Download Explaining anti-immigrant attitudes in the EU

Document related concepts

False consensus effect wikipedia , lookup

Transcript
Explaining anti-immigrant attitudes in the EU
A cross-country study on the determinants of intolerance
towards immigrants
Master Thesis Sociology
2011/2012, semester 2
Author:
Jaap Heirman
ANR:
S430270
Supervisor:
Inge Sieben
Place, Date:
Tilburg, August 2012
1
Abstract
Hostility between ethnic groups appears to be one of the most significant challenges facing European
countries today. Though research on anti-immigrant attitudes and its determinants has become quite
popular over the last decades, not every aspect of the subject has been addressed sufficiently in the
scientific literature. The present study aims to fill this research gap by, first, empirically examining
which individual-level and country-level characteristics explain one’s level of tolerance towards
immigrants in 27 EU countries and, second, to determine whether these relations are mediated by
perceptions of immigrants as an economic or a cultural threat. Hypotheses on which individual- and
country-level variables affect intolerance towards immigrants through perceived economic threat were
derived from Realistic Group Conflict Theory; hypotheses on which individual- and country-level
variables affect intolerance towards immigrants through perceived cultural threat were derived from
Symbolic Interaction Theory and the Cultural Capital Explanation. Using multilevel logistics
regression analysis, data from the 2008 European Value Study were analyzed in order to test the
hypotheses. The results from this analysis were somewhat mixed. At the individual level, the results
found were mostly in line with the predicted relationships. Conform Realistic Group Conflict Theory,
low education, low income and low job status levels significantly increased intolerance towards
immigrants, and this affect was mediated through perceived economic threat. Conform the Symbolic
Interaction Theory and the Cultural Capital Explanation, the effect of education on cultural threat was
mediated strongly by perceived cultural threat as well. At the country-level, findings on the effect of
several economic predictors and the size and culture of the immigrant population contradicted the
predicted relationships. Possible explanations for these unexpected results were sought in alternative
theories and in the methodological characteristics of this study.
Preface
This thesis was written as part of the master program Sociology. Writing the thesis has been
challenging at times, but has also been an educational and rewarding experience. I would like to thank
my supervisor Inge Sieben, as well as the staff and students of the master’s seminar Sociology for their
useful feedback and support.
2
Table of contents
1.
Introduction
p. 4
1.1.
Research Problem
p. 4
1.2.
Literature Evaluation
p. 6
1.3.
Research Goal and Question
p. 10
1.4.
What’s next?
p. 12
2.
Theoretical Framework
p. 13
2.1.
Realistic Group Conflict Theory
p. 13
2.2.
Social Identity Theory and the Cultural Capital Explanation
p. 16
2.3.
Conceptual Model
p. 20
2.4.
Control Variables
p. 21
3.
Methodological Framework
p. 23
3.1.
Data
p. 23
3.2.
Selection of Cases
p. 23
3.3.
Operationalization
p. 24
3.3.1.
The Dependent Variable: Tolerance towards Immigrants
p. 24
3.3.2.
The Independent Variables: Determinants of Tolerance towards Immigrants
p. 25
3.4.
Research Method
p. 28
4.
Results
p. 30
4.1.
Descriptive Statistics
P. 30
4.2.
Bivariate Analysis
p. 36
4.3.
Multilevel Regression Analysis
p. 38
4.3.1
The effect of individual-level variables on intolerance towards immigrants.
p. 40
4.3.1.
The effect of country-level variables on intolerance towards immigrants.
p. 42
5.
Conclusion and Discussion
p. 45
5.1.
Conclusion
p. 45
5.2.
Discussion
p. 47
7.
Literature References
p. 49
3
1. Introduction
1.1. Research Problem
“Few issues have had a greater impact on the politics and society of European nations than
immigration”, (Hollifield, 1997, p30). In the last century, immigrants have entered Europe in large
numbers, leading to a drastic change in the continent’s demographic build-up (McLaren, 2003). Mass
immigration to Europe started throughout the second half of the 20th century. The economy of Western
European societies was in the lift and there was a shortage of laborers. To address this issue, European
governments actively stimulated the immigration of low cost workers. Originally these migrant
workers were intended to stay only temporary (hence the term ‘guest worker’), yet not all migrant
workers eventually returned to their countries of origin. When the economy began to slow down in the
1970’s, active attempts were made by the European governments to reduce the number of migrants
coming into the countries. (McLaren, 2003). However, immigration policies failed to prevent an
increasing number of immigrants to settle in Europe (Meuleman, Davidov & Billiet, 2009). One of the
reasons was that most Western European governments still allowed family reunions, and expulsions of
immigrants were rare (McLaren, 2003). Moreover, when Western Europe stopped recruiting and
immigrants could no longer gain direct access to this part of the continent, Southern Europe began to
record significant immigration flows (Zick, Pettigrew & Wagner, 2008). In addition, these new
immigrant groups steadily grew in size due to the high birth- and low death rates of their young
population (Pettigrew, 1998). Although the immigrants filled an important need for unskilled labor,
they met with considerable resistance from the native population (Weldon, 2006). Tensions between
immigrant groups and the European majority started to especially grow when unemployment
intensified during the 1980’s (Pettigrew, 1998). Part of the majority group blamed the immigrants for
the social and economic problems their countries faced. Dissatisfaction regarding the immigrants
resulted in a rise of anti-immigrant violence and electoral successes of right wing anti-immigration
parties. (McLaren, 2003).
Nowadays, Europe is more multi-cultural than ever before in history (Coenders, 2001). Over 56
million immigrants have settled in Europe; more than on any other continent. In addition to
immigration from outside the continent, Europe has also experienced immigration from the south and
east to the north and the west. Besides first generation immigrants, millions of descendants of
immigrants and a vast array of groups like asylum seekers, illegal workers, international students, and
tourists make up modern Europe. (Zick, Pettigrew & Wagner, 2008). Despite Europe’s familiarity
with other cultures, the friction between immigrant and European majority groups has yet to dissolve.
Immigration and integration issues have been put high on the agenda of most European governments
(Scheepers, Gijsberts, & Coenders, 2002), yet efforts to improve interethnic relations have so far been
4
considered ineffective. A recent study on attitudes toward minorities in Europe reveals an increase in
hostility towards foreigners coupled with an increase in support for the implementation of restrictive
immigration policies (Semyonov, Raijman & Gorodzeisky, 2008). Furthermore, a substantial number
of people perceive immigration as having negative consequences, both economic and non-economic
(Meuleman, Davidov & Billiet, 2009). For instance, immigrants are often seen as a drain on social
services like welfare and unemployment insurance, or as detrimental to the host country’s culture
(Semyonov, Raijman & Gorodzeisky, 2008). In line with these sentiments, anti-immigrant populism is
still on the rise (Boomgaarden, 2006). Examples of this phenomenon can be observed all over Europe.
In the Netherlands, the anti-immigration party of Geert Wilders (the ‘Partij voor de vrijheid’) won 24
of the 150 seats in the 2010 general elections, making it the third largest party in the Dutch parliament.
Another example is the recent success of the True Finns party, a populist and nationalistic party in
Finland, who obtained 19% of the popular vote in the 2011 elections, making them the third largest
party in the Finnish parliament. Besides political shifts, other indicators of the problematic interethnic
relations in Europe are the numerous examples of ethnic prejudice and discrimination. 15% of the
respondents in the Eurobarometer of 2002 found ethnic discrimination of immigrants to be justified
(Zick, Pettigrew & Wagner, 2008) and an even greater number of people show subtle types of
prejudice towards minorities (Pettigrew, 1998). A field experiment of Adida, Laitin and Valvort
(2011) demonstrated the pervasiveness of anti-Muslim immigrant attitudes in Western Europe. With
experimental games, the researchers showed that the generosity of French majority members towards
Muslim immigrants decreased with an increase of Muslims in their midst. In a different study, they
found that Muslim immigrants in France were two and a half times less likely to receive a job
interview callback than their Christian counterparts (Adida, Laitin & Valvort, 2010). 1 Besides these
relatively subtle manifestations of prejudice, more extreme forms are known as well. In all European
nations, immigrant groups have experienced xenophobic violence, although sharp differences in
quantity exist among European nations (Pettigrew, 1998).
Concisely, hostility between ethnic groups seems to be one of the most significant challenges facing
European countries today. Scientific knowledge on anti-immigrant attitudes can help to address these
challenges more efficiently. The current research aims to add to our scientific knowledge on antiimmigrant attitudes by empirically examining intolerance towards immigrants in EU countries. Before
the research goal and questions will be elaborated further, the current state of the literature shall be
discussed in the following section. The aim of this review is not to provide a conclusive overview of
the literature on anti-immigrant attitudes, but rather to give the reader an idea of the current state of the
1
One should be careful to generalize these findings on Muslim immigrants to the entire group of immigrants.
Still, Muslims are a large group within the immigrant community in many European countries today.
5
research field and to point out missing elements in the existing literature, which will form the starting
point of the current research.
1.2. Literature Evaluation
Due to the transformation in demographics of European societies over the last decades, the issue of
ethnic relations in general and antagonistic reactions of the majority-group towards ethnic minorities
has gained increased relevance (Coenders, 2001). With this in mind, it should come as no surprise that
research on anti-immigrant attitudes has become quite popular. However, when reviewing the existing
literature, it becomes apparent that not all aspects of anti-immigrant attitudes and its determinants have
been addressed sufficiently. In the following paragraphs, the state of the literature will be discussed,
and missing elements in the existing literature will be pointed out.
Research on interethnic relations and anti-immigrant attitudes has a long standing tradition in the
social sciences, but until recently the research field has been dominated by a North American
perspective (Zick, Pettigrew & Wagner, 2008). One such North American study is that of Esses,
Dovidio, Jackson and Armstrong (2001). The aim of their research was to examine the role of
perceived competition on anti-immigrant attitudes in the USA and Canada. The researchers found that
respondents who scored high on Social Dominance Orientation hold less favorable attitudes towards
immigrants and immigration. Social Dominance Orientation refers to people’s believes whether
unequal social outcomes and social hierarchies are appropriate. Results also showed support for what
the researchers called, ‘the immigration dilemma’; immigrants who receive social services are
perceived negatively by members of the host country, yet immigrants who are economically successful
are perceived negatively by the members of the host country as well.
Another North American study is McDaniel, Nooruddin and Shortle’s (2010) research on the effects
of religion on anti-immigrant attitudes in North America. According to the authors, a person’s
religious identity can become intertwined with other identities such as a nation identity. American
history has showed this intermingling from its beginning. The conservative strain of American civil
religion, which the authors coined ‘Christian nationalism’, views America as holding a special
connection with god which requires it to be protected from outsiders and those who would do it harm.
The authors found a relation between religious conservatism and negative attitudes about immigrants
and that the adherents of Christian nationalism held the least favorable attitudes towards immigrants.
McDaniel, Nooruddin and Shortle’s study (2010) illustrates that, though North American studies can
provide insight in the interethnic relations of Europe as well, we must be cautious when generalizing
their results to other regions. Cultural opposition to immigrants appeared to be rooted in a particular
understanding of America’s origins as a Christian nation. While North America and Europe are similar
in certain ways, they are also very different in ways that could affect people’s perception of
6
immigrants (Zick, Pettigrew & Wagner, 2008). For instance, Europe has a vastly different background,
with a longer history of colonization, two wars that shaped interethnic relations and set the context for
migration and EU is sharply different in governance structure than Canada or the USA. In addition,
compared to Canada and the United states, European nations do not consider themselves to be
countries of immigration. (Zick, Pettigrew & Wagner, 2008) Considering these differences between
the two continents, the importance of European studies for understanding interethnic relations on this
particular continent becomes apparent. If we aim is to understand interethnic relations in Europe, we
cannot rely too heavily on the North American body of literature.
An inspection of the European literature shows that a large share of the studies is limited to the
examination of single countries. One example is the study that Verberk, Scheepers and Felling (2002)
conducted in the Netherlands. They examined, among other matters, the role of social class and
education on attitudes towards ethnic minorities. Their results showed that less educated people and
those who belong to the lower social classes are particularly likely to perceive ethnic minorities as a
threat. Perceived ethnic threat in turn had strong effects on the researchers’ measures of attitudes
towards ethnic minorities. These differences in attitudes were also shown to play a role in people’s
intended behavior towards immigrants. Respondents with a more negative attitude towards immigrants
generally maintained a larger distance between themselves and ethnic minorities in different domains
of social life. In addition, these respondents were more opposed against policies aimed at establishing
ethnic equality and were more likely to support ethnic discrimination.
An additional example of a study conducted in a single country is Zick, Wagner, van Dick and
Petzel’s (2001) study. The researchers aimed to explore the relation between ethnic attitudes and
attitudes towards the acculturation of ethnic minorities in Germany. Attitudes towards the
acculturation of ethnic minorities refers to people’s believe about the way that minorities should relate
to the culture of the majority. Their study distinguished three types of acculturation ideologies:
integration (immigrants valuing contact with the dominant culture while maintaining their ethnic
identity.), assimilation (immigrants valuing contact with the dominant culture while abandoning their
ethnic identity) and segregation (immigrants not expected to develop close relationships with the
dominant culture but maintaining their ethnic identity.) Based on experimental and survey data, the
researchers conclude that the more integrative a majority respondent’s acculturation attitude, the more
positive his or her behavior towards ethnic minority members.
Single country studies like the ones described in the previous paragraphs can tell us a lot about what
individual-level characteristics, like social class, education and acculturation attitude are related to
people’s perception of immigrants, and their intended behavior towards immigrants. However, being
7
restricted to a single country, these studies are unable to take the effects of contextual level-variables
on anti-immigrant attitudes into account.2
In order to examine the effect of country-level characteristics as well as individual level-characteristics
on anti-immigrant attitudes, cross-country studies are desirable. Though a large part of the literature
focuses on North America or on individual countries, this is not to say that there are no cross-country
studies of Europe on anti-immigrant attitudes. Over the last decades, a number of such studies have
been conducted. For example, Semyonov, Raijman and Gorodzeisky (2008) conducted a European
cross-country study on anti-immigrant attitudes. Their aim was to examine the extent to which
attitudes toward foreigners varied across twenty-one European countries, using data from the 2002
European Social Survey. Their findings showed that negative views tend to be more pronounced
among those who are economically and socially vulnerable and among those who hold conservative
ideologies. At the contextual level, the researchers found that negative attitudes towards foreigners are
likely to increase with the size of the foreign population and with support for right wing parties and
that negative attitudes towards foreigners are likely to decline with improved economic conditions.
Cross-country studies like these provide valuable insight in the influence of country-level
characteristics on anti-immigrant attitudes. However, a peculiar aspect of the cross-country literature
on anti-immigrant attitudes is that most of the studies directly examined the effect of individual-level
and country-level characteristics on anti-immigrant attitudes. The mechanisms that explain how certain
individual- and country-level characteristics could lead to negative immigrant perceptions remain
unspecified. An explanation that is suggested by several theories is that the perception of immigrants
as a threat mediates the relation between certain personal characteristics and country characteristics,
and a person’s attitude towards immigrants (Savelkoul et. al 2010). For example, individuals with a
low income might perceive immigrants as a threat for their own economic well-being and therefore
might be less tolerant towards them.3 Scheepers, Gijsberts and Coenders (2002) state that in many
studies, the perception of ethnic threat was proposed as the crucial mediating link between social
conditions and anti-immigrant attitudes, yet this link has hardly ever been tested empirically.
Additional research on the mediating effect of perceived ethnic threat would seem welcome in order to
assess how certain individual and country characteristics could lead to anti-immigrant attitudes.
A number of studies do shed light onto the mediating effect of perceived ethnic threat. In their study,
Savelkoul, Scheepers and Tolsma (2010) aimed to explain anti-Muslim attitudes in the Netherlands.
Contextual-level effects were determined by comparing different regions in the Netherlands. Their
2
Given that the research data was collected at a single point in time. Longitudinal single-country studies might
be able to determine the effects of contextual-level variables, yet these types of studies bring along additional
methodological difficulties (Meuleman, Davidov & Billiet, 2009).
3
The details of this mechanism shall be elaborated more in the theoretical section of this research.
8
results suggested that the size of the Muslim population fostered people’s perception of Muslims as a
threat, which in turn induced anti-Muslim attitudes. The effect of several individual-level
characteristics, like educational attainment and occupation status, on anti-Muslim attitudes was
mediated by perceptions of ethnic threat as well. The study of Verberk, Scheepers and Felling (2002)
that was discussed earlier this chapter, did also employ ethnic threat as a mediating variable. However,
being restricted to a single country, both studies suffer from the problems discussed earlier.
Scheepers, Gijsberts and Coencers (2002) did conduct a cross country research were ethnic threat was
employed as a mediating variable. The researchers examined in fifteen European countries to what
extent differences in support of ethnic exclusionism can be explained through individual
characteristics, contextual characteristics, and interaction effects between the two. In their research,
ethnic exclusionism was defined as opposition among European citizens to the granting of legal rights
to immigrants. Results showed that people with a low level of education, manual workers,
unemployed, and people with a low income are more likely to perceive ethnic out-groups as a threat.
Perceived threat in turn increased people’s support for ethnic exclusionism. At a contextual level the
researchers found that the larger the proportion of non-EU citizens in a country, the higher the ethnic
exclusionism. Few interaction effects were found.
Studies like the one of Scheepers, Gijsberts and Coenders (2002) are valuable for our understanding of
interethnic conflicts. Nevertheless, there is an ambiguity regarding the mediating effect of ethnic threat
that most studies who included this variable have failed to address. We have seen in studies that ethnic
threat is a core explanatory variable for immigrant attitudes (Scheepers, Gijsberts and Coencers, 2002)
However, the use of the concept ‘ethnic threat’ is rather vague, since it could refer to a number of
threats. Schneider (2008) suggests a division of ethnic threat in an economic and a cultural component.
Although there is widespread agreement on the existence of both forms of ethnic threat, the distinction
between the two often remains implicit in scientific research (Schneider, 2008). Most researchers
simply examine ‘ethnic threat’ as a one-dimensional concept. Since these two threats are highly
correlated, it is understandable they are often used as a single factor, but recent studies have found that
cultural and economic threats independently affect prejudice. An example of a study that made the
distinction between economic and cultural threat is Sniderman, Hagendoorn and Prior’s (2004)
research. They examined which factors influence citizens’ responses to immigrants in the Netherlands.
In their study, indicators of cultural and economic threat were both significantly related to measures of
prejudice but a perceived threat to the Dutch culture was found to be a stronger predictor for hostility
towards minorities. Sniderman, Hagendoor and Priors study shows that there are single country studies
on this matter. However, there has not been a thoroughly conducted cross-country research on the
distinction between economic and cultural threat. (Lucassen & Lubbers, (2012)
9
Recently, and attempt has been made by Lucassen and Lubbers (2012) to conduct a study that captures
all the missing aspects that we discussed in this literature evaluation. That is to say: Lucassen and
Lubbers conducted a cross-country study, with perceived threat as a mediating variable, and they
distinguished perceived economic threat from perceived cultural threat.One side note is that they did
not directly examine anti-immigrant attitudes, but instead examined determinants of support for the far
right (whose selling point often is their anti-immigrant standpoint). Their study revealed that perceived
cultural ethnic threats are stronger predictors for far right preferences than perceived economic ethnic
threats. In addition, they found that, economic threat and cultural ethnic threats were distinguishable
using factor analysis in 8 of the 11 countries they investigated. According to the researchers,
additional research that distinguishes economic from cultural threat would be a favorable contribution
to the existing literature, given how few studies have been conducted on the relative impact of
multiple types of ethnic threat.
To summarize: only a fraction of the studies in the research field are cross-country examinations of
Europe. Yet, if we want to know the relation between both country and individual characteristics and
anti-immigrant attitudes, these types of studies are essential. Moreover, as far as we are aware, besides
Lucassen and Lubbers (2012), none of the European cross-country studies have included perceived
threat as a mediating variable and made the distinction between the economic and the cultural
component of ethnic threat, though this seems important in order to assess exactly how certain
individual and country-level characteristics could lead to negative immigrant perceptions.
1.3. Research Goal and Question
The aim of this research is to fill the literature gap that was identified in the former section. I will
empirically examine how perceptions of immigrants as a cultural and as an economic threat mediate
the effect of individual and country characteristics on tolerance towards immigrants. Data from the
European Values Study (EVS) of 2008 will be used to investigate this matter. The EVS questionnaire
contains questions on both economic and cultural threat which allows me to examine their relative
impact on tolerance towards immigrants and whether the two aspects of perceived immigrant threat
are determined by different variables. In addition, using EVS data grants me the opportunity to
compare European countries and examine how country-level characteristics influence people’s
perception of immigrants as a cultural or economic threat. Although there is extensive collaboration
between European countries, the continent is still extremely heterogeneous with each country having
their own demographic, social and economic situation (Zick, Pettigrew & Wagner 2008). It is not
entirely clear how these contextual factors affect a person’s perception of immigrants. Country-level
data on immigrants is required to answer the research question, but can be difficult to obtain for
10
certain European countries. Therefore the analysis will be limited to the 27 EU countries, of which
sufficient country-level data can be accessed.
It is worth noting that anti-immigrant attitudes is a broad concept; multiple types of anti-immigrant
attitudes can be distinguished and certain types of anti-immigrant attitudes have been measured in
multiple ways in the existing literature. It is very well possible that these all have a different relation to
certain economic and cultural variables. Some of the types of anti-immigrant attitudes and behavior
that have been examined over the years are for example: ‘covert’ and ‘open’ prejudice (Pedersen &
Walker, 1997), preferred social distance (Smith & Dempsey, 1983), opposition to ethnic intermarriage
(Tolsma, Lubbers, Coenders, 2007), public opposition towards affirmative action policies (Coenders
& Schepers, 2003), denial of civil rights (Schuman et al., 1997), Solidarity towards immigrants
(Nielsen, 1985), discriminative behavior (Pettigrew, 1998) and ethnic mobilization and collective
action (Olzak, 1992).The type of anti-immigrant attitude that will be the focus of this study is
intolerance towards immigrants. Tolerance promotes a peaceful co-existence between various groups
and as such it is considered to be an important value in a multicultural society, whereas a lack of
tolerance can create friction between different social groups and can result in societal problems. To
examine its determinants seems a worthy research goal for any social scientist. By providing scientific
data on the determinants of intolerance towards immigrants, this study can provide an important
contribution to the political and societal debate on integration.
The research question that corresponds with the goal of examining how perceptions of immigrants as a
cultural and as an economic threat mediate the effect of individual and country characteristics on
tolerance towards immigrants, is the following:
‘To what extent are the effects of individual-level and country-level characteristics on an individual’s
level of tolerance towards immigrants mediated by his/her perception of immigrants as an economic
and a cultural threat?’
11
1.4.
What’s next?
In the second chapter, several theories on inter-ethnic relations and attitudes towards immigrants will
be discussed and will be used to derive testable hypotheses. In the third chapter, the research
methodology will be explained in detail. The research design, the data set, methods of analysis and the
operationalization of the core concepts will be discussed and methodological choices will be
elaborated. The hypotheses are tested using the 2008 EVS data set, and results from the analysis will
be discussed in the fourth chapter. Based on these results, the central research questions will be
answered in the fifth chapter. In this chapter, the strengths and limitations of this research will be
discussed and directions for future research will be suggested as well.
12
2. Theoretical Framework
Empirical studies reveal that attitudes towards immigrants are influenced by three major sources
(Semyonov, Raijman & Gorodzeisky, 2008). The first source is the characteristics of the respondent.
For example, a person’s educational attainment, religious affiliation and socio-economic status are
found to be correlated with anti-immigrant attitudes (Coenders & Scheepers, 2003; Manevska &
Achterberg’s, 2011). The second set of factors that affects anti-immigrant attitudes are characteristics
of the host societies. An example would be the economic situation in a country. The third set of factors
that affect anti-immigrant attitudes are the characteristics of the immigrant population. Among these is
for example the size of the total immigrant group in a country. Because both the second and the third
set of variables can be considered country-level variables, they will be referred to as such in the
following sections.
The relation between individual and country-level characteristics such as the ones mentioned in the
previous paragraph and anti-immigrant attitudes are explained by various theories in different ways.
Realistic Group Conflict Theory is one of the most frequently used theories to explain anti-immigrant
attitudes. Central in this theory is the presumption that a person’s perception of ethnic groups as an
economic threat affects his or her attitude towards these groups. Whether immigrants are seen as an
economic threat is in turn determined by certain individual and country characteristics that make this
person vulnerable to economic competition with immigrants. Other theories like Social Identity
Theory (Tajfel, 1981) and Manevska and Achterberg’s (2011) Cultural Capital Rxplanation focus on
perceived threat as a mediating variable between individual and country characteristics and antiimmigrant attitudes as well, but in contrast to Realistic Group Conflict Theory these theories stress a
cultural explanation instead of an economic one. A fear of losing one’s identity, losing cultural
resources or in other ways perceive immigrants as a threat to one’s own culture or identity can make a
person hold anti-immigrant attitudes.
In the following section, these theories will be discussed in detail and will be used to derive testable
hypotheses.
2.1. Realistic Group Conflict Theory
Realistic Group Conflict Theory was first introduced in the middle of the 20th century and has been
developed by various researchers (Campbell, 1965; LeVine & Campbell, 1972; Olzak, 1992). The
theory is known under multiple names (like Ethnic Conflict Theory or Group Conflict Theory) and in
multiple forms in the sociological literature, though the basic premise of the various interpretations of
the theory is the same: inter-group conflicts and negative inter-group attitudes are the result of
competition over scarce resources (Scheepers et al., 2002). Different social or ethnic groups in society
13
compete for the same resources on different types of markets (for example the labor market or the
housing market). This intergroup competition induces perceptions of threat which in turn lead to
negative attitudes towards competing groups (Savelkoul et al., 2010). Important to note is that not
actual competition but rather perceived competition affects one’s attitude of other groups (Esses
Jackson & Armstrong, 1998). Realistic Group Conflict Theory has its roots in economic thinking. The
theory assumes that individuals behave rationally in their battle for scarce resources4. An antagonistic
attitude could be considered a rational response of individuals in their struggle for scarce resources as
it reflects attempts to remove competition (Esses et al., 2001). These attempts may entail
discrimination and opposition to policies and programs that may benefit the out-group. One may
express negative opinions to convince both the in-group (one’s own group) and the out-group (groups
that are different from the self) of the competitor’s lack of worth or one may try to avoid or ignore
rivalling groups in an effort to remove competition (Esses et al., 2001). Perceived group competition is
likely to take the form of a zero-sum game. If a person thinks that the more the other group obtains,
the less is available for one’s own group and thus that any gains the other group makes comes at the
expense of one’s own group, one is more inclined to perceive the other group as competitors. When
members of the majority perceive that ethnic minorities acquire scarce resources, they may feel that
the majority population can no longer claim these goods and thus the majority group may develop
hostile attitudes towards the minority population (Verberk et al, 2002).
Some groups are more likely to be seen as competitors than others. Out-groups that are salient and
distinct from the in-group are especially likely to be seen as rivals (Esses et al., 2001). However, outgroups must also be similar to the in-group on dimensions that makes them likely to take resources
from the in-group. They must be interested in similar resources and they must be in a position to take
these resources (Esses et al., 2001). One of the core ideas of Realistic Group Conflict Theory is that
there is a relation between one’s position in society and one’s attitude towards immigrants (Hello,
2003). Social groups that hold similar positions to ethnic minorities may experience higher levels of
ethnic competition and as a consequence may hold more unfavorable views of ethnic minorities. In the
EU, the majority of immigrants are generally located in the lower strata of society. This means that
members of the majority group that are located in the lower strata as well, i.e. those with a low
educational level, a low income level, performing manual labor or who are unemployed, will have to
compete over resources with immigrants groups more so than members of the majority group that are
located in the higher strata of society. For example, the more educated have obtained an advantaged
position in society and will face less competition from ethnic minorities than the less educated (Hello,
4
Although one would perhaps expect this to be a struggle between individuals, people regard themselves as
members of in-groups and out-groups on the basis of, among other factors, their ethnic background.
Consequently, the battle for scarce resources is perceived as a struggle between groups (Hello, 2003).
14
2003). Therefore, highly educated are less likely to perceive immigrants as a threat to their socioeconomic well-being. Following the logic of Realistic Group Conflict Theory, the following
hypotheses can be formulated:
Hypothesis 1: Members of the majority group with a: low educational levels, b: low income levels, c:
performing manual labor or performing low skilled labor d: are unemployed, are more likely to be
intolerant towards immigrants. This effect is mediated through the perception of immigrants as an
economic treat.
In addition to being in the same societal strata as immigrants, there are other factors that determine the
amount of competition with ethnic minorities. Levine and Campbell (1972) state that when
competition over resources is present, proximity and contact will increase intergroup hostility.
Competition with immigrants is likely to be the highest for people that live in areas with a high
concentration of immigrants. Consequently, members of the majority that live in areas with a high
concentration of immigrants are more likely to view immigrants as an economic threat. Because high
concentrations of immigrants are usually found in urban areas, the following hypothesis can be
formulated:
Hypothesis 2: Members of the majority group who live in urban areas are more likely to be intolerant
towards immigrants. This effect is mediated through the perception of immigrants as an economic
threat.
Besides these individual-level variables, Realistic Group Conflict Theory can also be used as a
framework to predict the effects of contextual-level structural characteristics on antagonistic attitudes
(Tolsma, Lubbers, Coenders, 2007). One of these contextual-level characteristics is the size of the total
immigrant group in a country. Based on Realistic Group Conflict Theory, one would expect that the
larger the size of the immigrant group, the more the majority group will have to compete with them
over scarce resources and thus the more negative their attitude towards immigrants. Even if resources
are not necessarily low, but the relative number of minorities is high, people may still perceive ethnic
minorities as a threat to their socio-economic well-being (Hello, 2003). In addition, the actual
competition for scarce resources at the contextual level is likely to be higher if there are little scarce
resources (Hello, 2003). In other words, the economic situation of a country may influence a person’s
attitude towards immigrants. In times of high unemployment and in times of economic recession,
competition can expected to be higher since there are little scare resources. The actual competition for
scarce resources is not only higher if there are few scare resources, but also if there is a decrease in
scarce resources (Hello, 2003). Though, the economic situation of a certain country can be considered
excellent compared to other countries, if it degenerated compared to the situation in the last few years,
competition can still be expected to be higher since people perceive a decrease in scarce resources.
Thus, when the economy has shrunk, or unemployment figures have risen in the last years, actual
15
competition is expected to be high. Even if the relative number of people from an ethnic minority is
rather low, unhealthy economic conditions may cause people to perceive minority members as a
threat. Thus, the following country-level hypotheses can be derived from Realistic Group Conflict
Theory:
Hypothesis 3: The larger the total immigrant group in a country, the more likely that members of the
majority group are intolerant towards immigrants. This effect is mediated through the perception of
immigrants as an economic threat.
Hypothesis 4: The higher the unemployment rate in a country, the more likely that members of the
majority group are intolerant towards immigrants. This effect is mediated through the perception of
immigrants as an economic threat.
Hypothesis 5: The stronger the growth of the unemployment rate in a country, the more likely that
members of the majority group are intolerant towards immigrants. This effect is mediated through the
perception of immigrants as an economic threat.
Hypothesis 6: The worse the GDP in a country, the more likely that members of the majority group
are intolerant towards immigrants. This effect is mediated through the perception of immigrants as an
economic threat.
Hypothesis 7: The smaller the growth of GDP, the more likely that members of the majority group are
intolerant towards immigrants. This effect is mediated through the perception of immigrants as an
economic threat.
2.2. Social Identity Theory and the Cultural Capital Explanation
An individual’s perception of immigrants as an economic threat is not the only determinant of his or
her attitude towards immigrants. Anti-immigrant attitudes have been known to exist without economic
competition (Tolsma, Lubbers, Coenders, 2007). Esses et al. (2001) state that it is likely that the
majority group may see immigrants as competing over less tangible assets as well. For instance,
immigrants may be perceived to compete with the majority group over which culture and values are
the most ‘correct’. Non-economic determinants of anti-immigrant attitudes have been relatively
neglected by the literature until recently (Taylor, 1998). Similar to the fear of losing economic
resources, a fear of losing cultural resources can induce unfavorable views towards immigrants
(Tolsma, Lubbers, Coenders, 2007). Sniderman et al. (2004) found that when economic conditions are
good, considerations of group identity can even overshadow those of economic concerns. In their
research, hostility to minorities was best predicted by a perceived threat to the Dutch culture. Lubbers
and Güveli (2007) examined whether support for the ‘LPF’, in the former decade a right wing populist
16
party in the Netherlands, was based more on cultural or economic threats. Their study revealed that
perceived cultural ethnic threats were stronger predictors for LPF voting than perceived economic
ethnic threats, although economic threat was a relevant predictor as well.
Although group interest can clash over intangible goods like cultural identity and values as well,
Realistic Group Conflict Theory focusses foremost on conflicts over economic interest (Sniderman,
2004). Therefore the theory is less suitable for explaining the effects of cultural threats on antiimmigrant attitudes. A cultural interpretation of Realistic Group Conflict Theory can be used to
predict the mediating effect of cultural threat, but it seems that in order to fully understand the cultural
aspect of intergroup competition, it would be wise to turn to other theories as well.
One theory that can help explain the perception of immigrants as a cultural threat is Social Identity
Theory. Social Identity Theory was first introduced by Tajfel and Turner in 1979 and has been further
developed in various papers (Tajfel, 1982; Tajfel & Turner, 1985). Nowadays it is one of the most
frequently used theories in research on interethnic relations. Social identity Theorists perceive hostility
between ethnicities as a clash of cultural identities. The theory attempts to explain intergroup attitudes
and behavior by referring to the underlying psychological processes of developing and maintaining a
group identity. Important to note is that Social Identity Theory and Realistic Group Conflict Theory
are not mutually exclusive. It is likely that, to some extent, both concerns for economic well-being and
for identity and culture underlie reactions to minorities in Europe. Tajfel and Turner (1978) meant for
Social Identity Theory not to replace Realistic Group Conflict Theory, but rather to supplement it in
some respects. The complementary nature of these theories becomes apparent in, for example,
Scheepers, Gijsberts & Coenders (2002) work, where traditional Realistic Group Conflict Theory and
elements of Social Identity Theory were combined into a broad Ethnic Competition Theory.
Social Identity Theory states that the perceived distinction between one’s own group and other groups
lies at the basis of inter-group attitudes and inter-group behavior (Coenders, 2001). This distinction is
made through the process of ‘social categorization’ (Coenders, 2001). Social categorization means
“bringing together social objects or events in groups which are equivalent with regard to an
individual’s actions, intentions and system of belief” (Tajfel, 1981 p. 254). It is not merely a cognitive
tool to systemize the social world; social categorization also defines an individual’s place in society.
One of the basic premises of Social Identity Theory is that a core component of an individual’s sense
of self is based on which groups they belong to. An important concept here is ‘social identity’. This is
defined as “the part of an individual’s self-concept which derives from his knowledge of his
membership of a social group (or groups) together with the value and emotional significance attached
to that membership” (Tajfel, 1981, p.255). A person’s social identity may be positive or negative,
depending on the evaluation of social groups that contribute to one’s individual identity. Since people
strive for a positive self-concept, they are motivated to positively evaluate the groups that are at the
17
basis of their own social identity. According to Tajfel (1981), social identities are comparative in
nature. As Tajfel states: “the definition of a group (national, racial or any other) makes no sense if
there are no other groups around” (1981, p. 258). A positive social identity is to a large extent based
on favorable comparisons with members of other groups. In order for people to evaluate their own
group positively they are motivated to evaluate other groups negatively (Sniderman et al., 2004). The
result of this process of identification and contra-identification is ethnocentrism. The more strongly
individuals identify with their groups, the more bias they will show in favor of these groups against
salient out-groups (Duckitt, 1998). This explains why even in the absence of economic conflict, antiimmigrant attitudes may arise.
Coeders (2001) criticizes Social Identity Theory for lacking explicit testable notions regarding
variations in ethnocentrism between social categories and countries. Many people care about their
country’s national identity and culture but which types of individuals perceive immigrants as the
biggest threat to their culture/identity? In the literature, a relationship has frequently been observed
between educational attainment and opinions about minorities (Manevska and Achterberg, 2011).
People with a higher education are less prejudiced towards ethnic out-groups than those with a lower
education. A possible explanation for this observation is that people with a higher education are less
prone to in-group favoritism (Coenders and Scheepers, 2003). The relationship between educational
attainment and ethnic intolerance has been established consistently across time and in different
countries (Coenders & Scheepers, 2003). Martire and Clark (1982) found less anti-semitism among
the higher educated and Taylor et al. (1978) found that higher educated white Americans are more
supportive of racial integration. Considering that the higher educated may experience lower levels of
ethnic economic competition, there can be economic explanations for these observations but
researchers argue that there is a cultural explanation as well (Manevska and Achterberg, 2011).
Education in a cultural sense plays a role in questions of tolerance (Manevska and Achterberg, 2011).
Gabennesh (1972) stated that highly educated personas are better able to recognize cultural
expressions and to understand their meanings, are less inclined to reject deviant lifestyles and more
willing to value cultural diversity and to accept cultural differences. Following the logic of this
cultural explanation, the following hypotheses can be formulated:
Hypothesis 8: Members of the majority group with low educational levels are more likely to be
intolerant towards immigrants. This effect is mediated through the perception of immigrants as a
cultural threat.
In addition, certain characteristics of the immigrant population can be expected to make people more
prone to perceiving immigrants as a cultural threat. If the majority group perceives the immigrant’s
culture to be radically different from their own, it is likely that immigrants will be viewed as a cultural
threat more so than if the immigrant culture is perceived to be very similar. In other words, the larger
18
the cultural distance between the total immigrant group and the majority group, the more likely it is
that immigrants will be viewed as a cultural threat. This prediction is in line empirical studies like
Esses, Dovidio, Jackson and Armstrong’s (2001) research. They found that perceived similarity has a
role to play in the majority group’s formation of attitudes on immigrants. When a common identity
was introduced in one of the researchers’ experiments, (for example by emphasizing the shared
membership of a majority and a minority group member in a social category, this produced more
positive attitudes. Therefore, the following will be expected:
Hypothesis 9: The larger the cultural distance between the total immigrant group and the majority
group, the more likely that members of the majority group are intolerant towards immigrants. This
effect is mediated through the perception of immigrants as a cultural threat.
In addition, one would expect that the bigger the total immigrant group in a country is, the more likely
that immigrants are viewed as a cultural threat. According to Manevska & Achterberg, a greater share
of (non-Western) immigrants supposes a larger input of different cultures within society, leading to
intensification of the experienced cultural conflict and to greater amounts of experienced cultural
threat. Schneider (2008) found that the larger a culturally threatening out-group, the higher the average
perceived threat in a country. In line with these findings, the following will be hypothesized.
Hypothesis 10: The larger the total immigrant group in a country, the more likely that members of the
majority group are intolerant towards immigrants. This effect is mediated through the perception of
immigrants as a cultural threat.
In addition to the country-level, one would expect the effect of the size of an immigrant group on antiimmigrant attitudes to come into play at the individual-level as well. If people more directly
experience the effects of immigration in their everyday life, it is likely that they view immigrants more
as a cultural threat. Since high concentrations of immigrants are usually found in urban areas, I
hypothesize that:
Hypothesis 11: Members of the majority group who live in urban areas are more likely to be
intolerant towards immigrants. This effect is mediated through the perception of immigrants as an
cultural threat.
19
2.3. Conceptual model
Combined, the hypotheses form the following conceptual model.
Education
Perception of
economic threat
Income
Tolerance towards
immigrants
Job status
Employment
Perception of
cultural threat
Living in urban area
Individual level
% Immigrants
Contextual level
Unemployment rate
Rise in unemployment
GDP
Growth GDP
Cultural distance
Model 1
In order to examine whether the two types of perceived immigrant threat are determined by different
variables I will, besides testing the theoretically hypothesized relations, examine the impact of all
individual and country-level variables on both types of immigrant threat as well. For example: even
though job status is theoretically only linked with perceiving immigrants as an economic threat, I will
examine whether it is linked with perceiving immigrants as a cultural threat as well.
In addition to the effects described in model 1, could it be there are interaction effects between the
individual-level and country-level predictors? While it is a possibility, there is little reason to believe
that strong effects will be found. Manevska and Achterberg (2011) found no significant interaction
effects between education and country-level variables on perceived ethnic threat. Mayda (2006)
examined an interaction effect of education and GDP (gross domestic product) on anti-immigrant
attitudes and found that it was not significant. Scheepers, Gijsbert and Coenders (2002) examined the
whether the effects of individual characteristics on ethnic exclusionism varied along countries. While
20
they did find some differences in the effects of individual characteristics within counties, only one of
the ten hypothesized cross-level interactions reached significance. An additional argument is that the
theories from which the hypotheses in my research were derived do not explicitly argue for any
interaction effects. As a consequence, the focus of this research will be on direct effects and no
interaction effects will be examined.
Important to note is that, although there are several different immigrant groups in the EU and people
may hold a different opinion of one immigrant group than of the other, no distinction between
different immigrant groups within the EU will be made in the current research. There are several
reasons for this decision. The main reason is that the EVS data that will be used to test the hypothesis
did not distinguish between different immigrant groups either. In addition, it will be difficult to make a
cross-country comparison on anti-immigrant attitudes without aggregating the numerous ethnic groups
across the EU. Furthermore, according to Tajfel (1981), people perceive greater homogeneity among
out-group members than among in-group members. Individuals consider members of the out-group in
a relatively uniform manner as undifferentiated items in a unified social category. In other words,
people often generalize different groups of immigrants, treating them all as ‘immigrants’.
2.4. Control variables
In addition to the variables that are discussed in the previous section, a number of control variables
will be included in this research; the first one being age. Age will be a control variable because it has
frequently been reported to be significantly correlated with anti-immigrant attitudes as well as with
education and predictors of socio-economic status. There are multiple ways by which age could affect
anti-immigrant attitudes. One might expect young people to be more are more tolerant towards
immigrants because they are generally more open to new and foreign impulses (Hernes & Knudsen,
1992). However, it is also possible that that the younger generation, who is in the process of acquiring
a house and making a career, has to compete more heavily with immigrants and thus is more likely to
hold negative attitudes towards them than the older generation (Hernes & Knudsen, 1992). Though
arguments can go both ways, most empirical studies suggest that older people are more likely to
exhibit anti-immigrant attitudes than younger people.
The second control variable is gender. Gender, much like age, has been reported to be significantly
correlated with anti-immigrant attitudes and with education and predictors of socio-economic status.
Theoretical arguments concerning the direction of the effect of gender on anti-immigrant attitudes run
both ways. Though women are catching up, at present, men are still more active on the labor market.
This may mean that men experience more competition from immigrants and therefore should be more
prejudiced towards immigrants than women. Another line of reasoning is that women hold more
21
negative views against immigrants because immigrants generally come from cultures with less
equality between the sexes (Hernes & Knudsen, 1992) and because of this, women may feel more
threatened by the influence of this immigrant culture. An examination of the empirical literature shows
that generally men are more likely to exhibit anti-immigrant attitudes than women, though significant
effects are not always found (Scheepers, Gijsberts, & Coenders, 2002; Semyonov, Raijman &
Gorodzeisky, 2006; Tolsma, Lubbers, Coenders, , 2007).
Religion was chosen as a control variable since it may correlate with both the perception of
immigrants as a cultural threat and with several individual-level characteristics (Knoll, 2009). Based
on Symbolic Interaction Theory, we can argue that citizens who ascribe to a religion are likely to be
intolerant of those they perceive as symbolic threats to their religious identity. In the EU, where most
people that ascribe to a religious affiliation are Christian, Christians are expected to express higher
levels of anti-immigrant attitudes than nonbelievers and non-Christians. What about members of
marginalized religions, like Buddhism or Judaism? Knoll (2009) argues that members of minority
groups are more likely to emphasize with other marginalized groups, such as immigrants, and
therefore should express lower levels of anti-immigrant attitudes. In line with these expectations,
Scheepers (2002) found that Christians were indeed significantly more in favor of ethnic exclusionism
of legally established immigrants than non-believers and non-Christians. However, this relation is not
always found (Savelkoul et al., 2010) and in the literature, counterarguments have been made as well.
For example, it has been argued that religious individuals may hold more positive immigrant attitudes
compared to non-believers, because they are more likely to attempt to live according to the JudeoChristian value that teaches to ‘love thy neighbor’ (Knoll, 2009). It should also be noted that there are
some highly secular countries among EU nations, and it might be possible that people with no
religious affiliation are more intolerant towards immigrants since they may perceive (religious)
immigrants to be a threat to the ‘secular values’ of their society.
22
3. Methodological Framework
3.1. Data
The central question of this research will be answered by testing the hypothesized relations with data
from the European Values study5. The EVS is a large scale, cross-national and longitudinal survey
research program that covers a wide range of human values. The EVS started in 1981 and is repeated
every nine years. The data that will be used for this research is derived from the fourth and last wave
in 2008 in which Europeans from forty-seven countries participated. Data was gathered by structured
questionnaires. The dataset is an excellent resource for the present research since it fits the research’s
subject matter and it allows for a comparison of a large group of European countries. Data on countylevel variables will be derived from Eurostat, the directorate-General of the European Commission6.
3.2. Selection of Cases
Not all cases in the EVS were selected for the data analysis. The inclusion of a case was based on
several criteria. First, since the goal of this research was to examine attitudes towards immigrants in
the EU, only respondents from the 27 EU countries were selected for further analysis.7 Second, only
respondents that are part of the majority group in their present country were selected for analysis.
Other respondents were excluded since the hypotheses of the present study solely make predictions
about the majority group’s attitudes towards immigrants. Important to mention here is how the
majority group was defined in this research. Only respondents whose parents were both born in the
respondents’ present country were selected for further analysis. In order to assess whether this was the
case, the following questions were asked: “Was your father born in [country]?”, “(yes/no)” and “Was
your mother born in [country]?”, “(yes/no)”.8 While usually not considered as immigrants,
respondents born in their country of residence but with one or both parents born in another country are
likely to have strong bonds with the immigrant group, thus affecting their attitude towards immigrants.
Therefore, this group was excluded from the majority group analyses as well. Third, only respondents
between age 18 and 70 were selected for analysis. Realistic Group Conflict Theory states that negative
inter-group attitudes are the result of competition over scarce resources. Since generally only
respondents of working age have to compete with the immigrant group for economic resources,
5
See: www.europeanvaluesstudy.eu
6
See: ec.europa.eu/eurostat
7
Though Northern Ireland is part of The United Kingdom, the EVS has examined Northern Ireland independent
from the rest of The United Kingdom. Our country-level N is therefore 28 instead of 27.
8
[Country] being the respondent’s country of residence.
23
respondents younger than 18 and older than 70 were excluded from further analysis.9 Cases with data
missing that was required to determine if a case fit the inclusion criteria mentioned above (age,
member of the majority group, country of residence), were excluded. Finally: Missing values on the
dependent and the mediating variables might distort the research findings. Therefore, only respondents
with no missing values on the questions that were used to determine one’s tolerance towards
immigrants and one’s perception of immigrants as an economic and a cultural threat were selected for
analysis. After selection, a total number of 27.077 cases were available for further analysis.
3.3. Operationalization
3.3.1.
The Dependent Variable: Tolerance towards Immigrants
The dependent variable in this research is ‘tolerance toward immigrants’. This was measured with the
EVS question: “On this list are various groups of people. Could you please sort out any that you
would not like to have as neighbors?’’. One of the fifteen groups of people on the list was
‘immigrants’. The variable is measured dichotomous; respondents either stated they do not like to
have immigrants as neighbors or they did not mention immigrants as a group of people they would not
like to have as neighbors. One downside of this dichotomous operationalization is that there is little
room for nuance in the answers of the respondents. In addition, being measured by a single indicator
makes it very difficult to assess the validity of the operationalization. It is also worth mentioning that
the item used in the EVS is a relatively ‘strong’ way of measuring tolerance because it challenges
someone’s direct living environment. It is very well possible that people might tolerate immigrants in
their country but have mixed feelings towards having them as neighbors. In addition, it is important to
be aware of the susceptibility of research on anti-immigrant issues to social desirability effects. Some
respondents might have the tendency to answer in a manner that will be viewed favorably by the
interviewer. Since participants in the EVS were interviewed instead of having to fill out the
standardized questionnaire themselves, this research is especially susceptible to these particular
response biases. A side note to this matter is that although social desirability effects could affect the
research results, expressing negative feelings towards immigrants has become more socially
acceptable in the last years, given the public debate.
9
It is possible that younger and older respondents compete for economic resources with immigrants as well. For
instance, it is possible they compete for welfare or other tax expenditures. However, as we will see further down
this chapter, economic threat is operationalized in terms of immigrant threat to the labour market and thus we
will focus on the labour force.
24
3.3.2.
The Independent Variables: Determinants of Tolerance towards Immigrants
The variables ‘perception of immigrants as a cultural threat’ and ‘perception of immigrants as an
economic threat’ were measured the EVS question: “Please look at the following statements and
indicate where you would place you views on this scale”. Answers were given on a ten point Likert
scale. In order to acquire respondents’ perception of immigrants as a threat, answers were mirrored.
The statement “Immigrants take jobs away from natives in a country” measured perceived economic
threat. Similar to the dependent variable, perceived economic threat is measured by a single indicator,
which makes it difficult to assess the validity of the operationalization. In addition, perceived
economic threat is a fairly broad concept, constituting of multiple economic factors, and one can argue
whether this can be measured by a single item at all. Besides questions about the labor market,
questions regarding the perceived burden of immigration on health and welfare services or
immigrant’s contribution to tax income are frequently used to determine the perceived economic threat
as well (Schneider, 2008; Scheepers et all, 2002; McLaren, 2003). Still, we believe that the item used
in the EVS, while not covering the entire concept of economic threat, should be sufficient to measure
respondents’ perceived economic threat. Concerns over the impact of immigrants on the labor market
have always played a central role in the immigration debate (Borjas, 2003). Labor market concerns are
likely considered to be most important type of economic threat that immigrants might pose.
The second statement, “A country’s cultural life is undermined by immigrants” measured perceived
cultural threat. Like perceived economic threat, perceived cultural threat was measured by a single
indicator. While this research uses a relatively general question to assess the perceived cultural threat,
questions that are used to determine this concept in other studies are sometimes more specified. For
example, Lucassen and Lubbers (2011) used a four item scale to measure perceived cultural threat
with, among others, questions related to customs and traditions, religion and inter-ethnic tensions.
While the EVS data does not grant the opportunity for a more elaborate scale to measure cultural
threat, I believe the single item operationalization in the EVS should be sufficient to determine
respondents’ general perceived cultural threat.
Education was measured with the question: “What is the highest level you have completed in your
education?”. Respondents’ answers were recoded according to the international standard classification
of education (ISCED), ranging from ‘inadequately completed elementary education’ to ‘upper level
tertiary education’. To avoid complexity, these categories were further recoded into three categories:
lower education, medium education and higher education.
Income was measured with the following question: “Here is a list of incomes and we would like to
know what group your household is, counting all wages, salaries, pensions and other incomes that
come in. Just give the letter of the group your household falls into after taxes and other deductions.”
25
To account for country differences, respondents’ answers were recoded according to their purchasing
power parity (ppp) into three categories: low income, medium income, and high income.
In order to assess respondent’s job status, respondents were first asked the questions: “What is the
name and title of your main job?” and: “In your main job, what kind of work do/did you most of the
time?” Answers were recoded according to the EGP occupational class categories into eleven
categories, ranging from ‘self-employed farmer’ to ‘higher controller’. To avoid unnecessary
complexity, these eleven categories were further recoded into five categories: independent/selfemployed, high non-manual labor, low-non manual labor, high manual labor and low manual labor.
Employment was measured by the question: “Are you yourself gainfully employed at the moment or
not? Please select from the card the employment status that applies to you.” Ten answer categories
were presented in the EVS questionnaire. These categories included, besides unemployed and
employed categories, also categories like student and housewife. Although respondents that fall in the
latter two categories do not have a job, they do not compete with immigrants over resources on the
labor market and thus cannot be included into a category with unemployed respondents. The original
ten answer categories were therefore recoded into three new categories: unemployed, employed and
not unemployed.
The variable living in urban area was measured by the question: “Size of town where you live now”
Respondents could choose from eight answer categories, ranging from (1) under 2000 to (8) 500.000
and more. These answer categories were recoded into three categories. Lower urban environment,
middle urban environment and higher urban environment.
In order to assess the age of the respondents, the following question was asked: “Can you tell me your
year of birth please.” (Q87). The year of birth of the respondent was then calculated into age.
The control variable gender was measured by the following question: “Sex of the respondent, 1- Male;
2- Female” (Q86)
The control variable religion was measured by the question: “Do you belong to a religious
denomination?” ( yes/no), and the follow-up question “Which one?”. Answers were recoded into
three categories. Christianity (Protestantism and Catholicism) can be considered the dominant
religions in Europe and are given an independent category. While the Orthodox Christian faith can be
considered the dominant religion in several European countries as well, Coenders and Scheepers
(2003) found that Catholic and Protestant respondents scored significantly higher on several
dimensions of nationalism and ethnic exclusionism than Orthodox Christians. In their research,
Orthodox Christians bore more resemblance to respondents from the ‘other religion’ category in terms
of nationalisms and ethnic exclusionism. Therefore, Orthodox Christianity will not be included with
Protestantism and Catholicism but will instead be grouped in the variable other religion. Other
26
religions like Judaism and the Islam are grouped together in that category as well. The last category,
no religion, is for the respondents that do not belong to a religious denomination.
Data from the European labor force survey 2008, derived from Eurostat, was used as an indicator for
the country-level variable ‘percentage of immigrants’. This data solely reported on the population age
25-55 instead of the entire population, which can be considered a disadvantage of using this data.
However, an advantage was that we could analyze respondents along their type of background. In the
data, the distinction is made between native background, second generation immigrant and first
generation immigrant. The category ‘second generation migrants’ is further specified into people with
a mixed background (one parent born abroad) and people with a foreign background (two parents born
abroad). In this research, the first generation immigrants and both types of second generation migrants
are added up to obtain the percentage of immigrants in each of the countries. 10
Data on the unemployment rate of EU countries in 2008 was derived from Eurostat. The main source
used by Eurostat for the unemployment figures is the European Union Labour force survey.
The unemployment growth rate was measured over a timespan of five years preceding the year that the
EVS study was conducted. We chose a five year period instead of a shorter one because attitudes are
not likely to change overnight. Psychological research has shown that strong attitudes resist most
attempts at change (Olson, 1993). Most likely, it takes a significant amount of time before the
economic situation of a country affects one’s attitude towards immigrants and therefore, using a five
year period is fairly common in research on anti-immigrant attitudes (Scheepers, 2002).
Data on the GDP per capita in PPS (purchasing power standards), derived from Eurostat, was used to
measure the GDP of the EU countries in 2008. GDP per capita in PPS was expressed in relation to
EU-27=100 and is thus a relative measure of GDP.
Since GDP per capita in PPS was expressed in relation to EU-27=100, this measure is less suited for
comparison over time. Therefore, to examine the GDP growth we used data on the ‘real GDP growth
rate’ instead, which was also derived from Eurostat. Real GDP growth was measured during a five
year period as well. Again, the argument here is that this is fairly common in research on antiimmigrant attitudes (Scheepers, 2002) and that it takes time for the economic situation to start
affecting one’s opinion of immigrants.
One remark is that Finland was not represented in the European labor force survey 2008. Data on the number
of immigrants in Finland was estimated based on the percentage of total foreigners found in another report of
Eurostat (Statistics in Focus, 45/2010)
10
27
The country level variable ‘cultural distance’ was measured by the following question: “Please tell me
for each of the following whether you think it can always be justified, never justified or something in
between”:

Taking the drug marijuana or hashish;

Homosexuality;

Abortion;

Divorce;

Euthanasia (terminating the life of the incurably sick);

Prostitution;

Scientific experiments on embryos;

Artificial insemination or invirto fertilization;

Suicide.
Answers were given on a scale ranging from 1 (never) to 10 (always). The average score of the
immigrant group and that of the majority group were calculated separately for each individual country
on the combined set of items. The immigrant group score of a country was then subtracted from the
majority group score of that country in order to calculate the cultural distance between the majority
group and the immigrant group in a country.11 Since both a positive and a negative score represent a
cultural distance, the absolute values of these scores were taken. In order to assess the internal
consistency of the cultural distance scale, Chronbach’s alpha was examined. With an alpha of 0.840,
the internal consistency of the scale can be considered fairly high. In addition, the alpha if item deleted
was lower for every item on the scale, indicating that the internal consistency of the scale could not be
improved by dropping an item. One last remark about this variable is that the cultural distance score
of the immigrant group in Finland and Romania was based on a sample of respectively sixteen and
thirteen respondents. Since this can be considered quite small, this may have implications for the
representativeness of the cultural distance score of immigrants in these countries.
3.4. Research Method
Since data from the EVS was essentially gathered at one point in time, we can consider this to be a
cross-sectional survey design. The goal of this research was to examine how individual and country
characteristics affect an individual’s level of intolerance towards immigrants, and thus the unit of
analysis is the individual. Since these individuals are embedded in countries, multilevel analysis was
chosen. Because tolerance towards immigrants is measured dichotomous, logistic regression analysis
was required to examine how the probability of an individual being intolerant towards immigrants is
11
The immigrant group was defined in the same manner as the country-level variable ‘percentage of imigrants’
28
affected by various predictors. A step-wise logistic regression analysis consisting of six different
models was conducted to examine the data and test the hypotheses. The different models are be
elaborated below.
Model 1: The first model includes the intercept only. This model makes it possible to observe whether
individuals within a country resemble each other more closely, by assessing how much variance
between individuals is explained at the individual level relative to how much variance is be explained
at the country-level.
Model 2: Model 2 consists of model 1, the individual level variables (education, income, job status,
employment and living in urban area) and the control variables (age, gender and religion). This model
makes it possible to examine whether characteristics of the individual and his or her situation affect
the level of tolerance towards immigrants. In addition, model 1 makes it possible to detect whether
differences in intolerance towards immigrants in countries are due to differences in country
composition (composition-effects).
Model 3: Model 3 consists of Model 2 and the county-level variables (number of immigrants, GDP,
GDP growth, unemployment, unemployment growth and cultural distance). This model makes it
possible to examine whether characteristics of a respondents’ country affect his or her level of
tolerance towards immigrants.
Model 4: Model 4 consists of Model 3 and perceived economic threat. By comparing model 3 to
model 4, it is possible to examine to what degree individual- and county-level characteristics are
mediated by perceived economic threat.
Model 5: Model 5 consists of Model 3 + perception of immigrants as a cultural threat. By comparing
model 3 to model 5, it is possible to examine to what degree individual- and county-level
characteristics are mediated by perceived economic threat. In addition, by comparing model 5 to
model 4, it is possible to examine whether individual- and country- level variables are mediated more
by economic threat or cultural threat.
Model 6: Model 6 consists of Model 3 + perception of immigrants as an economic threat and
perception of immigrants as a cultural threat. This model makes it possible to examine to what degree
the effects of the individual and county-level variables on tolerance towards immigrants are mediated
by the perception of immigrants as a threat, both cultural and economic.
29
Because there is a possibility of multicollinearity, correlation tables will be examined to determine
whether concepts are empirically distinguishable. In addition, the different country-level predictors of
perceived economic/cultural threat and tolerance towards immigrants were put individually in the
multilevel logistic regression analysis instead of solely in the blocks that are described by the six
models. Of course, descriptive measures were examined first in order to gain a better understanding of
the data. The results of our analyses are displayed in the next chapter.
30
4. Results
4.1.Descriptive Statistics
In table 1, descriptive measures of the dependent variable, the two mediating variables, and age are
displayed. For these variables, the mean, standard deviation and the lowest and the highest score are
presented. Most of the other individual-level variables that were employed in this study were
measured with a nominal scale and for these variables, statistics like mean, minimum, and maximum
bear little significance. Instead we examined the relative sizes of the categories of these variables.
These are displayed in Table 2. Descriptive measures of the country-level variables are displayed in
Table 3. Table 4 shows the descriptive measures per country.
Table 1. Descriptive Statistics of individual characteristics
Mean
Min.
Max.
Sd.
N
Tolerance towards immigrants
0.17
0
1
0.38
27077
Perceived economic threat
5.47
1
10
3.09
27077
Perceived cultural threat
4.73
1
10
3.06
27077
44.41
18
70
14.815
27077
Age
Source: European Values Study 2008
The dependent variable, intolerance towards immigrants, was measured dichotomous; respondents
either stated they do not like to have immigrants as neighbors or they did not mention immigrants as a
group of people they would not like to have as neighbors. Of the 27077 respondents, 4731 did not like
to have an immigrant as neighbor whereas 22346 respondents did not mind. In other words, 17.5 % of
the respondents could be considered intolerant and 82.5% could be considered tolerant towards
immigrants.
The mediating variables, perceived economic threat and perceived cultural threat were both measured
on a ten point likert scale. 15.2 % of the respondents scored a 10 on perceived economic threat,
indicating that they entirely agreed with the statement that immigrants take jobs away from natives in
their country. 14.1% of the respondents scored a 1 and thus did entirely not agree with the same
statement. 10.6% did entirely agree and 20.5% did entirely not agree with the statement whether a
country’s cultural life is undermined by immigrants. Overall, respondents perceived immigrants more
as a threat to their economic situation than to their culture. A paired t-test showed that the difference in
means was significant at p<0.01. Scores on perceived cultural threat had about the same spread as
scores on perceived economic threat.
Table 2. Descriptive statistics of individual
characteristics
N
%
Job status
27077
100
Household income
Low manual labor
4418
16.32
High manual labor
3559
13.1
4950
N
27077
%
100
Low household income
6002
22.2
Medium household income
8084
29.9
18.3
High household income
7113
26.3
7580
28.0
Missing
5878
21.7
Independent Labor
1570
5.8
Employment
27077
100
Missing
5000
18.5
Employed
16324
60.3
Urban environment
27077
100
Not unemployed
9011
33.3
Low urban environment
10661
39.4
Unemployed
1625
6.0
Middle Urban Environment
8711
32.2
Employed missing
117
0.4
High Urban Environment
3295
23.2
Sex respondent
27077
100
Missing
1410
5.2
Male
14833
54.8
Educational attainment
27077
100
Female
12244
45.2
Low educational attainment
7610
28.1
Religious affiliation
27077
100
Medium educational attainment
13086
48.3
8086
29.9
High educational attainment
6381
23.6
Christian religious affiliation 15073
55.7
3918
14.5
Low non-manual labor
High non-manual labor
12
No religious affiliation
Other religious affiliation
13
Source: European Values Study 2008
The individual level variables with a nominal or categorical scale are displayed in table 2. Important to
note is that on a few variables, a large number of missings were reported. For instance, 18% of the
respondents reported a missing on Job status and 22% on household income. An important question is
whether these are missings of a systematic nature. In order to test whether this is the case, dummy
variables for categories with a high amount of missings were included in the multilevel-analysis. The
results from the regression analysis indicated that the respondents with a missing on employment did
not significantly differ from employed respondents, although they did most closely resemble the
unemployed category in terms of tolerance towards immigrants. Respondents with a missing on job
12
A relatively high percentage of respondents performed ‘high non-manual’ labor. However, the large number of
systematic missings could have distorted the relative percentages. In addition, keep in mind that there is no
‘middle’ labor category; the so called ‘lower controllers’ from the EGP class typology have been put in the high
non-manual labor group as well
13
The high percentage of the ‘other religious affiliation’ category is largely due to inclusion of the orthodox
Christian faith, which is particularly popular in Easter European countries.
1
status did also not significantly differ from the reference category. However, respondents with a
missing on income were found to systematically report higher levels of intolerance towards
immigrants compared to respondents with a low income and, especially, compared to respondents with
a medium or high income. This should be taken into account when interpreting the findings of the
regression analysis.
Table 3. Descriptive Statistics of country-characteristics14
Mean
Minimum
Maximum
Sd.
N
Unemployment rate (in %)
6.21
3.1
11.3
1.85
28
Unemployment rate over 5
-2.08
-11.90
1.80
3.07
28
GDP
99.96
44
279
44.33
28
GDP growth 2004-2008 (in
15.91
3.61
35.10
9.02
28
2.68
.10
8.91
2.12
28
17.49
.20
61.90
13.32
28
years (in %)
% change compared to 2003)
Cultural distance
Foreign background (in %)
Source: European Values Study 2008 & Eurostat
In table 3, descriptive measures of country-characteristics are presented. For now, we will focus on the
mean scores and standard deviations, since data on the minimum and maximum scoring countries are
presented in more detail in table 4.
We can observe in table 3 that the average unemployment rate in the EU countries in 2008 was just
over 6%, with a spread of almost 2%. On average the unemployment rate declined in European
countries with 2 % between 2004 and 2008, with a standard deviation of just over 3%. The mean GDP
score presented in table 3 bears little significance; Since GDP scores were expressed with respect to an
EU-27 score of 100, the average EU score should also be +/- 100. However, what we can state is that,
based on the standard deviation, there are relatively large differences in GDP among European
Nations. By looking at the real growth rate between 2004 and 2008 we find that, economically
speaking, this was a good period for Europe. On average, European countries reported a GDP growth
of 16 %, with a spread of 9%. The mean cultural distance in EU countries was 2.68. In other words, an
average discrepancy of 2.68 was observed between the scores of the majority population and the
scores of the immigrant population on the cultural scale. The spread between countries on the cultural
distance was just over 2 %. Finally: the average amount of inhabitants with a foreign background in
the EU countries was 17.49%. Note that with a standard deviation of more than 13% there is quite a
lot of spread in foreign background figures between countries.
14
Country statistics were not weighted for the number of respondents in each country. Instead, each country has
the same weight.
2
Table 4a. Descriptive statistics per country
N
Intolerance Perceived economic Perceived cultural
Towards
threat
Threat
Immigrants
Austria
1107 0.24
6.08
6.04
Belgium
1047 0.06
5.31
5.16
Bulgaria
974 0.17
5.63
3.54
Cyprus
719 0.26
7.21
6.21
Czech republic
1316 0.31
5.62
4.74
Denmark
999 0.06
2.49
3.91
Estonia
768 0.34
5.68
5.07
Finland
958 0.17
4.21
3.37
France
978 0.04
4.08
4.45
Germany
1449 0.13
5.99
5.46
Greece
1041 0.15
6.28
5.01
Hungary
1295 0.15
7.07
4.39
Ireland
613 0.14
6.48
5.40
Italy
1116 0.15
4.68
4.31
Latvia
806 0.23
6.18
5.12
Lithuania
1057 0.29
5.97
4.71
Luxembourg
511 0.15
4.40
4.11
Malta
1076 0.34
6.97
7.41
Netherlands
1072 0.14
4.46
4.48
Poland
1228 0.17
4.81
3.67
Portugal
1065 0.08
5.77
3.99
Romania
1037 0.21
4.92
4.02
Slovak Republic 1000 0.16
5.80
3.88
Slovenia
919 0.29
5.57
5.00
Spain
999 0.05
5.17
4.32
Sweden
713 0.07
3.40
3.83
Great Britain
939 0.15
6.50
5.96
Northern Ireland 275 0.23
6.32
(Source: European Values Study 2008)
5.23
3
Austria
Belgium
Bulgaria
Cyprus
Czech Rp.
Denmark
Estonia
Finland
France
Germany
Greece
Hungary
Ireland
Italy
Latvia
Lithuania
Luxembourg
Malta
Netherlands
Poland
Portugal
Romania
Slovak Rp.
Slovenia
Spain
Sweden
Great Britain
Table 4b. Descriptive statistics per country
Foreign Unemployment Unemployment GDP
Backgr. Rate
rate growth
Growth
25.30
3.8
-1.10
11.66
22.90
7.0
-1.10
8.66
.30
5.6
-5.30
28.04
24.40
3.6
1.10
17.77
7.20
4.4
-3.90
24.53
8.90
3.3
-2.20
6.72
35.60
5.5
-4.20
24.12
6.21
6.4
-2.40
13.46
26.60
7.8
-1.50
6.64
21.90
7.5
-2.30
9.06
11.70
7.7
-2.80
10.94
3.10
7.8
1.70
9.14
25.40
6.3
1.80
13.15
11.60
6.7
-1.30
3.61
29.20
7.5
-2.90
29.67
8.00
5.8
-5.60
31.30
61.90
4.9
-.10
18.92
6.40
5.9
-1.50
15.86
23.50
3.1
-2.00
11.55
3.20
7.1
-11.90
23.50
11.80
7.7
1.00
4.66
.20
5.8
-2.30
28.24
2.90
9.5
-8.70
35.10
16.70
4.4
-1.90
21.68
20.20
11.3
.70
12.63
25.80
6.2
-1.20
10.52
24.40
5.6
.90
7.23
N. Ireland
24.40
5.6
.90
7.23
GDP
124
116
44
99
81
125
69
119
107
116
92
64
133
104
56
61
279
79
134
56
78
47
73
91
104
124
112
Cultural
distance
2.54
4.61
.23
2.15
1.48
2.68
1.71
1.23
2.05
2.35
2.72
1.44
1.59
1.49
.10
1.43
1.70
8.91
.15
3.19
5.06
1.71
3.13
3.17
7.83
3.70
5.86
112
.96
(source: European Values Study, 2008)
Table 4 provides an overview of the descriptive characteristics per country. Data shows that there are
large differences between European countries in terms of tolerance towards immigrants. High levels of
intolerance were generally found more in Eastern Europe and the Baltic states. The highest levels of
intolerance were reported in Malta, where respondents were found more than 8 times more likely to be
intolerant towards immigrants than in France, the country with the lowest level of intolerance. Other
low levels of intolerance were reported in Spain, Belgium and Denmark.
Respondents from Denmark reported by far the lowest level of perceived economic threat, while
Cyprus reported the highest levels. In terms of perceived cultural threat, Finland scored the lowest and
Malta the highest. Economic and cultural threat levels did not always go together. For example,
Hungary reported high levels of perceived economic threat, but relatively low levels of perceived
4
cultural threat. Denmark, on the other hand, reported a much higher level of cultural threat, than of
economic threat. On average however, the two types of ethnic threat are strongly correlated, as we will
see further down this chapter.
Of all the EU countries, Romania had the least amount of inhabitants with a foreign background, 0.2%
of the population. Romania is closely followed by Bulgaria, where 0.3% of the population had a
foreign background. Luxembourg is the EU country with by far the highest percentage of inhabitants
with a foreign background (62%).
In 2008, Spain was the country with the highest unemployment rate (11.3%) and the Netherlands with
the lowest (3.1%). Ireland was the country with the highest unemployment rate growth (an increase of
1.8%) between 2004 and 2008 and Poland the country with the biggest decline (11.90%). With almost
2.8 times the EU-27 average, Luxembourg had the highest GDP. Not surprisingly, low GDP numbers
were generally found in Eastern Europe and high GDP numbers in Western and Northern Europe. All
European countries experienced an average growth in GDP between 2004 and 2008, but there are big
differences between countries. High growth rates were found mainly in Easter Europe and the Baltic
states, especially the Slovak Republic, Lithuania and Latvia. The lowest GDP growth rate was found
in Italy.
Sweden, Denmark and the Netherlands were found to be the countries where the culture of the
majority can be considered the most liberal. Since the minority group was relatively liberal as well in
these countries, the cultural distance was only moderate. The biggest cultural distance was found in
Malta and Spain and the smallest in Latvia. Perhaps to some surprise, in a number of countries, like
Malta and Portugal, the culture of the minority can be considered more liberal than the culture of the
majority. However, one thing to keep in mind is that this research does not make a distinction between
immigrants from other EU countries and immigrants from outside the EU. One methodological aspect
to keep in mind as well is that for some countries, the cultural distance score of the immigrant group
was based on quite a small sample, which could have had implications for the representativeness of
the cultural distance score of immigrants in these countries.
4.2. Bivariate analyses.
A statistical phenomenon which may pose severe difficulties for regression analysis is
multicollinearity. Multicollinearity refers to the situation where two or more independent variables are
highly correlated with each other. The presence of multicollinearity can result in unstable coefficients
(Singleton & Straits, 2005). In addition, highly correlated predictors are difficult to separate from each
other. We speak of multicollinearity when the correlation between two variables is more than .7. In
5
order to test for multicollinearity, a bivariate analysis was conducted. In the bivariate analysis, the
overall associations between the study variables were examined. In other words, the bivariate analysis
data represents both the direct and indirect effects between two variables. Table 5 provides an
overview of the correlation between the most important study variables. Correlations on the other
research variables have been examined as well, but are not presented in this chapter.
Table 5. Pierson Correlation Matrix
Intolerance
Intolerance towards immigrants
towards
Perceived Cultural
Perceived Economic
immigrants
threat
threat
1.00**
.197**
.202**
Perceived Cultural threat
.197**
1.00**
.547**
Perceived Economic threat
.202**
.547**
1.00**
Cultural Distance
-.022**
.125**
.058**
GDP
-.075**
.028**
-.114**
GDP growth
.129**
-.046**
.055**
Unemployment rate growth
-.050**
.105**
.068**
-.100**
-.066**
.056**
-.027**
**= significant at the 0.01 level (two tailed)
.089**
-.020**
Unemployment rate
Percentage immigrants
Table 5 shows that, as expected, perceived cultural threat and perceived economic threat are strongly
correlated. However, their correlation value is within the acceptable limit. Other values show no signs
of multicollinearity. Some values in the matrix are somewhat surprising, as they go against the
hypothesized relations. Although the strength of the relation is rather weak, cultural distance is
negatively associated with intolerance. In line with predictions, cultural distance is strongly related to
perceived cultural threat. Also in line with the hypothesis GDP is negatively correlated with tolerance
towards immigrants and with perceived economic threat. GDP growth is negatively correlated with
intolerance and with perceived economic threat. Surprisingly, both the growth in unemployment rate
and the unemployment rate in 2008 were negatively related to intolerance. Finally, the percentage of
immigrants is negatively correlated with intolerance and with perceived economic threat, which again
goes against our predictions. In line with the predictions, the percentage of immigrants is correlated
with perceived cultural threat. The nature of these relations will be discussed in more detail in the
following paragraphs.
6
4.3. Multilevel Regression Analysis
In order to test the hypothesized relations, a multilevel logistic regression analysis was conducted. The
results of this analysis are displayed in table 5a and 5b. Table 5a contains the results of model 1 and
model 2; table 5b contains the results of model 3 to 6.
Table 5a. Multilevel regression analyses on intolerance
towards immigrants (n=27077).
Model 1
Model 2
Coef.
S.E.
Coef.
S.E.
-1.680 0.128 -1.534
0.159
Intercept
Individual
characteristics
Low education (Ref)
Middle education
-0.156** 0.044
Higher education
-0.329** 0.059
Low income (Ref)
Medium income
-0.088
0.048
High income
-0.109*
0.053
Missings on income
0.128*
0.051
Low manual labor (Ref)
High manual labor
0.049
0.059
Low non manual labor
-0.007
0.058
High non manual labor
-0.165*
0.058
Independent/self emp.
-0.025
0.079
Missing on job status
0.077
0.064
Employed (ref)
Unemployed
-0.141
0.075
Not unemployed
-0.123*
0.041
Missing on employment
-0.538
0.285
Low urban env. (ref)
Medium urban env
0.065
0.041
High urban env.
Missing on urban env.
Age
Male (ref cat)
female
No religious (ref cat)
Christian religion
Other religion
Country level variance
% explained
(compared to empty
model)
0.669
-
0.092
0.016
0.075
0.005**
-0.13**
-0.068
-0.247**
0.684
-2.2%
0.046
0.089
0.001
0.036
0.044
0.092
0.094
*=P< 0.05; **= P< 0.01
7
Table 5b. Multilevel regression analyses on intolerance towards immigrants (n= 27077).
Coef.
Model 3
S.E.
Model 4
Coef.
S.E.
Coef.
Model 5
S.E.
Coef.
Model 6
S.E.
Intercept
-0.063
0.496
-1.79**
0.418
-1.74**
0.456
-2.129**
0.420
Individual
characteristics
Low education (Ref)
-
-
-
-
-
-
-
-
Middle education
Higher education
-0.159**
-0.330**
0.044
0.059
-0.092*
-0.167**
0.045
0.060
-0.097*
-0.178**
0.045
0.060
-0.073
-0.122*
-0.045
0.060
Low income (Ref)
Medium income
-0.088
0.048
-0.074
0.049
-0.084
0.049
-0.078
0.049
High income
-0.109*
0.053
-0.053
0.053
-0.085
0.053
-0.057
0.054
Missing on income
Low manual labor(Ref)
0.128*
-
0.051
-
0.123*
-
0.052
-
-0.085*
-
0.049
-
0.108*
-
0.053
-
High manual labor
Low non manual labor
High non manual labor
Missing on job status
Independent/self emp.
0.050
-0.004
-0.164*
0.076
-0.025
0.059
0.058
0.058
0.064
-0.079
0.054
0.043
-0.071
0.136
0.039
0.060
0.059
0.059
0.08
0.080
0.052
0.024
-0.108
0.109
-0.017
0.060
0.059
0.059
0.065
0.080
0.056
0.047
-0.064
0.135
0.024
0.060
0.060
0.059
0.065
0.081
Employed (ref)
Unemployed
-0.138
0.075
-0.196**
-0.77
-0.166*
0.077
-0.196*
0.077
Not unemployed
Missing on employ
Low urban env. (ref)
Medium urban env.
-0.129**
-0.538
0.067
0.041
0.285
0.041
-0.117**
-0.609
0.071
0.042
0.289
0.042
-0.12**
-0.629
0.064
0.042
0.289
0.042
-0.116**
-0.650
0.067
0.042
0.291
0.042
High urban env.
0.015
0.046
0.043
0.046
0.038
0.046
0.048
0.047
Missing on urban env.
0.075
0.089
0.125
0.089
0.038
0.046
0.116
0.090
Age
Male (ref cat)
0.005**
-
0.001
-
0.005**
-
0.001
-
0.005**
-
0.001
-
0.005**
-
0.001
-
female
No religious (ref cat)
-0.127**
-
0.036
-
-0.135**
-
0.036
-
-0.114**
-
0.036
-
-0.124**
-
0.036
-
Christian religion
Other religion
-0.063
-0.266**
0.044
0.090
-0.062
-0.228*
0.045
0.090
-0.081
-0.215*
0.045
0.009
-0.072
-0.206*
0.045
0.091
Country
characteristics
Foreign background
GDP
0.04
-0.005
0.001
0.003
0.002
-0.003
0.008
0.003
-0.001
-0.003
0.009
0.003
-0.001
-0.002
0.008
0.003
GDP growth
0.048**
0.013
0.041**
0.011
0.048**
0.012
0.043**
0.011
Unemployment
Unemployment growth
-0.191**
0.041
0.050
0.038
-0.199**
0.008
0.041
0.031
-0.161**
0.027
0.045
0.035
-0.178**
0.011
0.041
0.031
Cultural distance
0.006
0.043
0.001
0.036
-0.022
0.039
-0.016
0.036
0.166**
0.006
0.113**
0.007
0.100**
0.007
Intermediating
variables
Perceived economic
threat
Perceived cultural
Threat
0.156**
0.006
8
Country level
variance
% explained (compared
to empty model
0.412
38.4%
0.059
0.336
49.8%
0.050
0.372
0.055
0.337
44.4%
0.050
49.6%
*=P< 0.05; **= P< 0.01
By examining the country level variance component of model 1(presented in the bottom row of table
5) we can observe whether individuals within a country resemble each other more closely. The
country-level variance in the first model was 0.669. Since the variance at the individual level is
fixed15, we can calculate that 83% of the observed variance in intolerance towards immigrants can be
explained at the individual-level and almost 17 % of the variance is at the country-level. Comparing
the variance components of the other models with the variance component of the empty model
indicates whether the inclusion of new country-level variables improves our ability to explain cross
national variations in tolerance levels. We see that the country-level variables in model 3 explain over
38% of the country-level variance in intolerance towards immigrants. The inclusion of perceived
economic threat in model 4 increases the percentage of variance explained to almost 50% whereas the
inclusion of perceived cultural threat on top of all other country-level variables explains about 44%.
This implies that economic threat is a slightly better predictor of intolerance than cultural threat.
4.3.1.
The effect of individual-level variables on intolerance towards immigrants.
We hypothesized that members of the majority group with low educational levels are more likely to be
intolerant towards immigrants and that this effect is mediated both through the perception of
immigrants as an economic treat and a cultural threat. In line with the hypothesis, model 2 shows that
members of the majority group with low education levels indeed showed significantly higher levels of
intolerance towards immigrants than both majority members with a medium educational level [coef. 0.159; p<0.01] and majority members with a high educational level [coef. -0.330; p<0.01]. A
comparison of model 4 and 5 suggest that effect of education on tolerance is slightly more mediated
by economic threat than by cultural threat. The impact of a medium education on intolerance,
compared to a low education, was reduced to 58% by economic threat and to 61% by cultural threat.
The impact of a high education on intolerance, compared to a low education, was reduced to 51% by
economic threat and to 54% by cultural threat. If we examine model 6 we can observe that the
respondents with a low education and with a medium education did not significantly differ when
controlled for both types of ethnic threat. The impact of a medium education on intolerance, compared
to a low education, was reduced to 46% of its former strength. The difference between respondents
with a low education and with a high education is still significant however [-0.112, p<0.05]. This
indicating that, besides a mediated effect, there is a moderately strong independent effect of education
on intolerance as well.
15
The individual level variance is equal to π^2/3
9
The next hypothesized relation was that members of the majority group with low income levels are
more likely to be intolerant towards immigrants and that this effect is mediated through the perception
of immigrants as an economic treat. Model 2 shows that members of the majority group with a high
income indeed show significantly lower levels of intolerance towards immigrants [-0.109, p<0.05].
Members of the majority group with a medium sized income showed less intolerance as well, though
this difference is not significant. A comparison of model 4 and 5 suggests that effect of income on
intolerance is mediated stronger by economic threat than by cultural threat. The impact of a medium
income on intolerance was reduced to 0.84% when controlled for economic threat. The impact of a
medium income on intolerance, compared to a low income, was reduced to 95% when controlled for
cultural threat, suggesting that the mediating effect of cultural threat is minimal. The effect of a high
income, compared to a low income, on tolerance was even stronger mediated by the two types of
ethnic threat. Economic threat reduced this effect to 49% and cultural threat reduced it to 78% of its
former strength. In model 4, 5 and 6 effect of income on intolerance is no longer significant.
The next prediction was that members of the majority group that perform manual labor or perform low
skilled labor are more likely to be intolerant towards immigrants and that this effect is mediated
through the perception of immigrants as an economic treat. Model 2 shows that there was little
difference between respondents who performed low manual labor with respondents who performed
low-non manual labor, high manual labor or who worked independently. However, in line with our
hypothesis, there was a significant difference between respondents who performed low manual labor
and respondents who performed high non-manual labor [-0.164, p<0.05]. Comparing model 2 with
model 3 shows that if we control for perceived economic threat, the impact is reduced to 43%. By
comparing model 2 with model 4 we can observe that, if we control for perceived cultural threat, the
impact is only reduced to 66%. This suggests, in line with our hypothesis, that the effect of high nonmanual labor on intolerance is stronger mediated by perceived economic threat than by perceived
cultural threat. Finally, if we compare model 2 with model 6 we can see that controlling for both
mediating variables even results leads to a reduced strength of %39. In other words, the effect of job
status on intolerance towards immigrants is strongly mediated by perceived ethnic threat.
The next hypothesis stated that members of the majority group who are unemployed are more likely to
be intolerant towards immigrants and that this effect is mediated through the perception of immigrants
as an economic treat. Surprisingly, model 2 shows that unemployed respondents actually showed
lower levels of intolerance towards immigrants than employed respondents, though the difference was
not significant. However, with the inclusion of cultural and economic threat, the relationship actually
became significant, indicating a suppressor effect. The inclusion of economic threat in the model led
to a strength 142% and the inclusion of cultural threat led to a strength of 120%. Respondents form
the not unemployed category (students, housewives etc.) showed significantly lower levels of
intolerance [-0.123*, p<0.05], which is according the predictions.
10
The last predictor at the individual-level was living in an urban area. Hypothesized was that members
of the majority group who live in urban areas are more likely to be intolerant towards immigrants and
that this effect is mediated through the perception of immigrants as an economic and a cultural threat.
Model 1 shows that respondents living in large cities reported almost an identical level of tolerance as
respondents living in a low urban environment. Respondents living in a medium level urban
environment did show higher levels of intolerance then the other categories, but again no significant
effects were found. Our hypothesis on the impact of urban area on intolerance is not confirmed. Even
though the hypothesis is rejected, we can still examine the mediating effects of perceived threat.
Model 3 shows that when controlled for economic threat, the effect of medium urban environment on
intolerance towards immigrants grew to 106%. When controlled for cultural threat, the effect of
medium urban environment on intolerance towards immigrants decreased to 96%. When controlled for
both forms of threat, the strength of medium urban environment on intolerance stayed exactly the
same (100%). In other words, the effect of both perceived cultural and economic threat is very
minimal.
As for the control variables, age was shown to be significantly related to intolerance. Older
respondents were more likely to be intolerant towards immigrants [0.005; p<0.01]. Gender was found
to be significantly related to intolerance as well [coef. -0.127; p<0.010]. Females generally had fewer
problems with an immigrant as neighbor. No significant difference was found between Christians and
respondents without a religious affiliation. However, respondents with a religious affiliation other than
Protestant or Catholic were found to be generally more tolerant [-0.266; p<0.01].
4.3.2.
The effect of country-level variables on intolerance towards immigrants.
The effect of the country-level predictors can be observed in model 3. The first country-level
hypothesis was that the larger the total immigrant group in a country, the more likely that members of
the majority group are intolerant towards immigrants and that this effect is mediated both through the
perception of immigrants as an economic threat. In addition, we argued that this effect is mediated by
cultural threat as well. Our data did not show this relation. The size of the immigrant group mattered
surprisingly little for people’s willingness to live near an immigrant neighbor. Model 4 showed that,
when controlled for economic threat, the strength of the relationship diminished to 44%.16, but since
the strength of the effect is so small, calculating the change in effect when perceived cultural and
economic threat were controlled for in terms of percentages leads to somewhat arbitrary outcomes. In
terms of absolute numbers, the change in effect was very small. When controlled for cultural threat,
16
Based on unrounded figures
11
the effect of the size of the immigrant group on intolerance actually became negative. The change in
effect was -27% but again, if we compared to coefficients in absolute terms, the difference was very
small.
The second country-level hypothesis stated that the higher the unemployment rate in a country, the
more likely that members of the majority group are intolerant towards immigrants and that this effect
is mediated through the perception of immigrants as an economic threat. Model 3 shows that,
surprisingly, a higher unemployment rate is actually significantly related to lower levels of intolerance
[-0.266; p<0.01]. We also hypothesized that the effect of unemployment on intolerance would be
mediated by perceived economic threat rather than perceived cultural threat. A comparison of model 4
and 5 shows that the mediating effect of cultural threat is actually stronger than the mediating effect of
economic threat. A comparison of model 3 and 4 shows that the impact of unemployment on tolerance
was slightly increased to 105% when controlled for economic threat, thus showing a slight compressor
effect. When controlled for cultural threat, the impact of unemployment on tolerance was reduced to
84% of its former strength. Based on these observations we can state that the effect of the
unemployment rate on intolerance is mediated through cultural threat rather than through economic
threat.
The third country-level hypothesis stated that the stronger the growth of the unemployment rate in a
country, the more likely that members of the majority group are intolerant towards immigrants. A
higher unemployment rate growth was positively associated with intolerance, yet the effect was not
significant. We hypothesized that the effect of the unemployment rate in a country on intolerance
would be mediated by perceived economic threat rather than perceived cultural threat. A comparison
between model 4 and 5 show this is indeed the case. Compared to model 3, the strength of the
coefficient of unemployment rate growth in model 4 is reduced to only 20%. The strength of the
coefficient of unemployment rate growth in model 5 was reduced to 66%. Again, one should be
careful when interpreting these results. Since the effect of unemployment growth on intolerance was
quite small, a small change in strength results a large change in percentage.
The fourth country-level hypothesis stated that the lower the GDP in a country, the more likely that the
that members of the majority group are intolerant towards immigrants. A high GDP was indeed related
to lower levels of intolerance, though again, the effect was not significant. We also hypothesized that
the effect of a high GDP on intolerance would be mediated by perceived economic threat rather than
perceived cultural threat. Model 4 and 5 show that the effect was mediated by perceived cultural threat
as well, but slightly more by perceived economic threat. The impact of GDP on intolerance in model 4
was reduced to only 59% of its strength in model 3. In model 4, its strength was 65%. When we
controlled for both types of perceived threat, the strength was further reduced to only 50%
12
The fifth hypothesis stated that the smaller the growth of the GDP in a country, the more likely that
members of the majority group are intolerant towards immigrants and that this effect is mediated
through the perception of immigrants as an economic threat. Again surprisingly, a larger growth in
GDP actually seemed to be significantly related to higher levels of intolerance. We also hypothesized
that the effect of a high growth in GDP on intolerance would be mediated by perceived economic
threat rather than perceived cultural threat. Model 4 and 5 show that the effect was indeed mediated by
perceived economic threat and was almost not at all mediated by perceived cultural threat. The impact
of GDP growth on intolerance in model 4 was reduced to only 85% of its strength in model 3. In
model 4, its strength was still 99%. When we controlled for both types of perceived threat at the same
time, the strength was reduced to 90%.
The sixth and final country-level hypothesis stated that the larger the cultural distance between the
total immigrant group and the majority group, the more likely that members of the majority group are
intolerant towards immigrants and that this effect is mediated through the perception of immigrants as
a cultural threat. As we can observe in model 3, cultural threat was found to have little effect on
intolerance. Calculating the change in effect of cultural distance on tolerance when perceived cultural
and economic threat were controlled for in terms of percentages leads to somewhat arbitrary
outcomes. Although the strength of the effect on cultural distance on tolerance dropped to only 11% of
its former strength when controlled for economic threat in absolute numbers only a very small change
was found. A similar result could be found when controlling for cultural threat and for both types of
ethnic threat.
13
5. Conclusion and Discussion
5.1.Conclusion
Though research on anti-immigrant attitudes and its determinants has become quite popular over the
last decades, not every aspect of the subject has been addressed sufficiently in the scientific literature.
The present study aimed to fill this gap in the literature by empirically examining the following
research question:
‘To what extent are the effects of individual-level and country-level characteristics on an individual’s
level of tolerance towards immigrants mediated by his/her perception of immigrants as an economic
and a cultural threat?’
Based on the results we can state that, at the individual level, the variables that significantly affected
levels of tolerance towards immigrants seemed to be mediated to a higher degree by perceived
economic threat than by perceived cultural threat. Low income and low job status levels significantly
increased intolerance towards immigrants and these effects were for a large part mediated through
economic threat. Higher educated respondents scored significantly lower on intolerance as well and
this effect was mediated strongly by both perceived cultural and economic threat. At the country level,
only two predictors were found to significantly affect intolerance towards immigrants. Of those two,
GDP growth was only mediated slightly by economic threat and was not mediated by cultural threat.
Unemployment was mediated to a moderate degree by cultural threat.
Unfortunately, it is difficult to draw stronger conclusions from our study. The problem is that several
hypotheses were contradicted by our data. We predicted that respondents who live in an urban area are
more likely to be intolerant towards immigrants than respondents from non-urban areas.
The
argumentation behind this hypothesis, based on Realistic Group Conflict theory and Symbolic
Interaction Theory, was that respondents living in urban areas perceive higher levels of economic and
cultural threat due to a larger immigrant population in these areas. Therefore, respondents from urban
areas are likely to hold more negative attitudes towards immigrants. However, the results showed that,
respondents from moderate urban areas actually were more likely to be intolerant that respondents
from high urban areas, although this effect was not significant. A theory that may help to explain this
unexpected finding is Contact Theory (Allport, 1954). Contact theory states that intergroup contact is
an effective way to reduce prejudice between groups. In urban environments, there are generally more
immigrants and therefore there is more contact between the majority and the immigrant group.
According to Contact Theory, this increase in contact reduces levels of intolerance. The effect of
intergroup contact may have counteracted the effect of ethnic competition in urban environments,
leading to insignificant results. Another explanation can be found in a possible selection effect.
14
Intolerance towards immigrants was measured by asking respondents whether they would mind to
have an immigrant as neighbor. People are well aware that there are greater immigrant numbers in
large cities and those who do not want to live near immigrants may very well be less likely to move to
large cities than people who are more tolerant.
An unexpected finding at the individual level was that unemployed respondents were actually less
likely to be intolerant towards immigrants than employed respondents. This observation contradicts
with Realistic Group Conflict Theory and is quite difficult to explain. An explanation could be that the
unemployed empathize more with the immigrant group because they are in a disadvantaged position
themselves. It is also possible that the unemployed generally have more immigrants among their social
network and thus that the explanation can be sought in Contact Theory as well. However, this does not
explain why the observation that the unemployed are less intolerant towards immigrants contradicts
prior research (Coenders and Scheepers, 2003; Savelkoul et al., 2008).
At the country level we also found that the percentage of people with a foreign background in a
country mattered surprisingly little for people’s willingness to live near an immigrant neighbor. This
finding contradicts Group Conflict Theory, but is line with our findings on the effect of living in an
urban area. Again, Contact theory may help to explain this observation as well. Schneider (2008), for
example, argues that Realistic Group Conflict Theory is too narrow to explain cross national
differences in perceived ethnic threat. Contact theory should be taken into account, since out group
size increases competition but also contact opportunities.
The other findings at the country-level are more difficult to explain by means of a theory. An
explanation for why the results contradict the prominent anti-immigrant attitude theories is that
theoretical relations may have not been adequately tested yet in Eastern European contexts. Most
European cross-country studies have been conducted in Western Europe; while Eastern Europe has
largely been ignored. For instance, studies like the ones of Coenders et al. (2003), Scheepers et al.
(2002) and Lucassen and Lubbers (2011.), have all been conducted in Western Europe. It is possible
that theories are adequate predictors in a Western European context, but that their explanations fall
short in an Eastern European context.
Another possibility is that these unexpected findings are the result of methodological issues. In the
discussion section, a number of methodological limitations are presented. It may very well be that our
one or more of these limitations that possible have affected the internal validity of this study. Future
research may find out whether there is an aspect of truth to some of the unexpected findings.
15
5.2. Discussion
By examining which individual-level and country-level characteristics affect people’s tolerance
towards immigrants and by examining the mediating role of perceived levels of economic and cultural
threat in these relations, the current research contributed to the scientific literature. Unlike other crosscountry studies on anti-immigrant attitudes, this research included perceived threat as a mediating
variable and made the distinction between the economic and the cultural component of ethnic threat,
which is important in order to assess how certain individual and country-level characteristics could
lead to negative immigrant perceptions. In contrast to other studies as well, intolerance towards
immigrants was operationalized in a way that challenges people’s living environment. As such, this
study presents unique data on a complex subject.
Despite these merits, this study is not without limitations. First, there were difficulties with regard to
the operationalization of some key concepts. The first one was that perceived economic threat was
measured with a single item. Respondents were asked whether they agreed with the statement that
immigrants take jobs away from natives in a country. Though an important part, the economy of
country encompasses more than just the labor market. It may very well be possible that some
respondents do not perceive immigrants as a threat to the labor market but, for example, do perceive
immigrants as tax burden or perceive them as threat with regard to other aspects of the economy.
Perceived cultural threat was measured with a single item as well, which posed similar problems.
While the EVS granted the opportunity to employ a unique operationalization of the concept
‘intolerance towards immigrants’, a difficulty was that intolerance was measured with a single
dichotomous item. As a consequence, there was little room for nuance in the respondents’ answers.
Possibly, there was too little variance to adequately detect the effect of predictors. It is recommended
that future studies, if possible, adopt more elaborate measures of their key variables.
A second limitation is related to the cultural distance scale. Country scores on cultural distance
indicated the discrepancy between the average majority group scores and the average immigrant group
scores of a country. However, the number of respondents on which the immigrant group scores was
based was for some countries less than desirable. As a result, the immigrant group scores on cultural
distance were somewhat arbitrary for these countries. This might have affected the results of the
multilevel regression analysis with regard to the cultural distance predictor.
A third limitation of this research is that, as with all cross sectional research designs, it is difficult to
establish causality. For example, the correlation between urban environment and intolerance towards
immigrants may reflect the effect of intolerance towards immigrants on living in an urban environment
instead of the other way around. Compared to a cross sectional research designs, longitudinal research
designs are better able to control for causality, although even in longitudinal research the complex
nature of attitudes and behavior make it difficult to assess causal relationships. Still, longitudinal
16
research may provide us a better understanding of the development of anti-immigrant attitudes. Since
very little longitudinal cross-country research has been conducted on this subject, additional
longitudinal studies would be a welcome contribution to the literature.
A fourth limitation is related to a possible response bias. Participants in the EVS were interviewed
instead of having to fill out the standardized questionnaire themselves. While this has some
advantages, it has some downsides as well. Research on anti-immigrant attitudes is rather susceptible
to response biases. It is possible that some respondents had the tendency to answer in a manner that
they thought was viewed favorably by the interviewer. A problematic aspect of this phenomenon is
that some individuals and groups display this tendency more than others. Philips (1971) suggests that
the consistent findings of lower racial prejudice among the middle class compared with lower class
respondents may not reflect true class differences but rather a greater concern among the middle class
to give desirable answers. Since we examined the effect of, among others, education, job status, and
income on intolerance towards immigrants, it is possible that a response bias affected the research
findings.
17
6. References
Adida, C. L., Laitin, D. D., & Valfort, M. A. (2010). Identifying barriers to Muslim integration in
France. Proceedings of the National Academy of Sciences, 107(52), 1-7.
Adida, C. L., Laitin, D. D., & Valfort, M. A. (2011), 'One Muslim is Enough!' - Evidence from a Field
Experiment in France. IZA Discussion Paper No. 6122.
Boomgaarden, H. G., & Vliegenthart, R. (2007). Explaining the Rise of Anti-immigrant Parties: The
Role of News Media Content. Electoral Studies, 26(2), 404–17.
Borjas, G. (2003). The labor demand curve is downward sloping: Re-examining the impact of
immigration on the labor market. Quarterly Journal of Economics, 118, 1335–1374.
Campbell, D.T. (1965). Ethnocentric and other altruistic motives. Lincoln, NE: University of
Nebraska Pres.
Coenders, M. (2001). Naturalistic attitudes and ethnic exclusionism in a comparative perspective: an
empirical study of attitudes toward the country and ethnic immigrants in 22 Countries. (Doctoral
dissertation). Nijmegen: KUN.
Coenders, M., & Scheepers, P. (2003). The effect of education on nationalism and ethnic
exclusionism: an international comparison. Political psychology 24 (2), 313-343.
Esses, V. M., Jackson, L. M., & Armstrong, T. L. (1998). Intergroup competition and attitudes toward
immigrants and immigration: An instrumental model of group conflict. Journal of Social Issues,
54, 699–724.
Esses, V. M., Dovidio, J. F., Jackson, L. M., & Armstrong. T. L. ( 2001). The Immigrants Dilemma:
The Role of Perceived Group Competition, Ethnic Prejudice, and Nationality Identity. Journal of
Social Issues, 53, 389–412.
Gabennesch, H. (1972). Authoritarianism as world view. American Journal of Sociology, 77, 857–875.
Hello, E. (2003). Educational attainment and ethnic attitudes: how to explain their relationship.
(Doctoral dissertation). Radboud University Nijmegen.
18
Hernes, G., & Knudsen, K. (1992). Norwegians’ attitudes toward new immigrants, Acta Sociologica,
35, 123-139.
Hollifield, J. F. (1997). Immigration and Integration in Western Europe: A Comparative Analysis. In
Ucarer, E. M., & Puchala, D. J. (Eds.), Immigration into Western Societies: Problems and Policies,
(pp. 28-69). London: Pinter publications.
Knoll, B. R. (2009). And who is my neighbor? Religion and immigration policy attitudes. Journal for
the Scientific Study of Religion, 48, 313-331.
LeVine, R., & Campbell, D. T. (I 972). Ethnocentrism: Theories of conflict, ethnic attitudes and group
behavior. New York: Wiley.
Lubbers, M., & Güveli, A. (2007). Voting LPF: Stratification and the Varying Importance of
Attitudes, Journal of Elections, Public Opinion and Parties 17(1), 21–48.
Lucassen, G., & Lubbers, M. (2012). Who Fears What? Explaining Far-right-wing Preference in
Europe by Distinguishing Perceived Cultural and Economic Ethnic Threats. Comparative Political
Studies, 45(5), 547-574.
Manevska, K., & Achterberg, P. (2011). Immigration and Perceived Ethnic Threat: Cultural Capital
and Economic Explanations. European Sociological Review.
Martire, G. & Clark, R., (1982). Anti-semitism in the United states. New York: Preager.
Mayda, A. M. (2006). Who is Against Immigration? A Cross-Country Investigation of Individual
Attitudes toward Immigrants. Review of Economics and Statistics, 88(3), 510-530.
McDaniel, E. L., Nooruddin, I., & Shortle., A. F. (2011). Divine Boundaries: How Religion Shapes
Citizens’ Attitudes Towards Immigrants. American Politics Research, 39(1): 205-233.
McLaren, L.M. (2003). Anti-Immigrant Prejudice in Europe: Contact, Threat Perception, and
Preferences for the Exclusion of Immigrants. Social Forces, 81(3), 909–36.
19
Meuleman, B., Davidov, E., & Billiet, J. (2009). Changing attitudes toward immigration in
Europe, 2002–2007: a dynamic group conflict theory approach. Social Science Research, 38,
352–65.
Nielsen, F. (1985). Toward a theory of ethnic solidarity in modern societies. American Sociological
Review, 50, 133–49.
Olson, J. M., & Zana, M.P. (1993). Attitudes and attitude change. Annual Review of Psychology, 44,
117-154.
Olzak, S. (1982). Ethnic mobilization in Quebec. Ethnic and Racial Studies, 5, 253–275.
Pedersen, A., & Walker, I. (1997). Prejudice against Australian Aborigines: Old-fashioned and
modern forms. European Journal of Social Psychology, 27, 561-587.
Pettigrew, T. F. (1998). Reactions toward the new minorities of Western Europe. Annual review of
Sociology, 24, 77–103
Philips, D.L. (1971). Knowledge from what? Chicago: Rand McNally
Quillian, L. (1995). Prejudice as a response to perceived group threat: population composition and
anti-Immigrant and racial prejudice in Europe. American Sociological Review 60: 586-611.
Savelkoul, M., Scheepers, P., Tolsma, J. and Hagendoorn, L. (2010). Anti-Muslim attitudes in the
Netherlands: tests of contradictory hypotheses derived from ethnic competition and intergroup contact
theory. European Sociological Review,
Scheepers, P., Gijsberts, M., & Coenders, M. (2002). Ethnic exclusionism in European countries.
Public opposition to civil rights for legal migrants as a response to perceived ethnic threat. European
sociological review, 18(1), 17-34.
Schneider, S. L. (2008). Anti-immigrant attitudes in Europe: outgroup size and perceived ethnic threat.
European Sociological Review, 24, 53–67.
Schuman, H. (1997). Racial Attitudes in America: Trends and Interpretations. Cambridge, MA:
Harvard University Press.
20
Semyonov, M., Raijman, R., & Gorodzeisky, A. (2006). The rise of anti-foreigner sentiment in
European societies, 1988-2000. American Sociological Review, 71, 426-449.
Semyonov, M., Raijman, R., Gorodzeisky, A. (2008). Foreigners’ impact on European societies:
public views and perceptions in a cross-national comparative perspective. International Journal of
Comparative Sociology, 49, 5–29.
Singleton, Royce, and Bruce Straits. 2005. Approaches to Social Research. 4th ed. New York: Oxford
University Press.
Smith, T. W., & Dempsey, G.R. (1983). The Polls: Ethnic Social Distance and Prejudice. Public
Opinion Quarterly, 47, 584-600.
Sniderman, P. M., Hagendoorn, L., & Prior, M. (2004). Predisposing factors and situational triggers:
Exclusionary reactions to immigrant minorities. American Political Science Review, 98, 35–
49.
Tajfel, H., & Turner, J. (1979). An Integrative Theory of Intergroup Conflict. in Austin, W. G., & and
Worchel, S., (Eds.), The Social Psychology of Intergroup Relations. Monterey: Wadsworth.
Tajfel, H. (1982). Social psychology of intergroup relations. In Rosenzweig. M. R. & Porter, L. W.,
(eds.), Annual Review of Psychology, 33, 1-39. Palo Alto: Annual Reviews.
Tajfel, H., & Turner, J., (1985). The social identity theory of intergroup behavior. In Worchel, S., &
Austin, W.G., (eds.), Psychology of Intergroup Relations, 2, 7-24. Chicago: Nelson-Hall.
Tolsma, J., Lubbers, M. & Coenders, M., (2007). Ethnic competition and opposition to ethnic
intermarriage in the Netherlands: a multi-levelapproach. European Sociological Review, 24, 215–230.
Verberk, G., Scheepers, P. & Felling, A. (2002). Attitudes and behavioural intentions towards
ethnic minorities: an empirical test of several theoretical explanations for the Dutch case. Journal of
Ethnic and Migration Studies, 28(2): 197-219.
Weldon, S. A. (2006). The Institutional Context of Tolerance for Ethnic Minorities: A Comparative,
Multilevel Analysis of Western Europe. American Journal of Political Science, 50(2), 331–349.
21
Zick, A., Pettigrew, T. F., & Wagner, U. (2008). Ethnic Prejudice and Discrimination in Europe.
Journal of Social Issues, 64(2), 233-251.
Zick, A., Wagner, U., Van Dick, R., & Petzel, T. (2001). Acculturation and prejudice in Germany:
Majority and minority perspectives. Journal of Social Issues, 57, 541–557.
22