Right wing voting behaviour among native
Germans:
An analysis of group threat theory and
intergroup contact theory
By: Rik Hogeweg
Student Number: 12265500 Master: Sociology
Track: Migration and Ethnic Studies
First Reader/Supervisor: Dr. Sonja Fransen Second Reader: Prof. Dr. Willem-Jan Duyvendak Date of Submission: 08-07-2019
Summary
In the 21st century, Europe’s political domain saw a rise in anti-immigration parties. These parties use immigration as one of their core political standpoints to gain more electoral success and are often situated on the right side of the political spectrum.
Germany is among the countries where a right wing political party gained electoral success by adopting anti-immigration standpoints. The Alternative für Deutschland achieved impressive electoral results in the last national elections, mainly by focussing on immigration during the electoral campaigns.
Immigration was a hot topic in the national elections of 2017 due to a large scale immigration wave reaching into Germany, where 1.4 million immigrants entered Germany in 2014–2016. However, this inflow did not lead to right wing parties’ electoral success in all parts of Germany. In the federal state of Sachsen-Anhalt, right wing parties achieved more electoral success when more immigrants entered the state, but in Bremen, right wing parties showed little or no electoral success, even though this federal state saw a similar increase in immigrants.
To explain when an increase in the inflow of immigrants leads to increased electoral success for right wing parties, this study used data from the European Social Survey, the Bundeswahlleiter and the Bundesamt Germany. Additionally, it builds on previous research that found that more negative attitudes toward immigrants result in more right-wing voting behaviour. However, other research found that when there is more intergroup contact, there are more positive attitudes toward immigrants, which results in less right-wing voting behaviour. Using both theories from previous research, group threat theory and intergroup contact theory, this study attempts to explain when an increase in the outgroup size leads to either more perceived group threat or intergroup contact that leads to more or less anti-immigrant sentiment results in more or less right wing voting behaviour. The explanations found are the unemployment rate, Germany’s economic situation, social trust, whether native Germans would allow Muslims to come to Germany and Germans’ nativist feelings.
The findings show that increases in the number of immigrants in Germany influence both the perceived group threat by Germans and the amount of intergroup contact for Germans. Additionally, the number of immigrants in Germany does not affect Germans’ anti-immigrant sentiments, while perceived group threat and intergroup contact do. The latter effect shows a possible indirect effect of the outgroup size on anti-immigrant sentiment. This study’s results confirm this indirect effect for both perceived group threat and intergroup contact. When testing for the indirect effect of an increase in outgroup size, through perceived group threat or intergroup contact, on anti-immigrant sentiment and ultimately voting behaviour, this study found that increases in the number of immigrants in Germany do not indirectly influence Germans voting behaviour. The influences that this study thought to explain when an increase in the outgroup size in Germany ultimately increases the right-wing voting behaviour of Germans were also not significant.
In conclusion, the outgroup size in Germany matters in general, but not for voting behaviour. The explanations examined in this study cannot explain when an increase in the number of immigrants in Germany influences German voting behaviour. This study did show that anti-immigrant sentiment can change through the feeling of threat or the amount of contact with immigrants. Perceived group threat was found to negatively impact
anti-The explanations for when an increase in the outgroup size in Germany leads to different amounts of anti-immigrant sentiment or right wing voting behaviour are not obvious. Common theories that lead to more anti-immigrant sentiment and right-wing voting behaviour, like increased labour market competition, were not supported. Less common theories that lessen anti-immigrant sentiment and right wing voting behaviour, like social trust, were also not supported. Therefore, anti-immigrant sentiment and voting behaviour are linked to many complex explanations.
Index
1 Introduction ... 5
2 The German case ... 7
3 Determining changes in anti-immigrant sentiment... 8
3.1 Economic explanations ... 9
3.1.1 Labour market competition model... 9
3.1.2 Fiscal burden model ... 9
3.2 Group threat theory ... 9
3.3 Intergroup contact theory ... 10
3.4 Non-economic explanations ... 12 3.4.1 Sociopsychological explanations ... 12 3.5 Alternative explanations ... 13 4 Methods ... 16 4.1 Dependent variables ... 17 4.2 Independent variables ... 18 4.3 Control variables ... 19 5 Results ... 20
5.1 Perceived group threat and intergroup contact theory ... 20
5.2 When the outgroup size has an influence ... 27
6 Conclusion... 33
7 Bibliography ... 36
1 Introduction
Since the start of the 20th century, anti-immigration parties have grown in popularity in Europe (Meuleman, Davidov & Billiet 2009), including in Germany, Switzerland, Belgium, France and Austria (Garcia-Faroldi 2017, Mughan & Paxton 2006). These parties often use immigration as a factor in their political agenda and electoral success. Garcia-Faroldi (2017) finds that political parties do not only focus on immigrants from European countries that have not yet arrived, but also on immigrants that have already settled in their country.
An example is the Alternative für Deutschland (AfD) in Germany. In 2017, the AfD, as a new party, rose into parliament as the third biggest party in Germany (Mudde, 2017). This right wing party gained 13.3% of the votes in their first national elections. According to Mudde (2017), the main reason for this stunning result is that immigration was a prominent issue during the most recent national elections in Germany.
Immigration has become a prominent issue in elections in Germany due to the significant increase in immigrants. Germany saw an increase of 600.000 asylum applications in 2014–2016 (‘Migration to Europe in charts’, 2018). In these three years, just under 1.4 million immigrants applied for asylum in Germany. This did not include migrants other than those applying for asylum.
The data from the Bundesamt (2019) and the Bundeswahlleiter (2019) show that federal states with similar increases in outgroup population have seen different voting behaviour. First, Sachsen-Anhalt saw the highest increase in right wing voting behaviour and the highest increase in immigrant population. In contrast, Bremen had small or no increase in right wing voting behaviour despite a similar increase in immigrant population compared to Sachsen-Anhalt. This implies that outgroup size only influences voting behaviour under certain circumstances. To examine this, this study used Germany as a case study on when outgroup size affects anti-immigrant sentiment and voting behaviour.
Immigration described with the examples of Saschsen-Anhalt and Bremen has led to changing perceptions among native Germans, which can affect which party they vote for. Walczak, Van der Brug and De Vries (2012, p. 279) found that immigration increasingly explains party preferences over time. Right wing parties have often tapped into this perception.
People voting for anti-immigration parties have shown to have ideals, values and beliefs that connect with right wing ideologies (Chandler & Tsai 2001). According to Bridges and Mateut (2014, p. 424), these individuals are also more likely to have more nationalist feelings. They find that natives perceive people of a different race to have a negative impact on the culture of the destination country. As a result, some natives want to ban or restrict immigration.
The perceived consequences of immigration affect natives’ attitudes toward
migration. In contemporary society, attitudes toward migration have become more negative as the public perceives competition over scarce goods, higher taxes and threats to their own culture (Naumann, Stoetzer & Pietrantuono 2018, Meuleman, Davidov & Billiet 2009,
Hainmueller & Hiscox 2010). This has resulted in more right-wing voting behaviour as these more negative attitudes toward immigration are adapted by right wing parties to have electoral success.
To determine the effect of outgroup size on the relationship between perceived group threat, anti-immigrant sentiment and voting behaviour, I used the quantitative data analysis methods ordinal logistics regression and mixed effects ordered logistic regression.
The data used for this study was retrieved from three sources: the Bundesamt (2019), the Bundeswahlleiter (2019) and the European Social Survey round 7 from 2014.
The data shows that migration has consequences for voting behaviour. Immigration can lead to anti-immigrant sentiment, which affects voting behaviour. This connection has been explored by classical theories, such as group threat theory and intergroup contact theory. However, these theories cannot fully explain the phenomenon.
Group threat theory states that an increase in outgroup size leads to an increase in perceived group threat and anti-immigrant sentiment, which leads to an increase in right wing voting behaviour. However, the results in Bremen are contrary to this theory. Therefore, people in this federal state have less anti-immigrant sentiments and perhaps perceive less group threat.
Intergroup contact theory instead states that an increase in outgroup size leads to an increase in intergroup contact and a decrease in anti-immigrant sentiment, which would lead to a decrease in right wing voting behaviour. In Sachsen-Anhalt, however, an increase of the outgroup size has led to an increase in right wing voting behaviour. Therefore, people living in Sachsen-Anhalt could have more anti-immigrant sentiments and little or no intergroup contact.
Significant quantitative research has been conducted on the relationship between outgroup size, perceived group threat, intergroup contact, anti-immigrant sentiment and voting behaviour. However, the mediating effect of outgroup size on the relationship between perceived group threat, anti-immigrant sentiment and right wing voting behaviour or between intergroup contact, perceived group threat, anti-immigrant sentiment and right wing voting behaviour has not been examined quantitatively.
Therefore, this study examined multiple factors that might affect how outgroup size affects these relationships. These include individual, non-individual and control factors. The individual factors are perceptions of the country’s economic situation, the amount of trust individuals have in their social context, and how connected individuals feel to their country of residence. The non-individual factors are the region (divided into the federal states of Germany), unemployment rate per federal state and the percentage of Muslims per federal state. The control factors are age, gender, education, employment status and income.
The rest of this paper is structured as follows. Chapter 2 discusses provides the background of the German case study and discusses how economic factors affect anti-immigrant sentiment. Next, Chapter 3 discusses how group threat theory and economic explanations led the literature to find that perceived group threat leads to more anti-immigrant sentiment. Following this discussion, hypotheses are presented taking outgroup size, perceived group threat, anti-immigrant sentiment and voting behaviour into account. Next, intergroup contact theory is discussed as an opposing theory to group threat theory, which uses non-economic explanations for the change in anti-immigrant sentiment and hypotheses opposing those of group threat theory are presented. Finally, the non-economic explanations are grounded in the sociopsychological approaches. Chapter 4 discusses the methodology and operationalises the key concepts. Chapter 5 presents the results, after which Chapter 6 discusses the results and provides a conclusion.
2 The German case
In the last decade, a new migratory wave has begun from Africa, Asia and the Middle East into Europe. Many are refugees or migrants seeking employment in Europe. This
migratory movement has predominantly spread to the countries of western Europe.
Germany has been taking in many migrants, but the country is not known for its acceptance of foreigners from outside of the European Union.
Since the end of the Second World War, Germany has seen three large-scale waves of immigration (Ette, Hess & Sauer 2016). The first of these waves took place from the 1950s until the middle of the 1970s, when a recruitment stop put a halt to immigration into Germany. Steinert (2014) notes that, before the recruitment stop, foreign workers could only be employed in Germany as a last resort. They were meant to fill the gaps left by natives. In 1955, Italy was the first country to reach an immigration agreement with Germany, which meant that Italians could migrate to Germany for employment reasons. Following this, Spain and Greece signed agreements with Germany in 1960, and Turkey followed in 1961. Non-European countries, like Thailand, the Philippines, Egypt and Morocco also tried to reach immigration agreements with Germany, but were denied.
From the 1960s until the recruitment stop of 1973, Schönwälder (2004) found that the West German government actively sought to exclude migrants from Africa and Asia. As the government wished to control the migratory in-flow into the country, they created a barrier agreed upon by the federal ministries. Included in this barrier, were policies limiting employment migration. These policies made it impossible for employers to recruit unskilled labour migrants from countries outside of Europe (except for the United States, Canada, Australia, New Zealand and Israel).
These government policies did not arise from public concern (Schönwälder 2004, p. 256). Steinert (2014) finds that during the post-war period (1945–1960), West German society was unaware of the policy goals the government acted on. Even today, this information remains confidential. The reasons for these policies are to be linked to migratory movement control and ensuring that immigrants’ stay is temporary. On a side note, Turkey is seen as a country between Asia and Europe. This creates a situation where immigrants from Turkey can either be seen as Asians or Europeans. Except, in the 1960’s and 1970s, this never became the case as, because Turkey was then seen as part of Europe, so migration to Germany was not viewed as Turks being either Asian or European.
Immigration picked up again in the 1980s with a second large-scale migratory wave. Germany signed new bilateral labour migration agreements with countries like
Czechoslovakia, Poland, Romania, Bulgaria and former Yugoslavia (Rudolph 1994; Nienaber & Frys 2012). The economic boom ensured the growth of Germany’s economy and the labour market provided many job openings. Additionally, with the unification of Germany, an enormous internal migratory movement took place. From the middle of 1989 until late 1991, around 1.7 million people move from East to West Germany. However, after this internal migration flow, the labour market required even more workers. Therefore, German government signed the aforementioned agreements by which it opened paths for Central and East Europeans to enter the German labour market. These paths were project-tied employment, guest worker contracts, seasonal workers and cross-border commuters (Rudolph 1994, p. 3–5).
All these paths into the German labour market still saw immigrants as temporary. The chances of acquiring a second work permit or a residence permit were low.
Nonetheless, Coenders and Scheepers (2008) saw a trend between 1980 and 2000, during which German citizens became less resistant to foreigners integrating into German society. The third wave of migration into Germany has been taking place since the start of the 21st century. The European Lisbon Agenda forced German government to change their labour migration policy. These changes came through the Immigration Act (2005), Transposition Act (2007), Labour Migration Control Act (2009) and second Transposition Act (2012) (Ette, Hess & Sauer 2016, Burkert, Niebuhr & Wapler 2008). Through these acts, German accepted labour migration from non-European countries for the first time. Paths opened up for people from Africa, Asia, the Middle East and South America to enter the German labour market. These acts made it easier to acquire residence and work permits, to extend the duration of stay after migrants had their first permits and even to permanently settle in Germany.
The three large migratory waves into Germany were mainly for employment purposes. Germany has had a large need for workers, and foreign workers filled this demand. The German government achieved this through bilateral labour migration agreements and guest worker programs. After the second World War, Germany saw their native population grow and due to this growth, their need for foreign workers decreased. But in the contemporary labour market, the number of foreign workers entering the German labour market has started growing again putting pressure on the German labour market. As native Germans experienced a growing number of workers on the labour market and foreign workers finding more employment, their believes turned against foreign workers. This led to native Germans believing that foreign workers took employment opportunities they could have filled themselves. And, as more foreign workers enter the German labour market, native Germans want government to restrict immigration as they feel an increased competition over employment opportunities with foreign workers. Therefore, the last migratory wave has again put pressure on the German labour market. As a result, right-wing political parties have taken this matter into hand. These parties voice the anti-immigrant sentiment native Germans have felt more recently. This has led to right-wing political parties gaining electoral success in the most recent national elections.
Nevertheless, sociologists believe that labour market factors only partially explain anti-immigrant sentiment. There is a growing debate on whether non-economic factors better explain how and why anti-immigrant sentiment changes or has changed. In the following chapter, possible economic explanations for changes in anti-immigrant sentiment are described. These explanations are grounded in group threat theory. As these economic factors only partially explain anti-immigrant sentiment, the non-economic factors are described next sociopsychological approaches. These approaches compare group threat theory with intergroup contact theory.
3 Determining changes in anti-immigrant sentiment
Multiple explanations will be offered for the change in anti-immigrant sentiments across Europe. In this chapter, I discuss how the labour market competition model, the fiscal burden model and the sociopsychological approaches have been used to explain this change. The first two are economic explanations, the third is a non-economic explanation.
Additionally, I discuss how these explanations they have led to two classical sociological theories: group threat theory and intergroup contact theory.
the sociopsychological approaches, specifically on the mediating factor outgroup size, as this is a gap in previous research. Finally, I discuss alternative explanations and present the hypotheses.
3.1 Economic explanations
3.1.1 Labour market competition model
The labour market competition model states that competition over scarce labour opportunities on the labour market impacts anti-immigrant sentiment. Natives are against immigration of those with similar skill levels due to increased competition, but less opposed or unopposed to those with different skill levels. However, research has consistently
contradicted this model (Scheve & Slaughter 2001, Snidermand et al. 2004, Mayda 2006, Hainmueller & Hiscox 2010, Helbling & Kriesi 2014).
3.1.2 Fiscal burden model
The fiscal burden model states that anti-immigrant sentiment grows due to increased taxes. Natives with high incomes prefer immigrants with high income over immigrants with low income, as they would split tax payments more evenly. Immigrants with low income would benefit more from public services instead of adding to the supply of public services. Some previous research has found support for this model (Scheve & Slaughter 2001, Hanson et al. 2007, Facchini & Mayda 2009). However, Hainmueller and Hiscox (2010) found that high- and low-income natives prefer high skilled immigration over low-skilled immigration no matter what the situation is. The reason for this is that high skilled immigrants usually have high incomes. These immigrants therefore supply to public services benefiting natives in the need of such services, and share the tax payments with high income natives. Nevertheless, other explanations are possible.
3.2 Group threat theory
Both previous models state that anti-immigrant sentiment increases due to the threat level natives perceive. Whether in labour market opportunities, tax payments or public services, natives feel that an increase in immigrants decreases their chances of finding a job, receiving larger tax payments, or using public services.
These models are grounded in group threat theory, which has two steps. First,
competition over scarce resources can be either material (e.g. housing, labour market issues, tax payments, public services) or immaterial (e.g. language or culture) (Blalock, 1967).
Second, people belonging to the ingroup (the natives in the case of these models) who perceive a threat from the outgroup have more negative attitudes toward people from the outgroup (Quillian 1995). Stephan and Renfro (2002, p. 197) define perceived group threat as: ‘anticipation of negative consequences’. Schlueter and Scheepers (2010, p. 286) define anti-immigration attitudes as: ‘explicit preferences for denying to individuals or groups of people equality of treatment which they may wish’.
This theory links perceived group threat directly to anti-(im)migration sentiment (Coenders 2001, Scheepers et al. 2002). The current study examines the mediating effect of outgroup size on the relationship between perceived group threat and anti-immigrant sentiment. Outgroup size has previously been found to be a strong indicator of intergroup competition (Blalock 1967, Quillian 1995, Scheepers et al. 2002) and is regularly defined as the percentage of outgroup members relative to the total population in a given geographic context.
According to the general proposition that follows from group threat theory, an increase in perceived group threat and anti-immigration sentiment is expected in geographic contexts where the outgroup size increases. Following these more negative immigration sentiments, a rise in right wing voting behaviour is expected, as right wing parties more often voice anti-immigration sentiment (Mughan & Paxton 2006).
Nevertheless, perceived group threat does not always hold up. In Bremen, the outgroup size increased by over five percent between 2000 and 2017, but the increase in right wing voting behaviour was only six percent over the same time period. Therefore, the mediating effect of outgroup size on the relationship between perceived group threat, anti-immigrant sentiment and voting behaviour needs to be examined. In this thesis, group threat theory was used to ground the belief that an increase in outgroup size affects the relationship between perceived group threat and anti-immigrant sentiments, and the change in that relationship indirectly affects voting behaviour.
The mediating effect of outgroup size has not previously been studied. Studies on group threat examined the relationship between group threat and anti-immigrant sentiment but never examined mediating variables impacting this relationship. Therefore, the following hypotheses were formulated:
H1: When the outgroup size changes, the amount of perceived group threat increases. H2: When the outgroup size changes, the amount of anti-immigrant sentiment increases. H3: When the amount of perceived group threat increases, the amount of anti-immigrant sentiment increases.
H4: When the outgroup size changes, the amount of perceived group threat and anti-immigrant sentiment increases.
H5: When the outgroup size changes, the amount of perceived group threat increases, the amount of anti-immigrant sentiment increases and the amount of right-wing voting behaviour increases.
3.3 Intergroup contact theory
As shown in the examples of Bremen and Sachsen-Anhalt, group threat theory has not consistently shown the mediating effect of outgroup size on the relationship between perceived group threat, anti-immigrant sentiment and voting behaviour. Therefore, academics have started investigating other explanations why anti-immigration sentiment
Intergroup contact theory can explain what group threat theory failed to. Intergroup contact theory examines how negative attitudes between groups may be changed into more positive attitudes toward each other (Allport 1954), which lead to less right wing voting behaviour. The necessary contact between groups has to be face-to-face interaction. The theory states that intergroup contact leads to more positive intergroup attitudes. Believing that, when one has more contact with an individual from a different group, understanding of that person grows, which reduces the negative attitudes.
Allport (1954) and Pettigrew (1998) found that the circumstances under which these attitudes change need to be optimal, like having common goals, intergroup cooperation, equal status and authority support. However, Pettigrew and Tropp (2008) found that these optimal conditions are not necessary. The everyday intergroup contact at work or in the neighbourhood leads to more positive intergroup attitudes (Pettigrew & Tropp 2006). Nevertheless, sometimes attitudes do not change. The principle of self-selection, which means that ingroup members with negative attitudes toward outgroup members tend to avoid intergroup contact, support this (Schlueter & Scheepers 2010, p. 287.) But overall, intergroup contacts have shown to have beneficial effects on intergroup attitudes, namely that having intergroup contact reduces perceived group threat (Stephan & Renfro 2002).
As an increase in intergroup contact reduces perceived group threat, intergroup contact theory shows where it opposes group threat theory, which states that perceived group threat increases. Therefore, the mediating effect of outgroup size on the relationship between intergroup contact and anti-immigrant sentiment becomes relevant. While group threat theory states that an increase in outgroup size leads to an increase in both perceived group threat and anti-immigrant sentiment, intergroup contact theory states that an
increase in outgroup size leads to an increase in intergroup contact, but a decrease in anti-immigrant sentiment. Therefore, intergroup contact theory explains why in Bremen, an increase in outgroup size led to a decrease in both perceived group threat and
anti-immigrant sentiment. This also explains why the citizens of Bremen voted less for right wing parties than those in Sachsen-Anhalt, even though both federal states saw a similar increase in outgroup size.
In this study, intergroup contact theory is used to explain how an increase in the outgroup size affects the relationship between perceived group threat and anti-immigrant sentiment, and indirectly affects voting behaviour. Previous studies focussed on the relationship between intergroup contact and anti-immigrant sentiment. This thesis,
however, focusses on the mediating effect of outgroup size on this relationship. Therefore, following hypotheses were formulated:
H6: When the outgroup size increases, the amount of intergroup contact increases. H7: When the amount of intergroup contact increases, the amount of anti-immigrant sentiment decreases.
H8: When the outgroup size increases, the amount of intergroup contact increases, and the amount of anti-immigrant sentiment decreases.
H9: When the outgroup size increases, the amount of intergroup contact increases, the amount of anti-immigrant sentiment decreases and the amount of right-wing voting behaviour decreases.
3.4 Non-economic explanations
3.4.1 Sociopsychological explanationsTo study the effect outgroup size on the relationship between perceived group threat and anti-immigrant sentiment, this thesis draws on the sociopsychological approaches. These approaches attempt to explain how perceived threats to culture lead to anti-immigrant sentiments. Their main emphasis is on the role of group-related attitudes and symbols in shaping immigration attitudes. According to Hainmueller and Hopkins (2014), national identity, stereotypes and prejudice are some of the sociopsychological factors. One approach on perceived threat to national identity emphasises that native-born citizens should assimilate to the culture and become accustomed to the language in the destination country. A different approach emphasises how prejudice and stereotyping in combination with mass media can lead to more anti-immigrant sentiments, while local encounters with immigrants could lead to less anti-immigrant sentiments.
The sociopsychological approaches include a broad array of factors, but none of them focus on the effect outgroup size on the relationship between perceived threat,
anti-immigrant sentiment and voting behaviour. Therefore, I examine why voters in
contemporary societies in Western Europe vote for right wing extremist parties and whether this is due to an increase in outgroup size. Specifically, I ask whether citizens in
contemporary Western European societies influence the right-wing parties in their country. Van der Brug and Kritzinger (2012) studied generational differences in voting
behaviour, which may explain why voters have changed their mind and started voting for right wing parties. They found that the factors influencing voting behaviour have changed over time. The influence of long-term factors has decreased, and the importance of short-term factors has increased. In another study, Walczak, Van der Brug and De Vries (2012) found that long-term factors, like religion, social trust and the political left/right dimension have weak impact on voting behaviour. Instead, younger voters have replaced these influences with others, such as attitudes toward immigration. Additionally, Citrin et al. (1997) identified that voters decide what to vote based on what they believe will happen on a national level. Thus, national politics can influence attitudes toward immigration.
As described before, between 2013 and 2015, Germany saw a large increase in immigration and the rise of a new right wing populist party called AfD. To explain this change in voting behaviour, the main questions in this thesis are:
When does outgroup size affect the relationship between perceived group threat and anti-immigrant sentiment leading to an increase in right wing voting behaviour in Germany? When does outgroup size affect the relationship between intergroup contact and anti-immigrant sentiment leading to a decrease in right wing voting behaviour in Germany?
Previous research has focussed on two relationships. First, an increase in perceived group threat leads to an increase in anti-immigrant sentiment and indirectly increases right wing voting behaviour. This was shown through the labour market and fiscal burden models
This thesis tested whether those relationships hold true, and examined how outgroup size mediates these relationships. By testing opposing hypotheses from group threat theory and intergroup contact theory, this study determined the effect of outgroup size on perceived group threat, intergroup contact and anti-immigrant sentiment and the indirect effect of outgroup size on voting behaviour. Additionally, this study explored alternative explanations for how outgroup size affects the relationship between perceived group threat and anti-immigrant sentiment and the relationship between intergroup contact and perceived group threat. Alternative explanations include the unemployment rate by state, the economic situation in the country, the social trust experienced by natives, whether Muslims are allowed to come to a respondent’s country and nativism, which means how strongly individuals feels connected to their native country.
3.5 Alternative explanations
The first alternative explanation is the unemployment rate, which the labour market competition model attempts to use to explain anti-immigrant sentiment. One of Cochrane and Nevitte’s (2014) main questions was whether the unemployment rate affects anti-immigrant sentiment and how it links economic misery to anti-anti-immigrant sentiment. They found that far-right political parties shape public opinion about immigration through creating links between immigrants and domestic problems, including unemployment. Therefore, Cochrane and Nevitte state that when the unemployment rate is high, far-right political parties find more support for their ideals. To test this, the following hypothesis was formulated:
H10: When the unemployment rate and the outgroup size change, the amount of perceived group threat increases and the amount of anti-immigrant sentiment increases, leading to an increase in the amount of right-wing voting behaviour.
The second alternative explanation is the economic situation in a country. The economic situation in a country determines the fiscal burden of its population, as a poor economic situation fiscally burdens the population. For that reason, natives will have
increased anti-immigrant sentiment, as immigrants often put extra strain of the fiscal burden of the taxpayers, or at least the public perceives this to be true. Kuntz, Davidov and
Semyonov (2017) found that a shift in public views regarding the economic situation in a country likely results in higher levels of anti-immigrant sentiment. The insecure feelings caused by an unstable economy create a perceived threat on an individual level as well as an increase in anti-immigrant sentiment. This led to the following hypothesis:
H11: When the economic situation in a country becomes worse and the outgroup size changes, the amount of perceived group threat increases and the amount of anti-immigrant sentiment increases, leading to an increase in the amount of right-wing voting behaviour.
The third alternative explanation is the amount of social trust individuals have in others in a country. This falls under the sociopsychological approaches. According to Berning and Ziller (2017), individuals who have high social trust are more trusting of outgroup
members, are less likely to reject outgroup members and are less likely to vote for right wing parties. This led to the following hypothesis:
H12: When the amount of social trust increases and the outgroup size changes, the amount of intergroup contact increases and the amount of anti-immigrant sentiment decreases, leading to a decrease in the amount of right-wing voting behaviour.
The fourth alternative explanation is whether the individuals would allow Muslims to come to their country or not. As with social trust, this explanation falls under
sociopsychological approaches. Allowing Muslims to come to a country has more to do with their religion (Islam) than the fact that they are immigrants. A study on Muslim immigrants found that intergroup contact can alter this threat perception of native-born individuals. Creighton and Jamal (2015) found that misperceptions and negative stereotypes of the Muslim community/outgroup are stronger and more frequent when there is little to no intergroup contact. The opposite is true when intergroup contact became more frequent. This led to the following hypothesis:
H13: When the number of Muslims allowed to come to the country increases and the outgroup size changes, the amount of intergroup contact increases and the amount of anti-immigrant sentiment decreases, leading to a decrease in the amount of right wing voting behaviour.
The last alternative explanation is nativism. Betz (2019) asserts three types of nativism: economic nativism, welfare chauvinism and symbolic nativism. The current study focusses on symbolic nativism, which is ‘centred on the defence of fundamental traditions, values and historically evolved institutional arrangements that define a particular
community, its culture and identity’ (Betz 2019, p. 123). Nativism is used as a defence mechanism for native-born individuals to maintain their own beliefs, values and traditions. This led to the following hypothesis:
H14: When the amount of nativism decreases and the outgroup size changes, the amount of intergroup contact increases and the amount of anti-immigrant sentiment decreases, leading to a decrease in the amount of right-wing voting behaviour.
Table 1: All hypotheses Number Hypothesis
1 When the outgroup size changes, the amount of perceived group threat increases 2 When the outgroup size changes, the amount of anti-immigrant sentiment changes
3 When the amount of perceived group threat increases, the amount of anti-immigrant sentiment increases
4 When the outgroup size changes, the amount of perceived group threat and the amount of anti-immigrant sentiment increase 5 When the outgroup size changes, the amount of perceived group threat, the amount of anti-immigrant sentiment and the
amount of right-wing voting behaviour increases
6 When the outgroup size changes, the amount of intergroup contact increases
7 When the amount of intergroup contact increases, the amount of anti-immigrant sentiment decreases
8 When the outgroup size changes, the amount of intergroup contact increases, and the amount of anti-immigrant sentiment decreases
9 When the outgroup size changes, the amount of intergroup contact increases, the amount of anti-immigrant sentiment decreases and the amount of right-wing voting behaviour decreases.
10 When the unemployment rate and the outgroup size change, the amount of perceived group threat increases and the amount of anti-immigrant sentiment increases, leading to an increase in the amount of right-wing voting behaviour
11 When the economic situation in a country becomes worse and the outgroup size changes, the amount of perceived group threat increases and the amount of anti-immigrant sentiment increases, leading to an increase in the amount of right-wing voting behaviour
12 When the amount of social trust increases and the outgroup size changes, the amount of intergroup contact increases and the amount of anti-immigrant sentiment decreases, leading to a decrease in the amount of right-wing voting behaviour
13 When the number of Muslims allowed to come into the country increases and the outgroup size changes, the amount of
intergroup contact increases and the amount of anti-immigrant sentiment decreases, leading to a decrease in the amount of right wing voting behaviour
14 When the amount of nativism decreases and the outgroup size changes, the amount of intergroup contact increases and the amount of anti-immigrant sentiment decreases, leading to a decrease in the amount of right-wing voting behaviour
Note: Hypotheses 1–5 were formulated through group threat theory and hypotheses 6–9 through intergroup contact theory. Hypotheses 10– 14 are alternative explanations for the influence of outgroup size.
4 Methods
This study focusses on how outgroup size directly or indirectly affects the relationship between perceived group threat, anti-immigrant sentiment and voting behaviour, as well as the relationship between intergroup contact, anti-immigrant sentiment and voting
behaviour. For the empirical analysis of these relationships, data was collected from the Bundesamt Deutschland (2019), the Bundeswahlleiter (2019) and the European Social Survey.
The Bundesamt Deutschland is the German Federal Statistics Office, from which data on the foreign and total population in Germany between 2000 and 2017 was collected. This data is presented in Graph 1. The data collected from the Bundeswahlleiter consists of federal state election results between the years 1980 and 2018. This data is presented in Graphs 2–17. The European Social Survey is a cross-national survey on attitudes, beliefs, and behavioural patterns of diverse populations in Europe. For this study, the collected data consisted of the country specific data file on Germany for European Social Survey round 7 collected in 2014. This round of the European Social Survey asked specific questions on immigration.
To analyse the aforementioned relationships, the quantitative research methods ordinal logistic regression and multi-level ordered logistic regression were used. Ordinal logistic regression allows predictions to be made for an ordinal dependent variable given one or more independent variables. Multi-level ordered logistic regression allows similar predictions, except that it also allows a level dimension in the analysis. An example of why a multi-level analysis was needed is that outgroup size is included as a regional variable in the data, but voting behaviour is included as an individual variable.
The limitations of this study are the measurement of anti-immigrant sentiment, the measurement of alternative explanations and the period of data collection of this specific round of the European Social Survey. First, the measurement of anti-immigrant sentiment in the European Social Survey is problematic. As respondents were asked only one of four questions related to anti-immigrant sentiment, it was not possible to create an all-encompassing variable for anti-immigrant sentiment. Therefore, four separate variables were included and conclusions could only be reached for separate categories of anti-immigrant sentiment. This is further discussed in the section on dependent variables (Section 4.1)
Second, the period of data collection of this round of the European Social Survey is no ideal. As the latest migration wave into Germany is still taking place, it would have been better to use data from 2015 or 2016. However, since the European Social Survey only includes immigration as a rotating topic in the survey, the last possible round of the survey with immigration-related questions is from 2014.
Third, multiple alternative explanations were measured by one item. This could have biased the measurement due to measurement validity of the explanatory variables. The concepts of the economic situation in Germany, allowing Muslims to come to Germany and nativism might not be accurately represented by the items taken into account. However, as the items chosen are core to these concepts, using them was expected to reliably and viably measure the concepts.
As voting behaviour is the one of the two dependent variables, and citizens under the age of 18 years cannot vote, the age limits were restricted to 18 years and older.
Additionally, the main unit of analysis is German natives. For that reason, only German natives (people born and still living in Germany) were included in the analysis. Overall, the analysis included 1,526 individuals.
4.1 Dependent variables
The dependent variables used in this study are voting behaviour and anti-immigrant sentiment. Voting behaviour is defined as which party respondents voted for during the last federal state (Landtags) elections (Lubbers & Coenders 2017, p. 104). Voting behaviour was divided into three categories: left, centre and right according to the left/right dimension of the political spectrum. The item in the European Social Survey asked which party the
respondent voted for in the last national election. It was coded so that a respondent scored high when voting for a right-wing party and low when voting for a left-wing party in the last national election. Left wing parties have a more positive attitude toward immigration. The German political parties Die Linke and Die Grüne are included in this category. Centre parties have a more neutral attitude toward immigration. German political parties in this category are SPD, FDP and Piratenpartei. The right-wing parties have a more negative attitude toward immigration. Parties in this category are AfD and NPD.
The second dependent variable is anti-immigrant sentiment. This was defined as whether respondents would allow people from other countries to come to their country of residence. Four questions concerning anti-immigrant sentiment were used, and every respondent was asked one of these questions. For that reason, this study includes four variables concerning anti-immigrant sentiment. The four questions concerned whether the respondents allowed professionals from poor European countries to come to their country of residence, whether the respondents allowed professionals from poor non-European countries to come to their country of residence, whether the respondents allowed unskilled labourers from poor European countries to come to their country of residence, and whether the respondents allowed unskilled labourers from poor non-European countries to come to their country of residence. All four items were coded on a four-point scale, ranging from 1 (allow many to come here), to 4 (allow none). This means that for all four items, a high score indicates high anti-immigrant sentiments.
Table 2: Descriptive statistics of independent and dependent variables
Range Mean S.E.
Dependent variables
Voting behaviour 1-3 1.802 0.025
Anti-immigrant sentiment
Do not allow professional European 1-4 1.506 0.031 Do not allow professional non-European 1-4 1.617 0.035 Do not allow low-skilled European 1-4 2.178 0.048 Do not allow low-skilled non-European 1-4 2.434 0.050 Independent variables
Perceived group threat 1-3 1.620 0.033
Intergroup contact 1-3 2.639 0.030 Region 1-16 7.127 0.232 Social Trust 1-3 2.265 0.031 Nativism 1-4 3.277 0.034 Religion 1-4 1.957 0.039 Economic situation 1-3 2.519 0.033 Control variables Age 18-93 52.765 0.832 Gender Male 1/0 0.537 0.026 Education 1-3 2.493 0.030 Employment Has work 1/0 0.653 0.025 Income 1-3 2.291 0.036
Note 1: Scoring 1 on voting behaviour means Left, 2 means centre, 3 means right. Note 2: Scoring for region is in alphabetical order of Bundesländer.
4.2 Independent variables
The independent variables included in this study are perceived group threat, intergroup contact, outgroup size, region, unemployment rate, economic situation in the country, social trust, the religion of the outgroup members and nativism. These variables were split into two levels. First, the individual level included perceived group threat, intergroup contact, economic situation in the country, social trust, the religion of the outgroup members and nativism. The regional level included outgroup size, region and unemployment rate. The regional variables present results for regional statistics. Therefore, multi-level ordered regression was used to allow for conclusions on individual level when using regional level variables.
Perceived group threat was measured as the anticipation of negative consequences due to the presence of an outgroup (Stephan & Renfro 2002, p. 197). An example of an item that measured perceived group threat is: ‘Immigrants take jobs away in country or create new jobs’. This variable, composed through principal component analysis, was measured on a three-point scale, a high score (3) indicates that the respondent experiences high group threat.
Intergroup contact was defined as contact that a German native has with any individual belonging to an outgroup. This variable was composed through principal
component analysis. It was measured on a three-point scale where a high score (3) indicates that the respondent experiences a high amount of contact with individuals who are not in the ingroup. An example item is: ‘How often do you have contact with someone from a different race or ethnic group?’
Outgroup size was defined as the percentage of foreigners in a federal state. To calculate the outgroup size per state, data from the German Federal Statistics Office (das Bundesamt) was used. This information, as well as the increase in outgroup size between 2000 and 2017, is shown in Graph 1 (see appendix).
Region consists of Germany’s sixteen federal states. These are listed in an
alphabetical order, from Baden-Württemberg to Thüringen. Therefore, Baden-Württemberg scored a one on region and Thüringen a 16 . The distribution of regions is given in Graph 1.
Data on the unemployment rate per federal state was collected from the German Federal Statistics Office (das Bundesamt). The analysis used the unemployment rate as a percentage in every federal state (region) for 2014.
The economic situation was defined as being happy with the current state of the economy (in Germany). Only one question was included in the survey, This variable was measured on a three-point scale, where a high score (3) meant being happy with the current state of the economy in the country.
Social trust reflects a positive expectation about the trustworthiness of the generalised, abstract other, and a person’s level of social trust is thus a standard estimate of the
trustworthiness of an unknown other (Robinson & Jackson, 2001). Multiple items were used from the European Social Survey to compose this variable. An example item for this variable is: ‘Most people can be trusted, or you can’t be too careful‘. After principal component analysis, this variable was measured on a three-point scale, where a high score (3) means the respondent has a high amount of social trust.
The religion of the outgroup member used the item that asks whether respondents would allow few or many Muslims to come to their country. This question used a four-point scale, where a high score (4) means that the respondent does not allow any Muslims.
Nativism is the feeling that people are strongly connected to ‘their’ country. According to Lippard, ‘nativism is unique in that its ideological core rests on notions of nationalism and distinguishing between native and non-native’ (2011, p. 593). In the European Social Survey, only one item corresponds with nativism. It measures how close respondents feel to their country and is measured on a four-point scale. A high score (4) means that the respondents do not feel close at all to ‘their’ country, or that they are not nativists.
4.3 Control variables
The control variables included in this study are age, gender, educational attainment level, employment status and income level. Age was measured in years and had a range of 18 to 93. Gender was coded as a dummy variable, where a score of 1 meant the respondent is male. Educational attainment level was measured according to the ISCED classification (‘International Standard Classification of Education,’ n.d.). This means a respondent with low education level has completed education up to primary education. A high level of education means the respondent has finished tertiary education or higher. The European Social Survey dataset ISCED measures were included and recoded into a three-point scale variable. A high
score on education (3) meant the respondent had a high educational attainment level. Employment status was included as a dummy variable. A score of 1 meant that the
respondent had employment at the time of the survey. Lastly, income was measured on a ten-point scale, but recoded into a three-point scale. The top twenty percent of incomes scoring eight through ten, which are the highest, were coded as high (3) which meant that the respondent had a high level of income. Low income respondents were coded as low (1), representing an income between the scores of one through three.
5 Results
In this chapter, the results from the analysis are presented and discussed. Section 5.1 includes the results based on the relationships stated in group threat theory and intergroup contact theory. Section 5.2 includes the results on how outgroup size affects the
relationships.
5.1 Perceived group threat and intergroup contact theory
The first hypothesis (H1) stated that an increase change in outgroup size would lead to an increase in perceived group threat. Table 3 shows that native Germans perceive more group threat when the outgroup size changes confirming hypothesis 1. Additionally, natives with a higher educational level or a higher income perceive less group threat when the outgroup size changes.
A similar hypothesis (H6) was formulated for intergroup contact. As shown in Table 3, native Germans have increased intergroup contact when the outgroup size changes. This confirms the hypothesis. Additionally, a higher level of education increases the amount of contact.
Table 3: The effect of a change in outgroup size on perceived group threat and anti-immigrant sentiment
Perceived Group Threat
Intergroup Contact Anti-Immigrant Sentiment Do not allow professional European
Model 1 Model 2 Model 3
Coef. S.E. Coef. S.E. Coef. S.E.
Outgroup size 0.508 9.120 0.000 0.000 0.000 0.000 Age 0.007 0.006 -0.031*** 0.004 0.012* 0.007 Gender 0.246** 0.111 -0.055 0.086 0.156 0.191 Education -0.717*** 0.087 0.525*** 0.100 -0.992*** 0.214 Employment 0.069 0.142 0.167 0.116 0.577*** 0.163 Income -0.155*** 0.056 -0.082 0.054 -0.000 0.183 Number of observations 1526 1526 363 Wald chi2 170.71 233.14 34.10
Table 3 (continued)
Anti-Immigrant Sentiment Do not allow
professional non-European
Do not allow low-skilled European
Do not allow low-skilled non-European
Model 4 Model 5 Model 6
Coef. S.E. Coef. S.E. Coef. S.E.
Outgroup size 0.000 0.000 0.000 0.000 0.000 0.000 Age 0.012* 0.007 -0.001 0.008 0.019** 0.008 Gender 0.156 0.191 -0.155 0.122 0.029 0.175 Education -0.992*** 0.214 -0.551*** 0.158 -0.488*** 0.142 Employment 0.577*** 0.163 0.065 0.217 0.247 0.272 Income -0.000 0.183 0.024 0.190 -0.354*** 0.112 Number of observations 363 373 393 Wald chi2 34.10 17.47 36.94 Prob > chi2 0.0000 0.0037 0.0000 Log pseudolikelihood -326.2568 -470.78059 -507.60842 *p < 0.1; **p < 0.05; ***p < 0.01
Native Germans do not show more negative anti-immigrant sentiment when the outgroup size changes. As the second hypothesis anticipated the opposite of what was found, the results shown in Table 3 reject H2. Additionally, natives with a higher level of education have lower anti-immigrant sentiment.
Nevertheless, when native Germans do perceive more group threat, are less
tolerable of immigrants. In this sense, more perceived group threat leads native Germans to hold a higher amount of anti-immigrant sentiment, as shown in Table 4. This confirms H3. Table 4: The relationship between perceived group threat and anti-immigrant sentiment
Anti-Immigrant Sentiment Do not allow professional European Do not allow professional non-European Do not allow low-skilled European Do not allow low-skilled non-European
Model 1 Model 2 Model 3 Model 4
Coef. S.E. Coef. S.E. Coef. S.E. Coef. S.E. Perceived group threat 0.931*** 0.195 1.349*** 0.170 1.404*** 0.208 1.545*** 0.193 Age -0.017** 0.008 0.011 0.008 -0.007 0.008 0.017** 0.008 Gender 0.012 0.225 0.093 0.226 -0.302 0.206 -0.010 0.207 Education -0.281 0.206 -0.772*** 0.200 -0.185 0.180 -0.273 0.189 Employment -0.239 0.281 0.537* 0.277 0.018 0.260 0.304 0.255 Income -0.149 0.169 -0.015 0.168 0.134 0.143 -0.310 0.143 Number of observations 397 363 373 393 Wald chi2 36.47 85.95 68.52 94.63 Prob > chi2 0.0000 0.0000 0.0000 0.0000 Pseudo R2 0.0643 0.1312 0.0851 0.1129 *p < 0.1; **p < 0.05; ***p < 0.01
When native Germans have more contact with outgroup members, immigrants in this case, their amount of anti-immigrant sentiment decreases. This translates into native Germans being more tolerant of immigrants when they have more contact with them. This result contradicts the previous result. Nonetheless, the hypothesis that more intergroup contact lowers anti-immigrant sentiment holds true, as shown in Table 5.
Table 5: The relationship between perceived group threat and anti-immigrant sentiment Anti-Immigrant Sentiment Do not allow professional European Do not allow professional non-European
Do not allow low-skilled European
Do not allow low-skilled non-European
Model 1 Model 2 Model 3 Model 4
Coef. S.E. Coef. S.E. Coef. S.E. Coef. S.E.
Intergroup contact -0.957*** 0.210 -1.379*** 0.170 -0.588*** 0.197 -0.835*** 0.171 Age -0.023*** 0.008 0.003 0.008 -0.005 0.007 0.010 0.008 Gender 0.080 0.225 0.178 0.226 -0.176 0.210 0.043 0.206 Education -0.399* 0.210 -0.822*** 0.200 -0.463*** 0.176 -0.384* 0.180 Employment -0.183 0.285 0.797*** 0.277 0.141 0.258 0.207 0.248 Income -0.216 0.169 -0.084 0.168 0.037 0.145 -0.404*** 0.143 Number of observations 397 363 373 393 Wald chi2 38.25 76.45 25.11 53.42 Prob > chi2 0.0000 0.0000 0.0003 0.0000 Pseudo R2 0.0598 0.1243 0.0269 0.0550 *p < 0.1; **p < 0.05; ***p < 0.01
Table 6 shows that a changes in the outgroup size and increased perceived group threat led to more anti-immigrant sentiment (H4). This means that native Germans have more negative sentiments toward immigrants when the outgroup size changes, as they perceive more threat from outgroup members.
Intergroup contact theory implies that native Germans would be less negative toward immigrants when they have more intergroup contact as the number of outgroup members changes (H8). Table 7 confirms this hypothesis. Additionally, native Germans with higher educational levels hold less anti-immigrant sentiment.
Table 6: The effect of a change in outgroup size and an increase in perceived group threat on anti-immigrant sentiment Anti-Immigrant Sentiment Do not allow professional European Do not allow professional non-European
Do not allow low-skilled European
Do not allow low-skilled non-European
Model 1 Model 2 Model 3 Model 4
Coef. S.E. Coef. S.E. Coef. S.E. Coef. S.E.
Outgroup size 0.000 0.000 0.000 0.000 0.000 0.000 PGT 0.958*** 1.402*** 1.404*** 0.223 1.549*** 0.180 Age -0.019*** 0.011* -0.007 0.009 0.016** 0.007 Gender -0.016 0.101 -0.302** 0.149 -0.012 0.184 Education -0.269 -0.827*** -0.185 0.160 -0.273* 0.144 Employment -0.278 0.568*** 0.018 10.97 0.303 0.200 Income -0.181 0.006 0.134 0.192 -0.309*** 0.099 Number of observations 397 363 373 393 Wald chi2 55.84 133.41 150.01 571.06 Prob > chi2 0.0000 0.0000 0.0000 0.0000 Likelihood -326.08 -295.88 -437.05 -464.22 *p < 0.1; **p < 0.05; ***p < 0.01
Note: PGT is perceived group threat; likelihood is log pseudolikelihood.
Table 7: The effect of a change in outgroup size and an increase in intergroup contact on anti-immigrant sentiment Anti-Immigrant Sentiment Do not allow professional European Do not allow professional non-European
Do not allow low-skilled European
Do not allow low-skilled non-European
Model 1 Model 2 Model 3 Model 4
Coef. S.E. Coef. S.E. Coef. S.E. Coef. S.E.
Outgroup size 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 IGC -0.950*** 0.221 -1.385*** 0.271 -0.591*** 0.223 -0.832*** 0.140 Age -0.024*** 0.007 0.003 0.008 -0.005 0.009 0.010 0.010 Gender 0.061 0.197 0.174 0.198 -0.177 0.135 0.046 0.184 Education -0.380 0.257 -0.851*** 0.195 -0.464*** 0.143 -0.382** 0.165 Employment -0.216 0.326 0.815*** 0.153 0.144 0.184 0.206 0.274 Income -0.253** 0.113 -0.084 0.143 0.037 0.191 -0.406*** 0.105 Number of observations 397 363 373 393 Wald chi2 163.21 97.66 32.94 150.28 Prob > chi2 0.0000 0.0000 0.0000 0.0000 Likelihood -328.27 -299.24 -464.85 -494.54 *p < 0.1; **p < 0.05; ***p < 0.01
As mentioned in the introduction, right wing political parties have adopted anti-immigration sentiment as one of their main platforms. Therefore, it was hypothesised that increased anti-immigrant sentiment would lead to increased right-wing voting behaviour. When the outgroup size increases, H4 stated that increases in perceived group threat leads to increased anti-immigrant sentiment among native Germans, which leads to more right wing voting behaviour (H5).
Table 8 shows that while anti-immigrant sentiment against low-skilled European immigrants does increase right wing voting behaviour, this is caused more by perceived group threat than by anti-immigrant sentiment. These results contradict H5.
An interesting finding is that the threat perceived by native Germans from non-European immigrants is greater than the perceived threat from non-European immigrants. Another interesting finding is that native Germans have stronger sentiment against low-skilled European immigrants.
Even though right-wing political parties made adopted anti-immigration sentiment as a main agenda point, not every federal state shows similar patterns. This has led to the hypothesis that when the outgroup size changes, and the amount of intergroup contact increases, sentiment toward immigrants is more positive and native Germans will not vote for right wing parties as much (H9). Table 9 shows that intergroup contact between native Germans and non-European immigrants lowers the amount of right-wing voting behaviour. For European immigrants, intergroup contact does not affect voting behaviour. Perhaps this points toward an alternative explanation, that native Germans do not see European
immigrants as an outgroup. H9 is rejected, as the results do not show that intergroup
contact and anti-immigrant sentiment simultaneously lower the amount of right-wing voting behaviour.
An interesting finding is that native Germans hold anti-immigrant sentiment when it comes to low-skilled European immigrants. This result was also found for the previous hypothesis. Therefore, native Germans might hold more anti-immigrant sentiment toward low-skilled European immigrants than other immigrants.
Table 8: The effect of a change in outgroup size, an increase in perceived group threat and an increase anti-immigrant sentiment on voting behaviour
Voting behaviour
Model 1 Model 2 Model 3 Model 4
Coef. S.E. Coef. S.E. Coef. S.E. Coef. S.E.
Anti-immigrant sentiment
Do not allow professional European 0.376 0.323
Do not allow professional non-European 0.197 0.184
Do not allow low-skilled European 0.620*** 0.199
Do not allow low-skilled non-European 0.183 0.175
Outgroup size 0.000 0.000 0.000 0.000 0.000 0.000 0.001 8.125
Perceived group threat 0.605*** 0.230 0.843*** 0.245 0.346** 0.166 0.751*** 0.257
Age 0.003 0.005 0.009 0.010 0.027* 0.014 -0.009 0.006 Gender 0.288 0.288 0.215 0.219 0.103 0.225 0.916*** 0.286 Education -0.466** 0.183 0.005 0.241 0.191 0.189 -0.735** 0.348 Employment -0.492* 0.289 -0.346 0.323 0.323 0.226 0.195 0.205 Income 0.395*** 0.150 0.178 0.195 -0.049 0.188 0.341 0.234 Number of observations 397 363 373 393 Wald chi2 43.13 42.78 35.71 25.06 Prob > chi2 0.0000 0.0000 0.0000 0.0007 Log pseudolikelihood -271.89 -239.65 -239.71 -276.43 *p < 0.1; **p < 0.05; ***p < 0.01
Table 9: The effect of a change in outgroup size, an increase in intergroup contact and an increase anti-immigrant sentiment on voting behaviour
Voting behaviour
Model 1 Model 2 Model 3 Model 4
Coef. S.E. Coef. S.E. Coef. S.E. Coef. S.E.
Anti-immigrant sentiment
Do not allow professional European 0.433 0.324
Do not allow professional non-European 0.262 0.165
Do not allow low-skilled European 0.700*** 0.198
Do not allow low-skilled non-European 0.282* 0.155
Outgroup size 0.000 0.000 0.000 0.000 0.001 8.582 Intergroup contact -0.475* 0.268 -0.720*** 0.241 -0.137 0.222 -0.774*** 0.242 Age 0.002 0.007 0.005 0.009 0.027** 0.013 -0.015*** 0.005 Gender 0.324 0.291 0.267 0.222 0.149 0.222 0.952*** 0.271 Education -0.562*** 0.189 -0.047 0.210 0.135 0.188 -0.749** 0.377 Employment -0.432 0.300 -0.254 0.280 0.337 0.235 0.152 0.224 Income 0.365*** 0.143 0.142 0.249 -0.066 0.186 0.298 0.236 Number of observations 397 363 373 393 Wald chi2 35.47 31.84 42.11 172.26 Prob > chi2 0.0000 0.0000 0.0000 0.0000 Log pseudolikelihood -274.21 -242.83 -240.67 -276.08 *p < 0.1; **p < 0.05; ***p < 0.01
5.2 When the outgroup size has an influence
To answer the main questions, this section shows the results regarding when an increase in outgroup size affects the relationships stated in hypotheses 10-14. First, the results are discussed for effect of the unemployment rate, followed by the results for the effects of economic situation, social trust, allowing Muslims and nativism.
Following the reasoning from the labour market competition model and perceived group threat theory, it was hypothesised that a change in the unemployment rate and outgroup size would lead to an increase in right wing voting behaviour (H10). Table 10 shows that this hypothesis is only confirmed for low-skilled European immigrants. For the other types of immigrants, it is found that changes in unemployment rate and outgroup size do not affect the amount of right wing voting behaviour. Nevertheless, only increases in the perceived threat by native Germans increase their amount of right-wing voting behaviour. Therefore, H10 is rejected
Table 10: Effect of a change in unemployment rate and outgroup size on perceived group threat and anti-immigrant sentiment leading to more right-wing voting behaviour
Voting behaviour
Model 1 Model 2 Model 3 Model 4
Coef. S.E. Coef. S.E. Coef. S.E. Coef. S.E.
AIS Do not allow prof. Euro. 0.376 0.311 Do not allow prof. non-Euro. 0.201 0.182 Do not allow l-s. Euro. 0.605*** 0.197 Do not allow l-s. non-Euro. 0.184 0.186 Outgroup size 0.000 0.000 0.000 0.000 0.000 0.000 0.000 PGT 0.605*** 0.223 0.833*** 0.243 0.329** 0.164 0.751*** 0.269 Unemp. rate 0.000 0.000 0.000 0.000 83.579 54.921 0.000 Age 0.003 0.005 0.009 0.010 0.026* 0.014 -0.009 0.006 Gender 0.288 0.290 0.214 0.216 0.109 0.225 00.917*** 0.296 Education -0.466*** 0.179 0.004 0.237 0.145 0.179 -0.731** 0.345 Employment -0.493* 0.290 -0.358 0.319 0.316 0.224 0.198 0.223 Income 0.395*** 0.145 0.184 0.192 -0.059 0.188 0.337 0.251 Number of observations 397 363 373 393 Wald chi2 40.96 40.96 33.41 24.03 Prob > chi2 0.0000 0.0000 0.0000 0.0011 Likelihood -271.87 -239.43 -271.87 -283.20 *p < 0.1; **p < 0.05; ***p < 0.01
Note: AIS is anti-immigrant sentiment; l-s is low-skilled; PGT is perceived group threat; Unemp. rate is unemployment rate; likelihood is log pseudolikelihood.
Based on the fiscal burden model and group threat theory, it was hypothesised that a better economic situation in Germany and a change in outgroup size would increase the
amount of right wing voting behaviour (H11). Table 11 shows that the economic situation in Germany rarely affects the amount of right-wing voting behaviour of native Germans. Therefore, H11 is rejected.
However, some interesting results were found. First, the perceived group threat experienced by native Germans significantly influences the amount of right-wing voting behaviour of native Germans. Second, native Germans hold negative anti-immigrant sentiment only toward low-skilled European immigrants.
Table 11: Effect of an increased economic situation in Germany and a change in the outgroup size on perceived group threat and anti-immigrant sentiment leading to more right-wing voting behaviour
Voting behaviour
Model 1 Model 2 Model 3 Model 4
Coef. S.E. Coef. S.E. Coef. S.E. Coef. S.E.
AIS Do not allow prof. Euro. 0.428 0.336 Do not allow prof. non-Euro. 0.267 0.163 Do not allow l-s. Euro. 0.624*** 0.199 Do not allow l-s. non-Euro. 0.181 0.174 Outgroup size 0.000 0.000 0.000 0.000 0.000 0.000 0.000 PGT 0.643*** 0.241 0.896*** 0.260 0.381** 0.162 0.755*** 0.239 Eco. situation 0.306 0.221 0.418** 0.217 0.316 0.261 0.040 0.228 Age 0.003 0.005 0.005 0.009 0.027* 0.014 -0.010 0.006 Gender 0.242 0.280 0.047 0.264 0.049 0.240 0.911*** 0.299 Education -0.482*** 0.176 -0.024 0.244 0.132 0.203 -0.734** 0.332 Employment -0.500* 0.278 -0.300 0.332 0.326 0.230 0.193 0.220 Income 0.368** 0.157 0.127 0.192 -0.089 0.173 0.332 0.240 Number of observations 397 363 373 393 Wald chi2 41.67 107.10 41.64 24.03 Prob > chi2 0.0000 0.0000 0.0000 0.0011 Likelihood -270.54 -236.13 -238.55 -283.20 *p < 0.1; **p < 0.05; ***p < 0.01
Note: AIS is anti-immigrant sentiment; l-s is low-skilled; PGT is perceived group threat; likelihood is log pseudolikelihood.
Hypothesis twelve was based on the sociopsychological approaches and intergroup contact theory and stated that increased social trust and a change in outgroup size lead to lower amount of right wing voting behaviour. Table 12 shows that native Germans’ social trust does not affect the amount of right-wing voting behaviour by native Germans.
However, intergroup contact does lower the amount of right-wing voting behaviour, except when intergroup contact is achieved with non-European immigrants. Additionally, native Germans hold anti-immigrant sentiment against low-skilled European immigrants, leading to a higher amount of right-wing voting behaviour.
Table 12: Effect of an increased economic situation in Germany and a change in the outgroup size on perceived group threat and anti-immigrant sentiment leading to more right-wing voting behaviour
Voting behaviour
Model 1 Model 2 Model 3 Model 4
Coef. S.E. Coef. S.E. Coef. S.E. Coef. S.E.
AIS Do not allow prof. Euro. 0.425 0.320 Do not allow prof. non-Euro. 0.218 0.153 Do not allow l-s. Euro. 0.663*** 0.186 Do not allow l-s. non-Euro. 0.265* 0.154 Outgroup size 0.000 0.000 0.000 0.000 0.000 0.000 0.000 8.711 IGC -0.473* 0.268 -0.712*** 0.241 -0.122 0.216 -0.754*** 0.239 Social trust -0.040 0.148 -0.198 0.239 -0.299* 0.170 -0.267 0.199 Age 0.002 0.007 0.005 0.011 0.027** 0.013 -0.015*** 0.005 Gender 0.322 0.282 0.280 0.214 0.110 0.209 0.952*** 0.275 Education -0.560*** 0.192 -0.035 0.214 0.182 0.186 -0.718* 0.367 Employment -0.433 0.300 -0.235 0.310 0.344 0.237 0.149 0.232 Income 0.366** 0.143 0.157 0.184 -0.060 0.180 0.320 0.229 Number of observations 397 363 373 393 Wald chi2 35.44 30.66 41.46 168.84 Prob > chi2 0.0000 0.0002 0.0000 0.0000 Likelihood -274.20 -242.32 -239.79 -275.21 *p < 0.1; **p < 0.05; ***p < 0.01
Note: AIS is anti-immigrant sentiment; l-s is low-skilled; IGC is intergroup contact; likelihood is log pseudolikelihood.
Hypothesis 13 stated that when native Germans allow more Muslims to come to Germany and the outgroup size changes, it leads to a lower amount of right wing voting behaviour by native Germans. Table 13 shows that a lower score on allowing Muslims to come to Germany goes with a lower amount of right wing voting behaviour. This result has the strongest implications when it comes to anti-immigrant sentiment against European immigrants. As the results show that intergroup contact lowers the amount of right wing voting behaviour of native Germans, H13 is rejected. Some evidence was found that allowing more Muslims to come to Germany leads to more intergroup contact, which lowers the amount of right-wing voting behaviour of native Germans. Anti-immigrant sentiment remains unaffected.
