouffer’s sample was not completely representative for the American population, he used pre-determined target groups and he drew many inferences that could be called
‘interpretative’ at best (Janowitz 1956, 350-53). Especially, the use of pre-determined target groups (communists, atheists and socialists), which I will turn later to, turned out to a big issue in Stouffer’s research (Sullivan et al. 1979).
The shortfalls of Stouffer’s study made his results not less important or less shocking: substantial majorities in the American population were willing to curb the rights of
communists in many areas (Sullivan et all. 1979, 782). Communists should be restricted in their freedom of speech, they should not be allowed to teach in high schools or colleges and the government should be able to monitor them even if that would lead to privacy concerns (Sullivan et all. 1979, 782). The respondents were more tolerant towards the other target groups (socialists, suspected communists and atheists), but still, a considerable degree of intolerance persisted across target groups (Sullivan et all. 1979, 782). The results conflicted with the old assumption that there was a consensus in American society regarding procedural norms that allow extremist groups access to political institutions (Sullivan et all. 1979, 782).
Sullivan et all. (1979) later criticized Stouffer for misinterpreting tolerance: Stouffer had measured tolerance by asking respondents for their attitudes toward predominantly left-wing target groups (Stouffer 1954). Sullivan et all. improved Sullivan’s measure method, by introducing a “content-controlled method”: this means that the researchers did not provide the respondents with pre-determined target groups, but leave the choice of target groups to the respondent (Sullivan et all. 1979). The modification by Sullivan et all. was of major
and not an attitude towards a pre-determined supposedly unpopular group. Therefore, it was decided to use Sullivan’s content-controlled method in the survey of this research.
Stouffer’s and Sullivan’s research show general attitudes of Americans toward groups in society, but they do not establish linkages between partisanship and tolerance (Stouffer 1954, Sullivan et al. 1979). I hypothesize that party preference can help explain levels of political tolerance, as people that vote for the same party share certain characteristics (Goren 2005, 881). For instance, in the US, “the closer one feels toward social groups typically seen as Republican, such as business people or born again Christians, the more strongly one identifies with the Republican Party and the warmer one regards fellow partisans” (Goren 2005, 882). The question is: do core political values- such as political tolerance- shape party identification or the other way around? Goren has argued that party identification is more stable and enduring in the minds of citizens than core political values (Goren 2005, 881) This finding was consistent for the four core political values he looked at in the US: equal
opportunity, limited government, traditional family values and moral tolerance (Goren 2005, 883). Second, he found that party identification constrained beliefs about three of the four of his core political values (traditional family values excepted) (Goren 2005, 881). Finally, core political values did not turn out to constrain partisan identities (Goren 2005, 881). Although Goren did not take political tolerance as one of his ‘core political values’, his results do suggest “that partisanship shapes important long-term political predispositions as well” (Goren 2005, 894). In effect, by applying this rationale, partisanship could shape an abstract principle as political tolerance.
If partisanship could shape an abstract principle as political tolerance, what then could be one of the determinants of partisanship? I hypothesize that personality traits and political
preferences are related: as personality traits differ, political preferences may vary along. For the purposes of this research, it is helpful to make a distinction between dogmatic persons and non-dogmatic persons. In 1960, Rokeach defined dogmatism as “a closed way of thinking which can be associated with any ideology regardless of content, an authoritarian outlook of life, an intolerance toward those with opposing beliefs and a sufferance of those with similar beliefs” (Rokeach 1960, 4). Thus, it can be inferred that dogmatic people have difficulty with allowing different viewpoints to their lives and they tend to see the world in black and white terms. The Christian-right in the United States is an example of a group of people showing a dogmatic tendency: according to Cochran (1989) religion leads people “to divide the world into believers and non-believers, and often suggests the moral superiority of the former (Cochran 1989, 56). In their research on political tolerance among evangelicals in the US, Wilcox and Jelen (1990) signaled that Christian evangelicals were less tolerant of atheists, homosexuals and communists than the general American population. One of the explanations that is given for this intolerance by evangelicals is their “insistence on an infallible or inerrant Bible” (Guth et al. 1988, 380). The dogmatism tendency to divide the world in ‘good’ and ‘evil’ is something that not only characterizes Christian evangelicals. In 2006, right-wing PVV-leader Geert Wilders called Muslim immigration in the Netherlands ‘a tsunami of Islamists’7 by which he implicitly divided the world in ‘the good’ (native Dutch people) and
‘the bad’ (Muslim immigrants).
Woodberry and Smith came up with similar findings about tolerance among
Conservative Protestants in the United States: they argued that, even if scale biases would be resolved, CPs would still be less tolerant than other Americans (Woodberry and Smith 1998, 41). Two findings stand out: first, most of the intolerance seemed to result from
fundamentalism rather than from religious orthodoxy (Green et al. 1994; Kirkpatrick 1993). Second, tolerance was also influenced by how much a group felt marginalized and threatened (Green et al. 1994; Sullivan et al. 1982, 251). The importance of these findings is that
fundamentalism and marginalization of groups are independent of religion: put differently, secular groups can be fundamentalist and marginalized as well. The extreme policy platforms that PVV-leader Wilders regularly campaigns on can be called fundamentalist: in September 2009, he proposed a tax on religious scarves and in February 2012, he introduced a hotline for nuisance caused by East-European labor migrants.89 These proposals are clearly dogmatic:
they display a closed way of thinking- as if only East-European labor migrants could create nuisance- and intolerance toward those with opposing beliefs (in this case probably Muslims who prefer to wear a headscarf). Since PVV-voters in the Netherlands and Conservative Protestants in the US seem to share a degree of dogmatism and fundamentalism, I hypothesize that PVV-voters will express at least some sort of intolerance as well.
The trouble is that the PVV is only one party in Dutch politics, which could lead to a too narrow hypothesis if one wants to assess the state of tolerance in the Netherlands on a larger scale. Therefore it has to be found out if the PVV falls into a broader category of political parties, so that a more general hypothesis can be formed. An answer to this question can probably be found in Kitschelt’s influential works of the early 1990s. Kitschelt argued that a new characterization of left-right orientations in Europe was needed, because of
“structural changes in mature democratic capitalism” (Kitschelt 1992, 13). Kitschelt signaled that “sweeping social change had shifted the main axis of voter distribution from a simple alternative between socialist (left) and capitalist (right) politics to a more complex
configuration opposing left-libertarian and right-authoritarian alternatives” (Kitschelt, 1994: 30–1). Second, he argued that political parties associated with the political left increasingly
8 Redactie Politiek, “Wilders wil ‘kopvoddentaks’,” Trouw, 16 september 2009.
9 Jeroen Langelaar, “Ruim 46.000 klachten bij PVV-meldpunt Oost-Europeanen,” 18 februari 2012.
expressed stronger support for libertarian values than did political parties identifying with the political right (Kitschelt, 1994: 30–2). “Conversely, respondents on the political right are more likely to support authoritarian values” (Kitschelt, 1994: 30–2).
Some of the terms Kitschelt uses need further clarification. With left-libertarian values, Kitschelt means the values that were primarily popularized by Inglehart, such as autonomy and self-expression, personal liberty and quality of life (Inglehart 1977). ‘Right-authoritarian alternatives’ is an indication for economic security, law and order and a more hierarchical society (Kitschelt 1994). On this dimension, the PVV would score as a right-authoritarian party as it promotes a dominant Dutch culture, it opposes immigration and it supports law-and-order oriented measures, such as an increase in the number of police officers with 10,000 (Party platform PVV 2010). Moreover, the dogmatism that is apparent in many of the PVV proposals seems to be partly reflected in the values that Kitschelt’s right-authoritarians held high. An emphasis on law and order and a hierarchical society seem to fit in Rokeach’s definition of dogmatism in which the emphasis on law and order resembles “an intolerance toward those with opposing beliefs” and the emphasis on a hierarchical society stems from “an authoritarian outlook of life” (Rockeach 1960, 4). It is this dogmatic tendency of right-wing authoritarian parties that creates the expectancy that voters for these parties will express lower levels of tolerance. Therefore my hypothesis is: voters of right-authoritarian parties are less likely to be tolerant than voters of left-libertarian parties. In this hypothesis, political preferences (as a proxy of dogmatism) are thus expected to help explain political tolerance.
For my research about political tolerance in the Netherlands, a survey has been conducted. The survey consisted of questions about political tolerance, party preference and some demographic control variables, such as age, gender and social class. The results of the survey should show if there is a relationship of any kind between party preference and political tolerance. Moreover, the results will put my hypothesis to the test: do the survey data confirm my hypothesis hat right-authoritarian voters are less tolerant than left-libertarian voters?
With my study, I attempt to replicate Sullivan et all.’s (1979) research. Studies like Sullivan’s and Stouffer’s (who laid the foundation in the field) have, as far as I know, never been conducted in the Netherlands. The Netherlands form the ideal case for political tolerance research: the Dutch have long been known for being tolerant, so in that way, the country could be a ‘critical case’ (Erk 2011, 110). If people in the Netherlands fail to be tolerant, where else could tolerance then be expected? The literature review established the main hypothesis of this research: “People who vote for right-authoritarian parties are less likely to be tolerant than voters of left-libertarian parties.”
The survey that formed the basis of my statistical analyses was created and published on the website of uleidenss.eu.qualtrics.com. This website is specifically designed for students and professors at Leiden University to assist them in their survey research. The survey was launched on April 27 and it was closed on May 15. In this period of time, 196 persons completed the survey. The group that completed the survey primarily consists of students, friends, family and other relatives. I have put the survey link on my Facebook wall and several friends of mine have shared the link on theirs. Besides, I have e-mailed all my contact persons on Hotmail. Finally, I have asked people at the Faculty of Social Sciences and first year Political Science students at the Kamerling Onnes building to fill in my survey. In total, I have gathered 145 online respondents and 48 respondents in the different departments in Leiden. It must be emphasized that my sample is not representative of the Dutch population:
the respondents are predominantly younger, because most of them were students. My respondents are higher educated than average Dutch people, because, again, most of them were students. Finally, since there were so many Political Science students in my dataset, it can be held that my respondents were more politically interested than the average Dutchman.
As said before, the motivation to conduct a survey stems from earlier research in the field: both Stouffer (1954) and Sullivan et all. (1979) conducted a survey in their political tolerance studies in the United States. Stouffer conducted his study in the heydays of the Cold War. He was interested in tolerance in US society, because it was assumed that there was common agreement on giving political rights to all groups in society. To put this assumption to the test, Stouffer gathered survey data asking people for their attitudes towards
communists, socialists and atheists. Stouffer assumed these groups were the most
nonconformist and least-liked groups in US society. The results of his study were rather bleak: it turned out that tolerance was not widely celebrated in the US. Especially communists were the targets of the respondents: a majority was not willing to provide them with basic rights as the right to speak and participate in politics (Stouffer 1954). Sullivan et all. (1979) built on Stouffer’s work a quarter of a century later. Many things had changed in US society in the meantime: protest groups as the civil rights movement and the women’s right movement had emerged. At the same time, the Ku Klux Klan had reappeared on the scene: with political murders, they tried to defy the landmark Brown v. Board 1954 Supreme Court decision (Ball 2004). In the sixties and seventies there was speak of a diversification of possible target groups of intolerance: progressive people were abhorred by the extreme violence of the Ku Klux Klan, while conservative people were still suspicious of communists. Sullivan et all. (1979) recognized that these things had changed and they introduced the so called
opposed to the researcher. This new method highly increased the validity of the tolerance measure.
In this study, the content-controlled method that was introduced by Sullivan et all’s (1979) has been used. The high validity of this measurement method can be illustrated with the following example: a GroenLinks voter may pick a right-wing extremist as his target group, while a PVV-voter might pick a Muslim as his group. In both cases, tolerance, as opposed to agreeableness, will be measured. The survey starts by asking respondents to choose or fill in their most objectionable group in Dutch society. Consequently, the
respondents have answered questions that measure their tolerance towards these targets (see figure 1).
Figure 1: questions about political tolerance
Translated from Dutch to English
1. Which group in Dutch society do you consider most objectionable? (one can think of any group, including but not limited to: political, religious, ethnical, ideological groups)
Right-wing extremists Animal rights activists Communists
Immigrants Other (specify)
2. Should a (member of most objectionable group) be allowed the right to speak? Yes/no
3. Should a (member of most objectionable group)be allowed to teach at schools? Yes/no
4. Should a book of a (member of most objectionable group)be removed from the libraries?
Yes/no
5. Should the (members of most objectionable group)be impeded from
participating in Dutch politics? (on all levels: local, provincial and national) Yes/no
6. (Members of most objectionable group) should be allowed to participate in public rallies in your municipality
Yes/no
7. The (members of most objectionable group) should be monitored by the government (one can think of wiretapping or internet surveillance) Yes/no
On May 15, the survey was closed and the outcomes were ready for statistical analysis. So, first of all, a tolerance variable out of all the questions that tapped on tolerance has been created. All tolerance questions could be answered with ‘yes’ or ‘no’. Sometimes the ‘yes’ stood for the ‘tolerant answer’, at other times ‘no’ resembled tolerance. To account for these inconsistencies, 3 questions were recoded (about the possibility of political
participation, the removal of books from libraries and governmental monitoring of suspect groups), so that the ‘no’ answer would be coded as one (as tolerant) and the ‘yes’ answer as two (less tolerant). After the recoding, all the values of ‘1’ were a sign of tolerance and all the values of ‘2’ marked intolerance. Finally, the 6 tolerant questions were combined, so that they would form a(n) (in)tolerance scale. This new variable could have values ranging from ‘6’ (very tolerant) to ‘12’ (very intolerant). Besides, a libertarian-authoritarian party dimension was created. As I was unable to track down an existing libertarian-authoritarian Dutch party
categorization, I created one myself. This obviously makes the scale arbitrary and any results springing from this scale must therefore be interpreted with caution.
Figure 2: Libertarian-authoritarian scale Dutch political parties
1= most left-libertarian, 12= most righ-authoritarian 1= GroenLinks
2= Partij voor de Dieren 3= D66 4= PvdA 5= SP 6= VVD 7= CDA 8= ChristenUnie 9= SGP 10= Trots op Nederland
11= Onafhankelijke Burgerpartij (Hero Brinkman) 12= PVV
Finally, an ‘age category variable’ has been created out of the ‘age’ variable. I have made a 0 through 17 category for children, a 18 through 24 category for adolescents, a 25 through 44 for people in their early career, a 45 through 64 for people in their late career and 65+ for retirees. Figure 3 shows the distribution of age over the categories.
Figure 3: Age categories
Age category Frequency Percent
________________________________________________ 0-17 3 1,5 18-24 152 67,5 25-44 22 11,2 45-64 13 6,6 65+ 2 1,0 Missing 4 2,0 Total 196 100,0 Results
Before any statistical tests will be conducted, it is helpful to take a quick look at the distribution of respondents over the different political parties.
Figure 4: Which party would you vote for, if there were parliamentary elections tomorrow?
Frequency Percent Valid CDA 7 3,6 PvdA 20 10,4 VVD 36 18,8 GroenLinks 26 13,5 SP 11 5,7 D66 58 30,2 ChristenUnie 3 1,6 SGP 1 ,5
Partij voor de Vrijheid (Geert Wilders)
8 4,2
Partij voor de Dieren 4 2,1
Onafhankelijke Burger Partij (Hero Brinkman)
1 ,5
Andere partij (other party) 1 ,5
Weet niet (don’t know) 13 6,8
Geen stem (not vote) 3 1,6
Total 192 100,0
As one can see, the results of my research are restrained by a relatively small and unrepresentative dataset. Compared to the 2010 national election outcomes, there are far too few CDA, PvdA, SP and PVV voters, while D66 and Groenlinks voters are heavily
overrepresented (Kiesraad 2010). With that being said, the actual statistical analyses can be conducted.
To test the main hypothesis- right-authoritarian voters are less likely to be tolerant than left-libertarian voters-, a linear regression has been conducted. In this regression analysis the dependent variable was the tolerance score ranging from 6 (tolerant) to 12 (intolerant) and the different independent variables were the libertarian-authoritarian scale, gender, age categories and social class.
Figure 5: Model Summary
Model R R Square
Adjusted R Square
1 ,190a ,036 ,013
a. Predictors: (Constant), Social class, Age categories, Liberarian-authoritarian scale. A
Figure 4 shows the model summary of the regression analysis. The R-square is only . 036 which means that the four independent variables- the libertarian-authoritarian scale, gender, age categories and social class can only explain 3.6% of the variance in the dependent variable ‘tolerance score’. Moreover, the ANOVA table (not displayed) indicates that the amount of variance that can be explained with the regression model has an α=.192, which is clearly not significant at a p<.05 level. Therefore, it can be said that the regression model is not significantly better at predicting changes in political tolerance than having no model (Field 2005, 213).
Figure 6: Coefficients of the political tolerance regression model
Sig. B SE B β (Constant) 7,038 ,816 ,000 Libertarian-authoritarian ,055 ,048 ,090 ,251 Age categories ,271 ,203 ,103 ,185 Gender -,236 ,247 -,074 ,341 Social class ,191 ,163 ,091 ,240
a. Dependent Variable: Tolerance_score
The β-values of figure 5 indicate the change in standard deviations of political
standard deviations of one on the libertarian-authoritarian scale leads to a 0.09 increase in standard deviations on the political tolerance variable. This means that when one supports a more authoritarian party (moving up on the libertarian-authoritarian scale), one becomes more intolerant (the 0.09 increase reflects increasing intolerance, as the tolerance variable was rated from 6=tolerant to 12=intolerant). This coefficient is, however, insignificant (α=.251). The other coefficients can be interpreted in a similar way: the coefficient of age is positive, which means that when one increases in age category, political intolerance increases as well. The coefficient of gender is negative, which means that when one moves up on the gender variable (so from 1=male to 2= female), political intolerance decreases. Finally, when one moves to a lower class (1= highest class and 5= lowest class), political intolerance increases with 0.163 standard deviations. However, since all beta coefficients are not significant at a p<0.05 level, these changes in predictors do not significantly predict changes in levels of tolerance.
The simple linear regression model has shown that the libertarian-authoritarian scale, the age category variable, gender and social class are not significant predictors of the
tolerance scale variable. However, it is possible that these predictors have a significant effect on the separate tolerant variables, such as public speech and the removal of books from libraries. With binary logistic regression models the effect of the predictors can be tested for each tolerance question. Again, the variables ‘libertarian-authoritarian scale’, age categories, gender and social class will be used as predictors.
Since showing all binary logistic regressions would result in overly lengthy analyses, at this point it needs to be chosen which binary logistic regressions will be analyzed. I choose for two tolerance questions on which an overwhelming majority picked one answer (the ‘public speech’ and the ‘removal of books from library’ variables) and for two tolerance questions on which the respondents were most divided (the ‘public teaching’ and the
member of their most objectionable group should be allowed to speak in public, whereas 27 people answer that a member should not be allowed to do this. When respondents are asked if books written by members of their most objectionable group should be removed from
libraries, 28 people agree (14,7%) and 163 people disagree (85,3%). On the ‘public teach’ issue (“should members of your most objectionable group be allowed to teach?”), 105 respondents say ‘yes’ (54.6%) and 87 people say ‘no’ (45.3%). On the ‘government
monitoring’ issue (“should members of your most objectionable group be monitored by the government?”) 101 people answer ‘yes’ (52.6%) and 91 people answer ‘no’ (47.7%).
When one conducts a binary logistic regression with the binary dependent variable ‘public speech’, the Omnibus Tests of Model Coefficients provide the likelihood that the adding of the different variables- the ‘libertarian-authoritarian scale’, age categories, gender and social class- to the model has not significantly increased the ability to predict the decision made by the subjects (their ‘yes’ or ‘no’ to public speech). The table gives a Chi-Square value of 12.339 with p=0.015. Since the likelihood that there has not been an increase in predictive ability is smaller than α=0.05, it can be hold that the predictors have significantly increased the ability to predict the decision made by the respondents.
Of the individual predictors, only the ‘libertarian-authoritarian scale’ (p=0.009) was statistically significant on a α=0.05 level, with ‘sex’ being almost significant (p=0.058). The odds ratio of the ‘libertarian-authoritarian scale’ is 1.224 which means that a one point increase on the ‘libertarian-authoritarian scale’ is associated with the odds of saying ‘no’ to the allowance of public speech increasing by a multiplicative factor of 1.224. Put differently, as one’s support for authoritarian parties rises, the odds that the person in question will deny objectionable groups their right to speak in public, increase as well.
The next binary dependent variable is the ‘removal of books from libraries’ question. This time the Omnibus Tests of Model Coefficients give a Chi-Square value of 5.421 with a
p=.247. This means that the likelihood that the predictors have not increased the ability of the model to predict the respondents’ answers is .247, which is much higher than the standard significance level of α=0.05. So, it can be said that the predictors have failed to increase the predictive ability of the model. This is reflected in the odds table, in which all individual predictors fail to gain statistical significance. Only the ‘age categories’ variable comes close with p=0.099.
A binary logistic regression on the ‘public teach’ variable with the same predictors as mentioned earlier, results into Omnibus Tests of Model Coefficients giving a Chi-Square value of 2.773, df=4 and p=0.597. This means that the likelihood that the predictors have not increased the predictive ability of the model is .597. Since this value is much higher than the α=0.05 level, the predictors have, in this case, failed to increase the ability to predict the decision made by the respondents on the ‘public teach’ variable. This is, again, also reflected in the fact that none of the individual predictors’-‘libertarian-authoritarian scale’, age
categories, gender and social class- meets the conventional .05 standard for statistical significance.
Finally, a binary logistic regression with ‘government monitoring’ as dependent variable is conducted. This time the Omnibus Tests of Model Coefficients give a Chi-Square value of 5.859, df=4 and p=0.210. This means that the likelihood that the predictors have not increased the predictive ability of the model is .210. Since this value is again much higher than the α=0.05 level, the predictors have failed to increase the ability to predict the decision made by the respondents on the ‘government monitoring’ variable. Surprisingly enough, one of the individual predictors does turn out to be statistically significant. The variable ‘age category’ has an odds ratio of 0.571 and a p=0.048, which falls below the α=0.05 level. Therefore, it can be said that for an increase of one on the ‘age category’ variable, the odds of saying ‘no’ to the government monitoring question are almost cut by a half (.571) So, the
odds to disagree with government monitoring decrease, when one moves up on the age category variable.
Discussion
With the regression analyses being conducted, one can proceed with the implications of the test results. What do the test results tell about a relationship between political tolerance and the ‘libertarian-authoritarian scale’ and what is the role of the control variables?
First of all, the independent variables- -‘libertarian-authoritarian scale’, age categories, gender and social class- fail to help explain the political tolerance scale in the simple linear regression model. The R-square of .036 is very low: only 3.6% in the variance of the political tolerance scale could be explained by the different independent variables. However, these results could change if a more rigorous ‘libertarian-authoritarian dimension’ of Dutch political parties would be created. The creation of such a scale is clearly a topic for further research.
The coefficients in the linear regression model are in the expected directions, but all fail statistical significance. The β-scores of the libertarian-authoritarian scale, age category and social class are all positive, indicating that an increase in support of authoritarian parties, ageing and a decrease in social class (social class was coded from 1=high to 5=low)
correspond with higher levels of intolerance. Gender is the only variable with a negative β, indicating that a hypothetical change from male to female would correspond with lower levels of intolerance. For establishing true relations, it would be interesting to see if the coefficients could obtain significance, while maintaining their direction with larger and more
representative datasets.
The four binary logistic regressions have provided some statistically significant results: first of all, for the ‘public speech’ variable, the predictors did increase the predictive ability of the model. The libertarian-authoritarian scale was in this case the only individual
predictor that increased the predictive ability of the tolerance question at stake. This finding is important, as it is the only one that directly confirms the main hypothesis of this research. The odds ratio of the ‘libertarian-authoritarian scale’ was a positive 1.224, which means that a one point increase on the ‘libertarian-authoritarian scale’ is associated with the odds of saying ‘no’ to the allowance of public speech increasing by a multiplicative factor of 1.224. To say it more bluntly: when one’s support for authoritarian parties increases, one is more tempted to deny allowance of public speech to objectionable groups. In this specific case of public speech, the hypothesis that right-authoritarian voters are less likely to be tolerant than left-libertarian voters, can thus be confirmed.
The other three binary logistic regressions- with ‘removal of books from library’, ‘public teach’ and ‘government monitoring’ as binary dependent variables- gave rise to more mixed results: the combined predictors in these regressions failed to increase the predictive ability of the models. So, the predictors did not succeed in showing which answer category was most likely to be picked by the respondents. However, in the case of ‘government
monitoring’ the individual predictor ‘age-categories’ did gain statistical significance and with an odds-ratio .571, meaning that when people move up one category on the age-ladder, the odds that they support government monitoring of objectionable groups almost doubles.
The four binary regressions also show that the strength and significance of the
predictors are not related to the extent of division or agreement on a tolerance issue. One issue on which people predominantly agreed provided statistical results (public speech) and one issue on which people were divided showed a significant individual predictor (government monitoring). This works the other way around as well, as the ‘removal of books from library’ (issue with high agreement) and the ‘public teach’ (issue with high division) binary
To conclude, it can be hold that based on this dataset and these statistical analyses there is only marginal support for the main hypothesis of this research. Only on the public speech issue could the libertarian-authoritarian dimension significantly predict the outcome of the tolerance question. This does not mean that the hypothesis can easily be discarded: the right to speak in public is an important democratic right, and if the libertarian-authoritarian dimension can help showing how people regard this issue of tolerance, then that is hopeful for the main hypothesis (USinfo.org 2012). Also, as said before, a scientifically-based
libertarian-authoritarian dimension of Dutch political parties needs to be established to make truly meaningful statements about its relationship with political tolerance. Besides, a much larger and more representative data set will be required to make causal inferences. Given the clear limits in size and representativeness of this data set, the establishment of any causal inferences here was quite cumbersome.
Literature
Adorno, T., Frenkel-Brunswick, E., Levinson, D., & Sanford, P. 1950. The Authoritarian Personality. New York: Harper.
Ball, Howard. 2004. Murder in Mississippi: United States v. Price and the Struggle for Civil Rights. US: University Press of Kansas.
Cochran, Clark E. 1989. “Normative dimensions of religion and politics. In C. W. Dunn (ed.), Religion in American politics. Washington, DC: Congressional Quarterly Press, 39-50.
Crick, Bernard. 1973. Political Theory and Practice. New York: Basic Books.
Erk, J. (March 2011). The Famous Dutch (In)tolerance. Current history: a monthly magazine of world affairs, vol. 110 (734).
Field, Andy. 2005. Discovering statistics using SPSS. 2nd ed.. London: Sage. Goren, Paul. 2005. “Party Identification and Core Political Values.” American
Journal of Political Science 49 (4): 881–896.
Green JC, Guth JL, Kellstedt LA, Smidt CE. 1994. “Uncivil challenges?: Support for civil liberties among religious activists.” J. Polit. Sci. 22: 25-49.
Guth, James L., Ted G. Jelen, Lyman A. Kellstedt, Corwin E. Smidt, and Kenneth D. Wald. 1988. “The politics of religion in America: Issues for investigation.” American Politics Quarterly 16: 357-97.
Inglehart, Ronald. 1977. The Silent Revolution. Princeton, NJ: Princeton University Press.
Janowitz, Morris. 1956. “Communism, Conformity and Civil Liberties: A Cross Section of the Nation Speaks its Mind. by Samule Stouffer. Review by Morris Janowitz.” The Public Opinion Quarterly 20 (1): 350-353.
Kiesraad. 2010. “Uitslag verkiezing leden Tweede Kamer van 9 juni 2010.” 16 juni
http://www.kiesraad.nl/nl/Actueel/Nieuwsberichten/(2047)-Actueel-Nieuwsberichten-2010/Uitslag_verkiezing_leden_Tweede_Kamer_van_9_juni_2010.html (June 18, 2012).
Kirkpatrick, LA. 1993. “Fundamentalism, Christian orthodoxy, and intrinsic religious orientation as predictors of discriminatory attitudes.” J. Sci. Stud. Relig. 32 (3): 256-68.
Kitschelt, Herbert. 1992. “The Formation of Party Systems in East Central Europe”, Politics and Society 20: 7–50.
Kitschelt, Herbert. 1994. The Transformation of European Social Democracy. Cambridge: Cambridge University Press.
Kitschelt, Herbert. 1995. The Radical Right in Western Europe: A Comparative Analysis. Ann Arbor, MI: The University of Michigan Press.
Lawrence, David. 1976. "Procedural Norms and Tolerance: A Reassessment." American Political Science Review 70: 80-100.
McClosky, Herbert. 1964. "Consensus and Ideology in American Politics." American Political Science Review 58:361-82.
McKay, Hill, Buckler, Crownston and Wiesner-Hanks, eds. 2008. A History of Western Society, since 1300. Boston: Houghton Mifflin Company.
NRCNext.nl. 2012. “Next checkt: ‘Anti-immigratiepartijen als de PVV gaan 3 à 4
verkiezingen mee’.” 20 april. http://www.nrcnext.nl/blog/2012/04/20/next-checkt-anti-immigratiepartijen-als-de-pvv-gaan-3-a-4-verkiezingen-mee/ (May 23, 2012).
Nunn, Clyde A., Harry J. Crockett and J. Allen Williams. 1978. Tolerance for Nonconformity. San Francisco: Jossey-Bass
Oxford Dictionary. 2012. “Oxford Dictionaries: The World’s most trusted dictionaries.” http://oxforddictionaries.com/ (visited on June 17, 2012).
Prothro, James W., and Charles W. Grigg. 1960. "Fundamental Principles of Democracy: Bases of Agreement and Disagreement." Journal of Politics 22: 276-94.
PVV. 2010. “De agenda van hoop en optimisme. Een tijd om te kiezen: PVV 2010-2015.”http://www.pvv.nl/images/stories/Webversie_Verkiezings
ProgrammaPVV.pdf (June 17, 2012).
Rokeach, M. 1960. The Open and Closed Mind. New York: Basic Books.
Stouffer, Samuel. 1954. Communism, Conformity and Civil Liberties. New Brunswick, NJ: Transaction Publishers.
Sullivan, John L., James Piereson, and George E. Marcus. 1979. “An Alternative
Conceptualization of Political Tolerance: Illusory Increases 1950s-1970s.” American Political Science Review 73(3): 781-894.
Sullivan, JL, Piereson J, Marcus G. 1982. Political Tolerance and American Democracy. Chicago: University Chicago Press.
USInfo.org. 2012. “What is democracy? Rights.” http://usinfo.org/mirror/ usinfo.state.gov/products/pubs/whatsdem/whatdm3.htm (June 18, 2012). Wilcox, Clyde and Ted Jelen. “Evangelicals and Political Tolerance.” American
Politics Research 18: 25-46.
Woodberry, Robert D. and Christian S. Smith. 1998. “FUNDAMENTALISM ET AL: Conservative Protestants in America.” Annu. Rev. Sociol. 24: 25-56.
Appendix A
Questionnaire survey
Translated from Dutch to English
1. Which group in Dutch society do you find most objectionable? (one can think of any group, including but not limited to: political, religious, ethnical, ideological groups)
Right-wing extremists Animal rights activists Communists
Immigrants Other (specify)
2. Should a (member of most objectionable group) be allowed the right to speak? Yes/no
3. Should a (member of most objectionable group)be allowed to teach at schools? Yes/no
4. Should a book of a (member of most objectionable group)be removed from the libraries?
Yes/no
5. Should the (members of most objectionable group)be impeded from
participating in Dutch politics? (on all levels: local, provincial and national) Yes/no
6. (Members of most objectionable group) should be allowed to participate in public rallies in your municipality
Yes/no
7. The (members of most objectionable group) should be monitored by the government (one can think of wiretapping or internet surveillance) Yes/no
8. If there were parliamentary elections tomorrow, which party would you vote for? CDA PvdA VVD GroenLinks SP D66 ChristenUnie SGP
Partij voor de Vrijheid (Geert Wilders) Partij voor de Dieren
Trots op Nederland (Rita Verdonk)
Onafhankelijke Burger Partij (Hero Brinkman) Blanco (blank vote)
Andere partij(other party) Weet niet (don’t know) Geen stem (no vote) 9. What is your age? (Age in years) 10. Gender (male/female)
11. People sometimes speak of the existence of social classes. If you have to choose, with which social class would you affiliate?
The upper class
The upper middle class The middle class
The higher working class The working class
