The marginalization of ethnic minorities
through a group formation perspective
A comparison of marginalization between the
Suriname/Antillean and the Turkish community in Amsterdam
through the lens of group formation and group behaviour
Thesis by
Sipke Bontekoe
community in Amsterdam through the lens of group formation and group behaviour.
Author: Sipke Bontekoe Student number: 10785582
University: University of Amsterdam Thesis programme: Politics of identity Tutor: Mike Medeiros
Word count: 9561 (excluding this page, footnotes and appendixes) Date submitted: 23 June 2018
Abstract: This paper contributes to the analysis of the position of the older immigrant minority groups in Western Europe by analysing group formation and the behaviour of the formed minority groups. Council elections exit poll data and interviews are used to analyse the position of the Surinamese/Antillean group and the Turkish group in Amsterdam. Their shift in voting from established labour party PvdA to progressive activist parties DENK and BIJ1, shows that marginalization is not a salient explanation for the group positions of these minorities. Rather, these two ethnic minorities are currently found in an equilibrium where they are content enough to stay where they are, even though they are marginalized.
1. Introduction 1
2. Theoretical framework 2
2.1 Marginalization 2
2.2 Group formation 5
2.3 Behaviour 7
3. The case of Amsterdam 8
3.1 Selection 8
3.2 The Surinamese/Antillean and the Turkish community 9
3.3 Political behaviour 10 4. Hypothesis 12 5. Quantitative analysis 12 5.1 Data 13 5.2 Operationalization 13 5.4 Research design 15 5.5 Results 16 6. Qualitative analysis 20 6.1 Sources 20 6.2 Interview strategy 20
6.3 Van Gom and Sordam about the Surinamese/Antillean community 21
6.4 Yilmaz about the Turkish community 22
6.5 Interview results 23
7. Conclusion and discussion 24
Sources and Literature 26
Appendix 1: Logistic models results 35
Appendix 2: Cross tab Ethnicity and Voting behaviour 36
Appendix 3: SPSS package 37
Appendix 4: STATA package 37
Appendix 5: Interview transcription Van Gom and Sordam 38
Appendix 6: Interview transcription Yilmaz 54
1. Introduction
In 1963 Martin Luther King Jr. gave his famous speech: “I have a dream”. In many cities in the United States uproars broke out amongst blacks against their oppression. The marginalized position of blacks drove this minority to politicize and escalate the status quo (King 2010). West-European states did not see such uproars from visible minorities since World War II. The answer seems simple: in Western Europe visible minorities consist out of immigrant groups and their position is not as bad (Weldon 2006). But, looking at wealth distribution averages the position of these groups is unequal to that of natives (Hooijer & Picot 2015: 1880). With the new wave of immigration in the last five years it is necessary for West-European countries to take lessons from what has become of the older immigrant groups (Geddes & Scholte 2016: 1). From an ethical viewpoint these groups deserve an equal place in society and from an economic viewpoint it would be beneficial to have them economically integrated (Bonin 2017; Dworkin 2000; Edding 2013; Rawls 1971).
This paper contributes to a further analysis of the position of the older, immigrant, minority groups by taking a closer look at their marginalization. So far, research into the marginalization of these groups has mainly been limited to a comparison of wealth, health and education level averages between ethnic and native communities (e.g., Arango & Baldwin-Edwards 2014
;
Bonin 2017; Geddes & Scholte 2016; Hooijer & Picot 2015; Musterd & Ostendorf2013;
Nielsen & Krasnik 2010; Nolan& Whelan2010;
Phillips 2010; Townsend2014). To add to the current knowledge, this paper takes a different angle and seeks to analyse group formation and the behaviour of the formed minority groups. Through this approach, this paper will explain more about the position of immigrant minorities in society, specifically whether their position can be explained by marginalization or ethno-cultural factors (Guibernau & Rex 2010).To do so the case of Amsterdam has been selected. At the Amsterdam council elections of March 2018 there was a remarkable shift in voting within the three biggest, older, ethnic immigrant groups: Moroccan, Surinamese/Antillean and Turkish (Gemeente Amsterdam OIS 2018; Kranendonk et al. 2018; Vermeulen & Michon 2018). Ever since these immigrant groups started forming in Amsterdam they have been predominantly voting for the established labour party PvdA (Rath 1988: 635; Vermeulen & Michon 2018). However, in 2018 the two progressive activist parties DENK and BIJ1, with leadership from members of these immigrant groups, were able to draw many votes of the Moroccan,
Surinamese/Antillean and Turkish groups (Kranendonk et al. 2018; Vermeulen & Michon 2018).
This paper analyses to what extent marginalization is explanatory for this shift, which exposes more about the extent these groups are marginalized and thus their position in society. To narrow the scope for this paper, the comparison will be made between Surinamese/Antillean and the Turkish group in contrast to Dutch native voters. These two groups show the least similarities in background and culture, and in their recent shift in voting behaviour. The Surinamese/Antillean group was dispersed in their voting with only 18% in favour of BIJ1, whereas 74% of Turkish voters voted for DENK (Vermeulen & Michon 2018).
This all comes together in the following research question: What is the current role of marginalization in relation to political mobilization within the Surinamese/Antillean and the Turkish community in Amsterdam? This will be researched with a mixed method approach. The quantitative part exists out of a logistic regression analysis of exit poll data from the Amsterdam council elections in March 2018, where the relation between ethnic background and party choice is interacted with education and in- and out-group attitudes. In addition, two interviews with prominent people from the Surinamese/Antillean community and Turkish community create more insight in the perception of marginalization within these groups and explain some behaviour.
This paper first creates a theoretical framework concerning marginalization and group formation, whereupon is looked at the case of Amsterdam. On this basis, the hypothesis for this paper steers towards a positive relationship between marginalization and mobilisation. The results of the quantitative section show that marginalization plays a minor role in voting for BIJ1 or DENK. The interviews show this can be explained by disunity that finds it roots in relative contentness for the Surinamese/Antillean community, whereas for the Turkish community it can be explained by political preference. This paper concludes that these two ethnic minorities are currently found in an equilibrium where they are content enough to stay where they are, even though they are marginalized.
2. Theoretical framework
In this theoretical framework marginalization will first be conceptualized, distinguishing the objective and subjective part of the concept. Subsequently, theory on group formation
is discussed to end up at the behaviour of the formed groups, which mostly concerns multiculturalism.
2.1 Marginalization
In this paper marginalization is used as the factor of analysis. Recently, Robert J. Pauly used this term in his influential book “Islam in Europe: integration or marginalization?” to indicate a cultural, economic, ethnic, linguistic, religious, political, and social backwards condition (2016). Marginalization has mostly been used in the education and school segregation literature, where it indicates the position of being on the sideline instead of the centre and therefore in an unequal position (Beach & Sernhede 2011; Messiou 2006). This term has been chosen since it depicts not only an objective state of poverty or social mobility, but also a subjective and a normative side. The normative side of unjust inequality is essential for the problematization of this paper. In this part of the theoretical section the term marginalization will be conceptualized in its objective and subjective meaning (e.g., Olson 2017; Rudolf 2017). One might be poor for example, but only if a person perceives themselves poor, it influences group formation processes. Therefore, objective marginalization will be defined by segregation, poverty and social mobility. Subjective marginalization is equal to perceived marginalization, which can be bottom lined as perceived discrimination.
First, segregation is the residential overrepresentation of a certain group in one area, while they are significantly underrepresented in others (Kempen & Şule Özüekren 1998: 1633-1634). These groups often have an ethnic distinguishing factor, which is one of the main explanatory variables of segregation (Musterd 2005: 335). This is influenced by dependency of in-group network support systems and the motivation to communally express cultural identity like language (idem: 340). Another main explanation for segregation is housing choice (Kempen & Şule Özüekren 1998: 1636). In Western Europe, people with an ethnic background often have less resources due to job market discrimination or unrecognized skills, so high concentrations cheaper housing in certain areas facilitates the preference of these groups to live close to each other (idem: 1641-1642; Musterd 2005: 339-340).
Segregation is heavily intertwined with poverty, which is the second part of objective marginalization. It is poverty that leads cheap housing neighbourhoods with high percentages immigrants into a downward spiral. Cheaper housing attracts lower socioeconomic status families, with children henceforth going to lesser quality schools in these neighbourhoods (Kempen & Şule Özüekren 1998: 1633-34; Beach & Sernhede 2011:
258-262). The need for resources and a negative image explain the shift towards ghetto formation.
Once this downward spiral has set in, the neighbourhood contained social resources and declining financial resources form a social mobility trap (Kempen & Şule Özüekren 1998: 1641-1642). Social mobility is the third part of objective marginalization. Social mobility is often described as an upwards movement by an individual in societies hierarchal ladder through talent, training and network (Lipset & Bendix 1991: 1-5; Musterd et al. 2003). Social mobility can also be applied to groups and creates a contrast between the ruling class and lower social group positions (Lipset & Bendix 1991: 4-5). One of the main explanatory variables of little social mobility is the neighbourhood someone grows up in (Musterd et al. 2003: 878). This, of course, makes sense looking at the downward spiral of poverty in segregated areas. For decades governments have tried fighting crime and poverty in segregated neighbourhoods, but the results are limited (idem: 879-818; Miltenburg et al. 2018). With these three parts of objective marginalization it has become clear that objective marginalization is not prone to fluctuate quickly over time.
Next to objective marginalization, subjective marginalization is an indispensable part of the equation. Even if people are segregated and poor with little social mobility, they can be quite content about the community they live in. This could be, because of the cultural or religious identity they are able to presume, or the level of trust business owners can hire people with from within the same ethnic group (Kempen & Şule Özüekren 1998: 1635). If objectively marginalized groups would like to move up socially but find out they are unable to, an anger against the ruling class stirs (Young 2006). This can best be described as perceived discrimination (Geddes & Scholte 2016: 40). Most research with immigrants about this process has been done around their well-being and conclude that high levels of perceived discrimination correlate with high levels of stress, illness and depression (Finch et al. 2000; Aichberger et al. 2015; Keys et al. 2015). In this paper, perceived discrimination is an essential addition to objective marginalization, since mostly, solely objective marginalization is taken into account during research into marginalized group positions (e.g., Arango& Baldwin-Edwards2014
;
Bonin 2017; Geddes & Scholte 2016; Hooijer & Picot 2015; Musterd & Ostendorf 2013;
Nielsen & Krasnik 2010; Nolan& Whelan2010;
Phillips 2010; Townsend2014). Group formation is a process with subjective marginalization is its core and it becomes clear in this paper later on, that subjective marginalization can fluctuate broadly over short periods of time.2.2 Group formation
The second section of the theoretical framework is about group formation. Group formation finds its roots in social identities of common ground (Brewer 2001). People are prone to look for identities. They can be found on an individual level, but are more likely to be found on a relational or collective level, where individuals identify with others or certain groups (Blumer 1958; Brewer 2001: 118-119; Hogg et al. 1995). Individuals often feel part of more than one identity: someone can identify as a Christian, immigrant and psychologist at the same time for instance (Joireman 2003: 2). This intersectionality is hard to account for. Overall, religion, race and language seem to be the major determinants of the most salient identities for a person (Fox 2004; Hamilton 2001; Joireman 2003: 3-6; Medeiros 2017). To be able to work on a group level analysis this paper omits debates about intersectionality and focusses on the salient identity of ethnicity, while taking into account the downsides of essentialism (Wong 1998). In this second section of the theoretical framework, group position theory is proposed as the main theory on group formation, although other influential group formation theories are also taken into account to show group position theory in its context and they are helpful additions for a nuanced story.
In 1958, Blumer proposed with group position theory that race prejudice should be explained at the analysis level of groups, rather than the individual level. Since the prominent role of ethnicity in the first section, it makes sense to look at intergroup dynamics through the lens of race. Group position theory is formed around the idea of two striving groups: a dominant group and a subordinate group (Blumer 1958; Bobo & Hutchings 1996: 952). The dominant group feels superior, entitled and is suspicious against upcoming new groups, because they are merely focussed on the needs and emotions of their own group (Blumer 1958: 4; Inkelas 2003: 628). The subordinate group is formed during the fight against their oppressor (Blumer 1958: 5-6; Rabb & Lipset 1962). The formation of a subordinate group is a complex social process, but comes down to the need of gathering strength to be able to stand up against the dominant group (Blumer 1958: 5-6). Since they encounter inequality, they feel the need to find common ground with people around them to stand stronger together. This process simplifies the characteristics of a group often to bare minimums like: blacks or Jews (ibid).
According to group position theory, racial prejudice is a defensive reaction from the dominant group towards a rising subordinate group that challenges the position of the dominant group (Bobo & Hutchings 1996: 957). The minority forms their sense of inferior social position as a collective process through their perception of that social
position. In these processes big events are crucial to the group formation process. Leaders of the groups opportunistically form in-group discourse when events happen with collective significance and therefore can escalate rapidly (Blumer 1958: 6). According to Blumer, these group aggregate processes have more impact on racial prejudice than the individual experiences (ibid). Simultaneously, both groups undergo this process and slowly superiority and inferiority embeds into the mindsets of the groups until they become “chronic and impermeable to change” (idem: 7). Group positions can, nonetheless, dissolve when group leaders do not take big events as an opportunity to emphasise the cleavage between the groups (ibid).
Decades later social identity theory also had a major impact on the analysis of group behaviour (Hogg & Abrams 1988; Tajfel & Turner 1979). Social identity theory emphasises intergroup relations within a framework where multifaceted individuals evaluate the relationship between individual behaviour and social structure (Hogg et al. 1995: 225). This social structure is characterized rather by the cleavages of in-groups and out-groups, than a hierarchal order of groups (idem: 262). Another difference to group position theory is that it is believed that the individual makes sense of their position themselves rather than following the in-group discourse. The addition of this theory to group position theory allows the interviews to be evaluated with greater dept, because the interpretations of the insiders about in-group attitudes needs to take into account that they themselves might give themselves a slight different position than the rest of the group.
In the nineties Blumer’s group position theory was largely confirmed with newer data analysis methods (Bobo & Hutchings 1996; Bobo 1999). Moreover, it was taken one step further by Pratto, Sidanius, Stallworth and Malle to a theory they called social dominance theory (1994). Social dominance theory minimizes group conflict within a society by institutions that promote ideologies to bring the dominant group in a superior position (Pratto et al. 1994: 741; Sidanius et al. 2004: 847). This mainly happens through myths that legitimize and validate their superior position (Pratto et al. 1994: 742). However, the research and its conclusions were mainly applied to individuals, telling a story of personal preference towards hierarchal relationship structures and discrimination, but omitting largely intergroup relations (idem: 754-755). Furthermore, the individual approach shows that individual members of minority groups will either try to become more like the dominant group or move antagonistically against the dominant group, which builds in-group favouritism (Sidanius et al. 2004: 864-865).
These three theories show a broad mix of schools of thought within social identity studies and form therefore a diverse overview. First, the primordialistic view of social identities can be recognized in the unpreventable conflict that Blumer as well as Pratto, Sidanius, Stallworth and Malle assume. Primordialism portraits identities as unchanging and implies that different groups cannot live together without conflict (Muro 2005: 5). Second, within Blumer’s group formation of the subordinate group also the instrumental school can be recognized where identities are fluid, and certain identities can become salient when they are instrumentalized by entrepreneurs (idem: 6). Third, social identity theory depicts the school of constructivism, that also assumes a dynamic identity, but in this school identities are a product of social processes that anyone can influence (ibid).
2.3 Behaviour
All of the theories on group formation show competitive intergroup relations. All of them assume a competition for limited resources, which creates an in-group bias. Is multiculturalism therefore something that is impossible to work towards? And, how can we expect different ethnic groups to live in harmony? In this third part of the theoretical framework will be looked at the behaviour of groups.
Multiculturalism is characterized by a total acceptance of all cultures and their identity values (Kymlicka 2010: 98-99). Since the year 2000 academic voices argue the failure of multiculturalism in Europe (e.g., Kymlicka 2010; Vertovec & Wessendorf 2010: 1-8). That makes sense on the basis of the group formation theories examined. The failure of multiculturalism can be differentiated further to show the difference between the dominant group and minorities in a society. Maykal Verkuyten studied perception differences around multiculturalism between dominant and subordinate groups (2005). He found that members of the dominant group don’t mind more multiculturalism if they are higher educated, which probably means they are less threatened by ethnic immigrants taking over their jobs for example (Verkuyten 2005: 126, 135). Although, members of the dominant group who are lower educated and might feel threatened by ethnic immigrants predominantly expressed negative attitudes towards multiculturalism. These are both effectively in line with the theory proposed so far. Verkuyten’s study also shows that subordinate groups prefer more multiculturalism, since that would improve their social position. The word ethnic is used in this paper to indicate non-western immigrants, since the term non-western speaks out of the perspective from western values and therefore normalizes western norms, whereas these norms are in no shape or form better or worse
than other philosophical frameworks (Yanow & Van der Haar 2010; Yanow & Van der Haar 2013).
This creates a need for parties that fight for the position of ethnic immigrants, since the lives of marginalized people seem largely influenced by their marginalization, with ethnicity as one of its core explanatory factors. In African studies this has been a subject of research for political analysis already at length (e.g., Bratton et al. 2012; Elischer 2013; Eifert et al. 2010). This might result in ethnic parties that chiefly fight for the rights of one particular ethnic group and forms a theoretic basis for the success of BIJ1 and DENK.
With the conceptualization of objective marginalization, subjective marginalization and group formation the analysis can be commenced. Before the hypothesis is proposed to current state of Amsterdam in all of this will be clarified.
3. The case of Amsterdam
This section first goes over the reasons for selecting the case of Amsterdam. Afterwards, the context of the Surinamese/Antillean and the Turkish community in Amsterdam is given in light of their objective and subjective marginalization. Eventually, the context for this study is further elaborated upon by an examination of their political behaviour since the arrival of these groups in the Netherlands.
3.1 Selection
Amsterdam was selected as the case for this paper for several reasons. First of all, the Netherlands is selected over other West-European states. The proportional, electoral system without any electoral thresholds – other than that of at least one seat – enables smaller political parties to get seats in parliaments or councils (Gallagher & Mitchell 2005: 503; Vossen 2017). Also, there was a shift where ethnic immigrants voted for newer parties that aim at specifically those ethnic groups (NRC 2018). This creates the opportunity to see whether this political mobilisation took place with the motivation of injustice about their marginalized position. Second, a city is selected over the Netherlands as a whole. The higher proportions immigrant groups in the city create clearer analysis (CBS 2001). Third, Amsterdam is selected over other big cities like Rotterdam and Den Haag. In Amsterdam it was not only political parties supported by people groups from predominantly Muslim countries, like DENK and Nida, that got voted into the city
council. Also, political party BIJ1 won a seat in the city council of Amsterdam with very little Muslim voters (de Volksrant 2018; Kranendonk et al. 2018). This variety will enable the results of this study to be extrapolated in a wider sense, since unity in religion might influence their behaviour.
Within Amsterdam there are three big, older immigrant communities: the Moroccan, Surinamese/Antillean and Turkish community. The Surinamese/Antillean and Turkish group are selected for the analysis of this paper, because they show the least similarities in background, culture and voting behaviour, which creates the richest analysis (Geddes & Scholten 2016; Vermeulen & Michon 2018). People from Surinamese and from the Antillean are treated as one group in this paper. First of all, these cultures are alike, the countries have similar histories and they came to the Netherlands for the same reasons (Van Gom 2018, Heelsum 2002; Lucassen & Lucassen 2011). Moreover, this is done by many other researchers (e.g., Gemeente Amsterdam OIS 2015; Tillie 2004; Vasta 2007; Vermeulen & Michon 2018).
3.2 The Surinamese/Antillean and the Turkish community
Let’s take a closer look at the Surinamese/Antillean and the Turkish community. Of the 800,000 Amsterdam residents, the Surinamese/Antillean group counts 80,000: 68,000 Surinamese and 12,000 Antilleans (Gemeente Amsterdam OIS). Most Surinam immigrants came to the Netherlands in the seventies after Surinam became an independent country and was no longer a colony of the Netherlands (Heelsum2002). Most Antilleans came to the Netherlands in the eighties and around the year 2000 (Lucassen & Lucassen 2011). The majority of both of these groups were lower educated labour immigrants looking for work. Moreover, 5% of Amsterdam residents are from Turkish descent (Gemeente Amsterdam OIS). Most Turks came to the Netherlands in the sixties and seventies as labour immigrants and were only going to stay for a couple of years (Prins 1996: 7). However, most of them ended up staying and brought over their families also.
These two groups live with a high segregation rate on the outskirts of the city of Amsterdam (Musterd & Van Gent 2015). Nevertheless, although many different policies were designed to take these groups out of their isolation, the rate of segregation is still high (Musterd & Van Gent 2015: 75; Bolt et al. 2008: 1379). Segregation is most salient in primary education. The quality of education at schools with a high proportion children from ethnic migrant families are significantly lower than schools with little ethnic migrants’ children (Gemeente Amsterdam OIS 2015a). People with a Surinamese/Antillean and Turkish background are more than twice as often classified
‘poor’ by the Amsterdam government than Dutch natives (Gemeente Amsterdam OIS 2015b: 31-33). Also, the Netherlands performs poorly in comparison to other European countries with regards to discrimination in the labour market (Dancygier & Laitin 2014: 45). People from poor neighbourhoods find themselves climbing the social ladder of hierarchy less high than from rich neighbourhoods (Musterd et al. 2003).
From an objective conception of marginalization, these rapports show that ethnic immigrant groups in Amsterdam have a weak position. But, what can be said about subjective marginalization? The Surinamese/Antillean group did their own research recently about discrimination and concluded that discrimination against their group is still prevailing (Esajas 2017). This research shows several signs of prepositions, but this can also be seen as a measure of subjective marginalization. The Turkish group likewise experiences moderate levels of discrimination according to questionnaires by SCP and IPSOS (Grimbergen 2010; SCP 2017: 21-24; Etnobarometer 2017). All of these sources show a clear us/them dichotomy, where a weaker out-group connections correlate with high discrimination perceptions.
Does this mean that multiculturalism is failing in the Netherlands? Many scholars argue it is (e.g., Entzinger 2006; Entzinger 2014; Sniderman & Hagendoorn 2007; Vasta 2007; Verkuyten & Thijs 2002). In the eighties, the Netherlands implemented policy that promoted multiculturalism. But, as more immigrants arrived, the integration policy shifted in the late nineties to assimilation instead, to protect the Dutch identity, mostly against Islam. Furthermore, as it became clear that immigrants ended up in significantly lower social positions, immigrants needed to assimilate to decrease the inequality (Vertovec & Wessendorf 2010: 3; Vasta 2007; 722).
3.3 Political behaviour
This section has painted thus far a negative picture of the older ethnic immigrant groups in Amsterdam, but all this academic literature depicts the story from one perspective only. So far, we looked at comparative research of averages. This needs to be complemented through the lens of group formation and the behaviour of the formed groups. In this paper this is done by looking at the voting behaviour of the Surinamese/Antillean and Turkish groups.
The Surinamese/Antillean and Turkish communities have predominantly been voting for Partij van den Arbeid (PvdA) since they arrived, which is the established Dutch Labour Party (Rath 1988: 635; Vermeulen & Michon 2018). This is mostly because of class, but for the Turkish community also because some Turks moved up in the ranks of the
party (Rath 1988: 636). But later on, even while the PvdA was working together in coalitions with liberal proponents of assimilation policies like the VVD, they kept on voting for the PvdA. At the city council elections in 2010, 54% of the Surinamese/Antillean and 58% of the Turks voted for PvdA in Amsterdam (Van Baalen 2018; Vermeulen & Michon 2018). If these groups perceive themselves as marginalized they would be voting for the party they believe will improve their position most, according to the literature.
So, it makes sense that in 2014 this started shifting, mostly amongst the Surinamese/Antillean group towards other parties. In 2018 there was a major shift in the Turkish community. This could be because of the downfall of the PvdA, but interestingly it seems that most of these voters started voting for new, activist-prone, progressive parties with leadership of the same ethnicity as these people they are attracting (Boom 2017; Kranendonk et al. 2018; Vermeulen & Michon 2018). Amongst the Surinamese/Antillean group Silvana Simons from BIJ1 won 18% of votes in contract to 74% of the Turkish voters who voted for DENK (Vermeulen & Michon 2018). Both of these parties verbalize the same message: marginalized groups deserve a better position. Both of these parties do not mention they solely represent those particular ethnic groups, but they do predominantly attract people with particular ethnic backgrounds. Therefore, in this paper, they will be categorized as ethnic activist parties. For the Turkish group it seems that DENK attracted voters to the ballot box who did not vote in 2014, since the turnout rate for the Turkish community was ten percentage points higher in 2018. It also needs to be taken into account that the bigger shift in the Turkish community might be explained by the organized nature of the Turkish community in Europe (e.g., Doomernik 1995; Østergaard‐Nielsen 2001; Yukleyen 2009).
The political party BIJ1 started when Sylvana Simons left the political party DENK in 2016 after a disagreement (NRC 2016a). At the time the party was called Artikel 1 and that year in the national elections they didn’t make it into parliament. In the council elections of 2018 BIJ1 drew enough votes for one seat in Amsterdam with a programme of equality and justice (Het Parool 2018). Furthermore, DENK started in 2014 when Tunahun Kuzu and Selçuk Öztürk stepped out of the PvdA to start their own party (Vossen 2017: 252). DENK stands for a diverse multicultural society and defend the rights of the marginalized, mostly aiming at ethnic migrants. DENK is known for the relationship they maintain with Erdogan’s AK party in Turkey. At the moment DENK has 3 seats in the national parliament and 21 council seats nationally, most of them in the bigger cities (Nu.nl 2018).
4. Hypothesis
On the basis of the theory and the current state of Amsterdam in its context of the past, what can we expect to find about marginalization? This paper mostly sets out to nuance a story that has already widely been told, yet there will be hypothesis tested. Foremost, this paper analyses to what extent marginalization is explanatory for the shift from PvdA to BIJ1 and DENK, since this will show the underlying reasons for current group positions. The following main hypotheses are suggested:
H1: Marginalization is a major explanatory variable for the Surinamese/Antillean group to vote for BIJ1.
H2: Marginalization is a major explanatory variable for the Turkish group to vote for DENK.
Besides the main hypotheses, two minor hypotheses will be reflected upon in the conclusion. If marginalization is a determining factor in H1 and H2, the following is also expected:
H3: The Surinam/Antillean community is less marginalized than the Turkish community.
This is expected, since there was a bigger shift in the Turkish group to ethnic activist parties. Furthermore, will be analysed if we can make a claim about mobilisation:
H4: Marginalization is a mobilizing factor.
5. Quantitative analysis
To test the hypothesis, a mixed method research approach is chosen to add in-group knowledge to the quantitative data. In this first section the data, operationalization, research design and results are presented for the quantitative part of this paper, whereupon the same will be done for the qualitative part.
5.1 Data
For the quantitative section, the exit poll data of the Amsterdam city council elections of 2018 is used (Vermeulen & Michon 2018). The date is produced by a long existing joint venture between the University of Amsterdam and Gemeente Amsterdam OIS. At different locations throughout Amsterdam 32 researches took questionnaires at voting stations, which resulted in 3300 respondents. Since 1994 this same method has been used to research voting behaviour and compare different council elections (ibid).
The robustness of the data is mainly determined be long recognized tradition of quality exit polls since 1994 of the Amsterdam city council elections (e.g., Van Heelsum 2005; De Vreese & Semetko 2002; Zeedan 2018). Notable is the high percentage voters of Turkish descent in the sample. In Amsterdam they make up 5% of the population, while in this sample they make up 10% of respondents. This is not a problem, since the other groups have representative percentages in the sample (Gemeente Amsterdam OIS). Moreover, on other factors the sample seem to be representing the population quite well (SCB 2016a). There is an education bias in the sample: voters were higher educated than city averages per ethnic group, which is also true for the population (CBS 2017: 6; Hillygus 2005). With the Surinamese/Antillean and Turkish group this was within 8 percentage points difference per education category (SCP 2016: 62). Though, native residents showed a substantially bigger education bias. Higher educated Dutch make up 49% of the Amsterdam population, but 69% of the Dutch respondents can be categorized in that highest bracket, which is a difference of 20 percentage points (Gemeente Amsterdam OIS 2012: 3). This points towards a mobilization of the marginalized.
5.2 Operationalization
This dataset is utilized to measure objective marginalization, subjective marginalization and behaviour for both the Surinamese/Antillean and the Turkish group. First, objective marginalization will be measured by the level of education, which is one of the questions in the exit poll questionnaire. The education level proxy can be used because of the correlation between education level, segregation, socioeconomic status and social mobility. A high correlation can be found between the level of education and segregation (e.g., Alba & Logan 1993; Coleman 1966; Darling-Hammond 1995; Johnson 2010; Williams & Collins 2001). It is generally believed that segregation cleavages follow to a big extent education level cleavages. The same is true for the relationship between socioeconomic status and the level of education (e.g., Brayne & Calloway 1990; Denton &
Massey 1988; Samir & Lutz 2017; Winkleby et al. 1990). And finally, the level of education, segregation and poverty are fine predictors of social mobility (e.g., Deary et a. 2005; Lucas 2001; Xie & Goyette 2003).
Second, in the quantitative section subjective marginalization will be measured by group attitudes. The questionnaire contained the question “Please indicate on a scale from 1 (no connection) to 7 (strong connection) how connected you feel to the following groups” and then proposed: Dutch, Amsterdam residents and your own ethnic group. This generated three separate attitude measurements per respondent. The extent that ethnic immigrants feel connected to Dutch people might not only an indicator for segregation, but also an indicator for discrimination (Levin et al. 2003). When they feel more discriminated against by the dominant group, they will indicate as less connected to the Dutch group (Dion 2014). Here we would expect a negative relation between feeling connected to Dutch natives and Amsterdam residents, and voting for their ethnic party – as they are considered as out-groups, and a positive relation with the in-group part of the question. However, about 53% of both the Surinamese/Antillean and the Turkish community who voted at the council elections indicate a 6 or 7 and feel strongly connected to Dutch people. There is a high chance that people who feel less connected to Dutch people didn’t vote, because they care less about the Netherlands and their leadership. When we look at the level of connectedness to their own ethnic group we recognize higher percentages in the highest two brackets, respectively 65% of the Surinamese/Antillean and 63% of the Turkish community. This depicts a higher loyalty to their in-group and a moderate level of in- out-group dichotomy, but the small difference with the connectedness to Dutch natives predicts that it will be hard to see clear signs of subjective marginalization in the logistic models.
Third, behaviour and mobilization will be measured by voting behaviour during the council elections. This cannot be seen as analysing the mobilisation or the behaviour of the Surinamese/Antillean and Turkish community. But, what it does allow for, is to see the role of marginalization when they move – as they did from the PvdA to BIJ1 and DENK. The fact that voting behaviour can be seen as a decent indicator for group attitudes adds to the validity of this study, but is not leveraged in this paper (e.g., Bloemraad & Schönwälder 2013; Bird et al. 2010; Geys 2006; Martiniello 2005; Maxwell 2010; Tillie 2004).
5.3 Research design
The quantitative part of this research investigates whether marginalization is salient as an explanatory factor over ethnicity. The following variables are used. The dependant variable is “party choice” and has three categories: 1) other parties: being all parties other than BIJ1 and DENK – this also is the reference category; 2) BIJ1; and 3) DENK. The first independent variable is “ethnicity”. Respondents are identified as either Dutch, Surinamese/Antillean, Turkish or “other”, where Dutch is the reference category1. The
second independent variable is highest completed education and consist out of four categories2, as it is usually done in the Netherlands (SCP 2016). For the education
variable the lowest category is the reference category. The third independent variable is a continuous variable and controls for “age”. The fourth independent variable is “gender” and divided up in: 1) male – which is also the reference category; and 2) female. The fifth, sixth and seventh independent variables measure subjective marginalization. They indicate the strength of the connection of respondents with Dutch natives, Amsterdam residents and their ethnic in-group. For these three independent variables the lowest score 1 means no connection and the highest score 7 means strongly connected. The lowest score 1 is the reference category.
It would have been the correct method to use a multinomial logistic regression to measure the explanatory power for each of these independent variables, since the research design satisfies all assumptions and conditions for this method3 (Starkweather
1 For the dependant variable “party choice” the first category is Dutch and includes all respondents with both parents born in the Netherlands – this is also the reference category. The second category is the Surinamese/Antillean group and consists out of all respondents where either their father, mother or they themselves are born in Surinam or on one of the Antillean islands. The third category is the Turkish group and consists out of all respondents where their father, mother or they themselves were born in Turkey. The fourth category contains all other respondents and consists mostly out of people with Moroccan descent. This category also includes 8 respondents who are in two of the first three categories. Furthermore, the categories for ethnicity can be made up like this, because more than 90% of third generation immigrants are still younger than eighteen years old and therefore are not in this sample, since in the Netherlands the minimum voting age is eighteen years old (CBS 2016a).
2 The four categories used for education are: 1) Lower education: being primary education – which is also the reference category; 2) Medium lower education: lower secondary education (vmbo); 3) Medium higher education: higher secondary education (havo/vwo) and lower tertiary education (mbo); 4) Higher education: higher tertiary education (hbo/wo) (SCP 2016).
3 The assumptions and conditions for a multinomial regression are satisfied. First, the dependant variable “party choice” is a nominal variable with more than 2 categories, the main independent variables are nominal variables with more than 2 categories. Second, the observations are independent of each other. Third, the dependant variable has mutually exclusive categories and is exhaustive. Fourth, there is no multicollinearity. All independent variables have a maximum Pearson’s or nonparametric correlation of 0,3. Only independent variables 5, 6 and 7 – the connectedness to Dutch natives, Amsterdam residents and members of own ethnic group – have correlations of maximum 0.6 amongst each other. Finally, all variables are categorial, which means that there are no outliers, high leverage values or highly influential points.
& Moske 2011: 1). However, the multinomial logistic models didn’t converge, because some categories in the crosstab “party choice” and “ethnicity” where to small4 (See
appendix 2). Therefore, a logistic regression was applied with a marginal regressions analysis to create a clear picture of the influence of each independent variable. This paper follows in that sense the example of Medeiros & Noël, who did the same in their study about party identification in Anglo-American democracies (2014).
5.4 Results
These logistic regressions yield the margins models graphically presented in figures 1.1, 1.2 and 1.3 including all variables. For the exact numbers, please see the results tab Appendix 1. Overall can be seen that ethnicity plays the most important role and has substantially more explanatory power than marginalization, which contradicts H1 and H2. Age and gender hardly play a role, which contributes to the robustness of the data.
Figure 1.1 shows to what extent the independent variables are explanatory in comparison to the others in predicting whether someone would vote for BIJ1 instead of all other parties, including DENK. Being Surinamese/Antillean strongly determines voting on BIJ1 instead of another party. There is no line for the Turkish group in this graph, because there were no Turkish respondents in the sample who voted for BIJ1. Marginalization seems to play a small role. First of all, education is not significant. Also,
aaaaaaa
Figure 1.1: Margins models of logistic regression voting for BIJ1
4 In the sample only 9 Dutch respondents voted for DENK, 9 ethnic “other” respondents voted for BIJ1 and no one of the Turkish community voted for BIJ1.
Figure 1.2: Margins models of logistic regression voting for DENK
Figure 1.3: Margins models of logistic regression voting for other parties
feeling more connected to Dutch natives makes people less likely to vote BIJ1, as expected. However, in-group attitudes are not significant. Overall, this shows that marginalization is not explanatory with respect to ethnicity for voting BIJ1 and does not confirm H1. If there were more Surinamese/Antillean respondents some of the indicators might have been significant.
For figure 1.2 the same model is run for voting for DENK instead of all other parties, including BIJ1. Again, ethnicity is by far the most salient explanatory variable. The explanation power of being Turkish in voting for DENK in model 2 is twice as strong as for being Surinamese/Antillean and voting for BIJ1 in model 1. Marginalization plays a bigger role in model 2 than in model 1. In model 2 marginalization indicators are all
significant but very small with a negative effect of “education”, a negative effect of feeling connected with Dutch natives and a positive effect for the connection with other people from the Turkish community. Lower educated Turks are more likely to vote for DENK than the higher educated. Feeling strongly connected to Dutch natives does reduce the likelihood of voting for DENK, and feeling connected to other Turks does increase the likelihood. These are all in line with the theory, but the little explanatory power contradicts H2.
In figure 1.3 it can be seen that both Surinamese/Antillean and Turkish people are less likely to vote for parties other than BIJ1 and DENK compared to Dutch natives. Turkish more so than Surinamese/Antillean. This can also be seen in figure 1.2, since people from the Surinamese/Antillean community are more likely to vote for DENK than for other parties, including BIJ1. This does not mean that Surinamese/Antillean people are more likely to vote for DENK then for BIJ1, since the effect for voting BIJ1 in model 1 is bigger and BIJ1 only is a small part of all of the other parties.
To have a closer look at education as an explanatory variable, margins plot models were run as graphically presented in figure 2.1 and 2.2. Instead of the commonly used 95% Cls, an 84% Cls is used to approach a 0.05 significance level difference between the
aaaaaaa
Figure 2.2: Margins plot of the predictions for education per ethnic group in voting for DENK
two categories (MacGregor-Fors & Payton 2013). This change created the possible curvilinear relationship in figure 2.2 to be significant for the category higher educated. Figure 2.1 shows that there is a good chance that the lowest educated are more likely to vote for BIJ1, but since it is significantly clear that it is not a curvilinear relationship, we cannot draw any hard conclusions out of this. That is what one might expect, since there are only 59 Surinamese/Antillean respondents who voted for BIJ1. However, amongst the 240 Turkish respondents who voted for DENK there are is a higher chance of a curvilinear relation and this is significant for the higher educated bracket. This shows that higher educated Turks are less likely to vote for DENK and does point towards a role for marginalization in explaining the voting shift. These margins plot models show that there is some explanatory power, but this is not enough to confirm H1 and H2.
The quantitative results show that marginalization has a very modest role to play in comparison with ethnicity. Objective marginalization does confirm what was expected on basis of the literature whereas subjective marginalization is very hard to differentiate.
6. Qualitative analysis
To validate the quantitative data, and create decent insights into the group formation process and behaviour of the Surinamese/Antillean and Turkish communities, the qualitative part is essential for this study. The qualitative part exists out of two interviews with insiders from both communities. They will first be introduced to validate their knowledge about these communities. After a short notice on interview strategy, reports are drawn first from the interview with Norman van Gom and Max Sordam about the Surinamese/Antillean community, and second, from the interview with Numan Yilmaz about the Turkish community. This section closes with a summary of its results.
6.1 Sources
From the Surinamese/Antillean community was spoken with Norman van Gom and Max Sordam, who were amongst the early arrivers from Surinam in the seventies as first-generation migrants. Norman van Gom is the director of a Suriname community centre in the Amsterdam-Zuidoost and director of Surinamese radio station MART. Max Sordam is a linguistic academic and cultural historian. He has received decorations of the Dutch Order of Orange-Nassau and the Surinamese Order of the Yellow Star for his work in building bridges between the Surinamese and Dutch cultures. He also wrote a book about the Suriname culture in the Netherlands: “Imprisoned in between cultures: backgrounds, attitudes and cultural behaviour of the Surinamese”.
Additionally, Numan Yilmaz was interviewed for his insights about the Turkish community. His parents moved from Turkey to the Netherlands in the early eighties before he was born, which makes him a second-generation migrant. Yilmaz was born in Amsterdam, where he grew up and studied. He occupies one of the 45 seats in the Amsterdam city council on behalf of the political party DENK, which occupies three of those seats. Yilmaz got himself into the council by breaking the preferential voting threshold (Het Parool 2018). He is the only one in the Amsterdam city council who has done so. This shows that he is in touch with the Turkish community in Amsterdam, as people voted for him instead of politicians higher on the list.
6.2 Interview strategy
To influence the respondents as less as possible, the following interview strategy has been put into place. Before the interview the concept of marginalization was explained in only three sentences to avoid nudging towards over-emphasizing the proposedly bad condition
of their in-group. To avoid the downsides of essentialism it was not assumed that the groups, as formulated in this study, are seen as defined groups by insiders. Therefore, in the interviews first questions were raised about how the groups are made up.
6.3 Van Gom and Sordam about the Surinamese/Antillean community
Frist, the interview with Norman van Gom and Max Sordam, who are both Surinamese. They stated that the Surinamese and the Antillean people can be seen as one cultural group in the Netherlands although this group is not united (Van Gom 2018). Surinam itself is not a homogenous country to start with. It consists out of several people groups like the Indo-Surinamese (or Hindustani) and Javanese. Moreover, they do not have one overarching religion. This disunity probably explains the dispersion of voting behaviour of this group. What they do share is the oppression of the Dutch culture in their countries during the colonial time to live like the Dutch, including social norms and language (Sordam 2018). The Surinamese and Antilleans do not feel Dutch. They want to partake in the Dutch society, but do not want to become Dutch.
Van Gom is convinced that at least 80% of Surinamese/Antillean people feel that the discrimination from Dutch natives has brought the community where it is right now: marginalized. Both Van Gom and Sordam are convinced that the Surinamese/Antillean community is socioeconomically worse off on average than Dutch natives. They perceive social mobility is achievable only to a certain extent, since it is capped by discrimination. Most Surinamese believe this will never diminish. For the social mobility that does take place “a Surinamese person needs to work three times as hard as a white person”, mostly because of stigmatization at schools by white teachers (Van Gom 2018). Segregation was not mentioned by Van Gom or Sordam. It seemed that the Surinamese/Antillean community is very happy to live together and act the way they want to act amongst each other, without having to adopt Dutch social norms.
In the interview there was no mention of any big events that group position theory describes as crucial to the formation of a subordinate group (Blumer 1958: 5-6), which might also explain the disunity of this group. Sordam admits that there is no political organisation that has reached or mobilized the Surinamese/Antillean people as a whole, because of their disunity and lack of leadership (Sordam 2018). The fact Van Gom and Sordam clearly speak about the failure of multiculturalism from the perspective of the Surinamese/Antillean group contradicts multiculturalism literature, since the Surinamese/Antillean community does not prefer to vote for ethnic parties according to
the data (Verkuyten 2005). Their “marginalization is not bad enough for them to unite, which is mostly because of the Dutch welfare system” (Van Gom 2018).
6.4 Yilmaz about the Turkish community
The second interview with Numan Yilmaz was fundamentally different. The bottom line of this interview was that the marginalization of the Turkish community is a purely political process and has very little to do with poverty, segregation or education (2018). It substantially shows subordinate group formation through big events. According to Yilmaz, the Turkish community can be viewed as a homogenous group: “Turkish people are loyal to their Turkish roots with its values, traditions and language” (Yilmaz 2018). This loyalty creates homogeneity, where lower educated Turks are closer with higher educated Turks as opposed to lower educated people from other ethnic groups. The strong connection with their roots does not withhold them from considering themselves Dutch. However, their unity should be nuanced, since one cleavage can be found, characterized by politically conservative and politically progressive. The conservatives make up about 70% of the Turkish community in Amsterdam and are pro-Erdogan. The progressive side make up about 30% and are against the Erdogan government. Social foundations follow this cleavage, but this cleavage is non-existent in neighbourhoods and sport clubs. DENK voters are predominantly from the conservative group, but not solely. Even though the existence of this cleavage “the Turkish community can really be seen as one”, according to Yilmaz (2018).
Yilmaz explains that the Turkish community as a whole is well integrated and hardly marginalized, until a couple of years ago (2018). Segregation and socioeconomic status were not a problem, and social mobility was viewed positively and justly. This changed through a couple of political incidents. To start with, PvdA members of parliament Rob Oudkerk and Henk Spekman both separately, made belittling statements about Moroccans (TROUW 2002; Vrij Nederland 2008). Although these two incidents were aimed at Moroccans, the Turkish community felt very much rejected also. Another incident was when PvdA Minister Asscher published an erroneous study that stated that 80% of Turkish youth sympathizes with IS (RTL Nieuws 2014). These incidents made the Turkish community reconsider voting for PvdA. The tipping point occurred when the Dutch prime minister Mark Rutte told Turkish people to go back to their own country if they could not appreciate living in the Netherlands (NRC 2016b). He did so after a group Turkish Dutch people aggressively pleaded against the deportation of a pro-Erdogan Turkish minister who was in Rotterdam to speak at a pro-Erdogan conference just before
the Turkish elections (NRC 2016b). This caused Turkish people to lose their trust in their freedom in the Netherlands and caused a quick escalation of the subjective marginalization. Their subordinate group was already united, so these big events with collective significance were not absorbed by first forming unity. They impacted the direction of this group straightaway. For Turkish, people the descent of the PvdA cannot be separated from these events.
It cannot be claimed on the basis of Yilmaz’ words that the Turkish community views multiculturalism as a failure, but if the current attitudes maintain it moves that way. Yilmaz is conversely convinced that voting for DENK translates in voting for integration and not for isolation, because it shows that “we can amplify our voice through decent democratic engagement and partake in society that way” (Yilmaz 2018).
Yilmaz’ reasoning about the movement from PvdA to DENK within the Turkish community can align with the data through the following addition. It were predominantly the lower educated Turkish people who voted DENK. A high percentage of Turkish people voting for DENK is also is pro-Erdogan (Etnobarometer 2017). The cleavage of conservative and progressive within the Turkish community therefore seems to correlate with education level. That would mean that the significant marginalization indicators found in the quantitative analysis for the Turkish community cannot be interpreted with marginalization as theoretized in this paper.
6.5 Interview results
The interviews predominantly confirm the data and create useful insights into the mechanisms behind the data. Both groups struggle with their own identity and with the question to what extend they are Dutch. The Surinamese/Antillean community perceives their marginalization in the way expected, although they seem to be more content with the current situation than expected. In combination with their disunity this can explain their current behaviour. This might confirm H1, because the lack of marginalization correlates with the little mobilisation for BIJ1. The Turkish community by way of contrast, is more unified. One clear cleavage can be distinguished on the basis of political preference and education. Their voting behaviour can be explained by marginalization, but not as presented in the theory. In both interviews any signs of the respondents dispositioning themselves outside of their community could not be recognized. With social identity theory in mind, this contributes to the validity of the interviews.
7. Conclusion and discussion
To conclude this paper, marginalization is not a salient explanation for changing voting behaviours within the Surinamese-Antillean and Turkish community from PVDA to BIJ1 and DENK. This answers the main research question: what is the current role of marginalization in relation to political mobilization within the Surinamese/Antillean and the Turkish community in Amsterdam? Through the lens of group formation and group behaviour H1 and H2 cannot be confirmed.
This is first because the quantitative data for the Surinamese/Antillean community is not significant for objective marginalization and not significant or hardly explanatory for subjective marginalization. The interview with Van Gom and Sordam confirms this. They emphasized marginalization and disunity in the Surinamese/Antillean community. There are signs that objective marginalization plays a part in ethnic voting, but not significantly to make any claims. Second, for the Turkish community, the quantitative data shows that marginalization plays a more important role for them than for the Surinamese/Antillean community, but is negligible in comparison to the ethnic factors. According to an analysis of Yilmaz’ insights this argument can again be downgraded, since the quantitative indicators for objective and subjective marginalization are likely to show the political cleavage in the Turkish community, rather than the role of marginalization as theoretized in this paper.
H3 and H4 cannot be confirmed either. First H3, the Turkish community did face a higher rate of marginalization than the Surinamese/Antillean as a probable causation of their bigger move to DENK. However, the Turkish community mobilized because of the escalation of subjective marginalization through events with collective significance that were instrumentalized, according to Yilmaz (2018). This heavily leans on the validity of the interview with Yilmaz, because it cannot be concluded from the subjective marginalization indicators in the data. Which might mean that this interview shows the lack of validity of the operationalization of subjective marginalization. Other arguments that strengthen the contradiction of the hypothesis, is that the unity of the Turkish community through their organized nature and their religion made the community move as a whole. Finally H4, it cannot be claimed on the basis of this research that marginalization is a mobilizing factor, since this study has not found high marginalization rates. The turnout rates of the Surinamese/Antillean and the Turkish community amongst the lower educated is substantially higher than that of Dutch lower educated, which points towards a mobilisation of the marginalized. However, no curvilinear
relationship was found for the Surinamese/Antillean community in margins plot model 2.1. Also, the ambiguity of the Turkish high subjective marginalization contradicts this hypothesis.
Although the answers on all hypothesis and assumptions are unsure, this paper still has the potential to nuance the current group position analysis of immigrant minorities in Western Europe. According to the comparative approach literature clear signs of objective marginalization can be distinguished and should not be ignored. However, these two communities in Amsterdam do not seem to bother all that much about those facts. Van Gom mentioned in the interview that the Dutch welfare state benefits generate a certain level of contentness that creates no need for collective action within the Surinamese/Antillean community (2018). Has the Dutch government created an equilibrium where ethnic minorities are marginalized, but looked after just well enough for them not to unite and mobilize on the basis of objective marginalization? Without objective marginalization backing up the subjective marginalization that caused the Turkish group to vote for DENK, the party will need a continuity of big events with collective significance to keep the subjective marginalization bubble inflated and the Turkish community mobilized. Welfare state benefits do not take people out of a marginalized position and policies that set out to help ethnic immigrants do not seem to create equal positions in society for these groups either (Entzinger 2014: 61-82; Hooijer & Picot 2015; Joppke 2007). Does that mean we have ended up in the society of social dominance theory (Pratto et al. 1994)? In that case Western European countries might have created a limbo for ethnic minorities: the policies of the West cannot break the cycle, and while goodwill and discriminating forces balance out, the position of ethnic minorities will not change anytime soon.
Acknowledgements
I want to thank assistant professor Mike Medeiros (department of Political Science at University of Amsterdam) for all the help, especially for his suggestions concerning the quantitative part of this paper.
Interviews
Gom, van, N. (2018). Interview conducted on the 23rd of May 2018.
Sordam, M. (2018). Interview conducted on the 23rd of May 2018.
Yilmaz, N. (2018). Interview conducted on the 29th of May 2018.
Literature
Aichberger, M. C., Bromand, Z., Rapp, M. A., Yesil, R., Montesinos, A. H., Temur-Erman, S., & Schouler-Ocak, M. (2015). Perceived ethnic discrimination, acculturation, and psychological distress in women of Turkish origin in Germany. Social psychiatry
and psychiatric epidemiology, 50(11), 1691-1700.
Alba, R. D., & Logan, J. R. (1993). Minority proximity to whites in suburbs: An individual-level analysis of segregation. American journal of sociology, 98(6), 1388-1427.
Arango, J., & Baldwin-Edwards, M. (2014). Immigrants and the informal
economy in Southern Europe. London: Routledge.
Baalen, van, C., Wielenga, F., & Wilp, M. (Eds.). (2018). Een versplinterd
landschap: Bijdragen over geschiedenis en actualiteit van Nederlandse politieke partijen.
Amsterdam: Amsterdam University Press.
Beach, D., & Sernhede, O. (2011). From learning to labour to learning for marginality: school segregation and marginalization in Swedish suburbs. British
Journal of Sociology of Education, 32(2), 257-274.
Bird, K., Saalfeld, T., & Wüst, A. M. (Eds.). (2010). The political representation of
immigrants and minorities: Voters, parties and parliaments in liberal democracies.
London: Routledge.
Blumer, H. (1958). Race prejudice as a sense of group position. Pacific
Sociological Review, 1(1), 3-7.
Bobo, L. D. (1999). Prejudice as group position: Microfoundations of a sociological approach to racism and race relations. Journal of Social Issues, 55(3), 445-472.
Bobo, L., & Hutchings, V. L. (1996). Perceptions of racial group competition: Extending Blumer's theory of group position to a multiracial social context. American
Sociological Review, 61(6), 951-972.
Bloemraad, I., & Schönwälder, K. (2013). Immigrant and ethnic minority representation in Europe: Conceptual challenges and theoretical approaches. West
European Politics, 36(3), 564-579.
Bonin, H. (2017). The Potential Economic Benefits of Education of Migrants in the
Bolt, G., Van Kempen, R., & Van Ham, M. (2008). Minority ethnic groups in the Dutch housing market: Spatial segregation, relocation dynamics and housing policy.
Urban studies, 45(7), 1359-1384.
Boom, W. (2017). De neergang van de PvdA. Amsterdam: Nieuw Amsterdam. Bratton, M., Bhavnani, R., & Chen, T. H. (2012). Voting intentions in Africa: ethnic, economic or partisan?. Commonwealth & Comparative Politics, 50(1), 27-52.
Brayne, C., & Calloway, P. (1990). The association of education and
socioeconomic status with the Mini Mental State Examination and the clinical diagnosis of dementia in elderly people. Age and Ageing, 19(2), 91-96.
Brewer, M. B. (2001). The Many Faces of Social Identity. Political Psychology, 22(1), 115-125.
CBS (2001). Waar wonen niet-westerse allochtonen in Nederland?.
https://www.cbs.nl/nl-nl/nieuws/2001/48/waar-wonen-niet-westerse-allochtonen-in-nederland-
Last consulted 7 June 2018.
CBS (2016a). Jaarrapport Integratie 2016. Den Haag.
CBS (2016b). Wie zijn de derde generatie?. https://www.cbs.nl/nl-nl/nieuws/2016/47/wie-zijn-de-derde-generatie- Last consulted 10 May 2018.
CBS (2017). Politieke betrokkenheid in Nederland. Den Haag.
Coleman, J. S. (1966). Equality of educational opportunity. Washington: National Center for Educational Statistics (DHEW/OE).
Dancygier, R. M., & Laitin, D. D. (2014). Immigration into Europe: Economic discrimination, violence, and public policy. Annual Review of Political Science, 17, 43-64.
Darling-Hammond, L. (1995). Inequality and Access to Knowledge. In the
Handbook of Research on Multicultural Education, p465-83.
Deary, I. J., Taylor, M. D., Hart, C. L., Wilson, V., Smith, G. D., Blane, D., & Starr, J. M. (2005). Intergenerational social mobility and mid-life status attainment: Influences of childhood intelligence, childhood social factors, and education. Intelligence, 33(5), 455-472.
Denton, N. A., & Massey, D. S. (1988). Residential segregation of blacks,
Hispanics, and Asians by socioeconomic status and generation. Social Science Quarterly, 69(4), 797.
Dion, K. L. (2014). Responses to perceived discrimination and relative
deprivation. Relative deprivation and social comparison, 171-192. Psychology Press. Doomernik, J. (1995). The institutionalization of Turkish Islam in Germany and the Netherlands: A comparison. Ethnic and Racial Studies, 18(1), 46-63.
Dworkin, R. (2000). Sovereign virtue: The theory and practice of equality. Harvard: Harvard University Press.
Edding, F. (2013). The Refugees as a Burden a Stimulus, and a Challenge to the
West German Economy (Vol. 4). Springer Science & Business Media.
Eifert, B., Miguel, E., & Posner, D. N. (2010). Political competition and ethnic identification in Africa. American Journal of Political Science, 54(2), 494-510.
Elischer, S. (2013). Political parties in Africa: Ethnicity and party formation. Cambridge: Cambridge University Press.
Entzinger, H. (2014). The rise and fall of multiculturalism: The case of the
Netherlands. In Toward assimilation and citizenship: Immigrants in liberal nation-states (pp. 59-86). London: Palgrave Macmillan.
Esajas, M. (2017). Je moet twee keer zo hard werken. Racisme, discriminatie en de
sociaaleconomische positie van Surinaamse Nederlanders. Amsterdam: Surinaams
Inspraak Orgaan.
Etnobarometer (2017). Turkse Nederlanders over Nederland en Turkije: Opinie-onderzoek Ipsos in opdracht van de Volkskrant. IPSOS.
Finch, B. K., Kolody, B., & Vega, W. A. (2000). Perceived discrimination and depression among Mexican-origin adults in California. Journal of Health and Social
Behavior, 295-313.
Fox, J. (2004). Religion and State Failure: An Examination of the Extent and Magnitude of Religious conflict from 1950 to 1996. International Political Science
Review, 25(1), 55-76.
Gallagher, M., & Mitchell, P. (Eds.). (2005). The politics of electoral systems. Oxford: Oxford University Press.
Geddes, A., & Scholten, P. (2016). The politics of migration and immigration in
Europe. London: Sage.
Gemeente Amsterdam OIS (S.A.). Achthonderdduizend Amsterdammers.
https://www.ois.amsterdam.nl/visualisatie/bevolking.html Last consulted 7 June 2018. Gemeente Amsterdam OIS (2012). Fact sheet Hoogopgeleide Amsterdammers. Gemeente Amsterdam OIS (2015a). Tegengaan van segregatie in het
basisonderwijs. Amsterdam: Dienst Maatschappelijke Ontwikkeling.
Gemeente Amsterdam OIS (2015b). Amsterdamse Armoedemonitor 2015. Amsterdam.
Gemeente Amsterdam OIS (2018). Herkomst.
