Intolerance Toward Immigrants in Switzerland: Diminished Threat Through Social Contacts?


  • We are grateful to the two anonymous referees and the editors for their helpful comments and suggestions.


Intolerance toward immigrants has recently reached noticeable highs in Switzerland. Referring to the conflict theory, the perception of a specific group as a threat tends to lead to intolerance toward that group. The expectation of a negative relationship between threat and tolerance is nevertheless not assumed to be universally valid for all members of society. In particular, the contact theory assumes that intercultural encounters should promote positive attitudes toward culturally different individuals and groups. Using a new and unique data set, our Heckman selection models reveal that Swiss who view rising immigration to mean a loss of economic privileges and an erosion of Swiss cultural values are less tolerant toward immigrants. Moreover, our results indicate that contact with immigrants may moderate this effect. However, not all group settings are able to reduce the perceived threats in a similar way, and not all sorts of social contact are able to foster tolerance toward immigrants.


Switzerland, for a variety of reasons, is traditionally thought to be a rather tolerant society. Some consider the Alpine country to be a “paradigmatic case of political integration” (Deutsch, 1976), where historical developments helped to incorporate different religions and languages without destroying cultural identities (Linder, 2010; Vatter, 2008). For others, Switzerland represents the prototype of a consensus democracy and is therefore predisposed to “kinder and gentler” ways (Lijphart, 1999). In addition to these expert perspectives, over 75 percent of the European population saw Switzerland as “a good example of a multicultural society and the peaceful co-existence of different groups and languages,” as already indicated by the 1997 Eurobarometer survey.

At the same time, Switzerland also displayed intolerance, as 71 percent of the European respondents stated that the Swiss do not like immigrants (Eurobarometer 1997). This sentiment reflects the traditional approach of Swiss immigration policies aimed at protecting “Swiss culture” against foreign influences (Manatschal, 2011; Niederberger, 2004; Skenderovic, 2009).1 This skepticism toward immigrants has recently peaked: new developments point to a distinct cultural shift in Switzerland from the tolerant consensus democracy to a nation characterized by increasing animosity toward immigrants (Bornschier, 2010; Giugni and Morariu, 2010). A rising number of popular votes against immigrants and on immigration issues have been held in the last decade (Kriesi et al., 2005). These culminated when an initiative against immigrants was accepted by the majority of the Swiss population for the first time in Swiss history.2 Almost three years ago, a clear majority voted in favor of a popular initiative concerning the existence of minarets in Switzerland, an issue that attracted worldwide attention (Antonsich and Jones, 2010). Although the aim of the initiative was to ban minarets from Switzerland, it was also a symbolic act to prevent the further growth of non-Swiss cultures in Switzerland. Prejudice and stereotyping is not however limited to Muslim immigrants from the former Yugoslavia, the Balkan states, or other non-Western European states: Discrimination can also be observed toward highly educated and qualified Western Europeans who come to Switzerland in search of employment. In particular, in many parts of the Swiss-German speaking population we can observe some hard feelings toward their German neighbors, likely stemming from the influx of highly-qualified Germans seeking employment in Switzerland as well as the stereotypes associated with the differences in the languages and cultures (Helbling, 2011). In this respect, intolerance toward immigrants emerges from a perceived cultural and economic threat brought in by immigrants with lower qualifications as well as by highly educated and highly qualified immigrants.

It is here that this investigation finds its starting point: We are interested in what drives intolerance toward immigrants in Switzerland. According to the key insight of the literature on tolerance, intolerance should mostly arise from perceptions that immigrants threaten important values or constitute danger to the economic situation (Gibson 1992, 569; Stouffer, 1955; Sullivan et al. 1993, 78-79, 186). The expectation of a negative relationship between threat and tolerance is nevertheless not assumed to be universally valid for all members of society. In particular, the contact theory assumes that intercultural encounters should lead to positive attitudes and behavior towards culturally different individuals and groups (Pettigrew, 1998). However, whereas casual and initial perception of or contact with out-group members often initially evokes anxiety and hesitation, contact theorists assume that repeated and continuous encounters with out-group members in varied settings result in a learning process about the out-group, which in turn corrects negative views of this group and reduces prejudice (Pettigrew, 1998; Pettigrew and Tropp, 2011). By evaluating social substructures (e.g., circles of friends, workplace, neighborhood, civil society organizations) in which individuals are embedded, we seek to gain insight into how intolerance is fostered. Since people spend most of their everyday lives within such substructures, we place these social contexts at the center of our analysis of tolerance formation, for it is there where encounters take place and relations as well as attitudes are formed (Mutz and Mondak, 2006).

We demonstrate that social ties, it appears, can effectively overcome the feeling of being threatened by immigrants. Not all group settings, however, are able to reduce the perceived cultural and economic threats in a similar way, and not all sorts of contact situations are able to foster tolerance toward immigrants. Altogether, these findings contribute to the literature on tolerance and to the broader literature on social contexts and anti-immigrant attitudes.

To date, several analyses have examined relationships between threat perceptions, interpersonal contact and attitudes toward immigrants in Switzerland (see Cattacin et al., 2006; Giugni and Morariu, 2010; Manzoni, 2007; Stolz, 1998, 2001; Wimmer, 2002). Using a new and unique data set, our paper contributes to this research in several important ways:

  • – First, with regard to the essentials of contact theory, we apply a more fine-grained measurement of both the type as well as the context of interpersonal contact. In this vein, we take advantage of the detailed questions on personal ties in our survey to explore the effects of inter-group relations and contact. Based on this direct contact measurement, we are able to distinguish between regular and casual contacts, thereby satisfying a basic condition of the contact hypothesis, which has often been neglected in previous research (Allport, 1954; Pettigrew, 1998; Pettigrew and Tropp, 2011). Moreover, the contact questions capture the degree to which one regularly interacts with immigrants in different social substructures. We can thus differentiate between the effects of the contact quantity in four specific social substructures (circle of friends, neighborhood, at work, in associations). Overall, these detailed measures of social contacts allow us to establish a more comprehensive framework for the relationship between threat, contact, and intolerance toward immigrants in Switzerland.
  • – Second, we propose a new measurement of intolerance in Switzerland. Whereas previous research equates intolerance with negative attitudes toward a group or with support for restrictive exclusionary policies, we define tolerance as the willingness to grant particular rights to groups found objectionable. We thus go beyond the question of anti-immigrant attitudes by referring to both a rejection as well as an acceptance component of tolerance (Forst 2003; Gibson, 1992, 2006; Mondak and Sanders 2003; Scanlon 2003). In this vein, our intolerance conceptualization makes it possible to clearly distinguish this concept from widely employed ones such as xenophobia, ethnocentrism, or negative attitudes toward immigrants. In addition, this conception is expected to deliver deeper insights with regard to the role of social contacts in explaining attitudes toward immigrants. Most recently, Gibson (2013), in discussing the importance of tolerance research within the social sciences, underscored the necessity of adequately measuring this concept.
  • – Third, we use Heckman selection models to adequately capture the two-steps underlying the tolerance concept (rejection and acceptance). Although Mondak and Sanders (2003, 2005) already indicated that selection models would be a methodologically adequate procedure to analyze tolerance and intolerance, to date this method has not been applied in the research on tolerance.

The remainder of our article is structured as follows: The second section outlines the conceptualization of intolerance. In the third section we present our theory linking dissimilarity, threat, and contact to the formation of intolerance. In the fourth section we elaborate on the methodology used and subsequently subject our main hypotheses to systematic empirical testing. The most important findings are summarized and discussed in the fifth section. Finally, a conclusion completes the article.

Conception of Intolerance

While intolerance is simply defined as the absence of tolerance (see Forst, 2003; King, 1998; Mondak and Sanders, 2002; Sullivan et al., 1993), the definition of tolerance is still “in a state of disarray” (Ferrar, 1976, 63). This disarray is rooted in the discrepancy between the general publics’ understanding and the philosophical definition of tolerance. In the public sphere tolerance is mainly understood as a general positive belief in terms of an absence of prejudice, racism, or ethnocentrism. In this sense, the UNESCO declares tolerance as the positive recognition of human rights and civil liberties (Unesco, 1995); or even as “harmony in difference”. Sniderman et al. (1989) refer to this approving kind of tolerance as principled tolerance. Principled tolerance thereby originates from the incorporation of democratic principles and the application of them to all societal and political groups without any exception.

In contrast to this mere positive and approving view, the philosophical interpretation of tolerance states that tolerance is the willingness to “put up with” others who are different from oneself (e.g. Forst, 2003; Gibson, 1992, 2002; Mutz, 2001; Sullivan et al., 1993; Stouffer, 1955). Or more precisely, “one is tolerant to the extent one is prepared to extend freedoms to those whose ideas one rejects, whatever these might be” (Sullivan et al., 1979, 784; Mendus, 1999, 3; Mutz, 2001). This definition implies that disapproval or objection necessitate tolerance (Sullivan et al., 1993, 4). One can only be called tolerant if he or she first rejects a group and then grant them certain societal rights even though he or she finds this group objectionable. In this regard, tolerance is a multi-dimensional concept (Ferrar, 1976) consisting of both a rejection component and an acceptance component. Scanlon (2003, 187), for example, mentions that “tolerance is an attitude that is intermediate between wholehearted acceptance and unrestrainable rejection”. More generally, this tolerance dimension incorporates accepting the disagreeable, whereas the everyday dimension of tolerance is based on a positive belief and approval without prior rejection.

In this paper, we will follow the philosophical interpretation of tolerance for two reasons: first, we are interested in why citizens permit or do not permit societal rights to groups they disapprove of or even oppose to; and not why they approve the acceptable (Ferrar, 1976, 64). Today's true societal challenge is to tolerate even those you object. This tolerance dimension is a means to overcome the societal conflicts of increasing diversity and diverging values. Tolerance, in this sense, is the ability to bear something potentially difficult. Second, by including the rejection component we are further able to distinguish tolerance from other concepts such as indifference or affirmation. Individuals not stating any objection against a group or attitude are considered as mostly affirmative as they do not fulfill the necessary condition of rejection. Moreover, the opposite of tolerance is intolerance, and not indifference (Forst, 2007; Orlenius, 2008): Only those who grant full political and social rights without exception to the groups they object to can be characterized as tolerant (Mondak and Sanders, 2002). In contrast, those advocating any sort of restrictions on political or social rights are considered as intolerant, whereby they may differ in terms of their depth of intolerance. Intolerance becomes deeper when the respondent denies more rights to his or her objectionable group. The degree of intolerance is most commonly constituted by the number of social rights a group is guaranteed, despite of the general objection toward them (Nevitte, 1996, 59).3 Although the differentiation between tolerance and intolerance may be dichotomous, the concept of tolerance and intolerance as a whole is continuous (see Gibson, 2005), ranging from true tolerance to deep intolerance.4

Perceived Threat, Contact, and Intolerance: Theory and Hypotheses

“It is hardly possible,” John Stuart Mill observed, “to overstate the value […] of placing human beings in contact with other persons dissimilar to themselves, and with modes of thought and action unlike those with which they are familiar” (Mill, 1848, 594). Although Mill wrote about life in the mid-19th century, observers of contemporary political culture continue to be concerned with social changes that have taken place over the past decades, such as the breaking apart of traditional relationships (e.g., family, clubs), steady waves of immigration, or the unprecedented rate at which information is exchanged through new channels of communication (e.g., new media)—all of which contribute to a mounting confrontation with diverging life-styles, opinions, and attitudes (Kirchner et al., 2011). How these processes may affect tolerance is not entirely certain: While they do lead to cultural and ethnical diversity, different ethical perspectives and values are placed in competition with each other, thereby increasing the demand for tolerance. The literature on tolerance and intolerance formation has placed threat at the center of its discourse (Gibson 1992, 569; Iglic, 2010, 718; Sniderman et al., 1989; Stouffer, 1955; Sullivan et al., 1993). Sullivan et al. (1993, 78-79), for example, argue: “In deciding how tolerant or intolerant to be toward a target, there are several obvious considerations. The first and foremost is probably perceptions of threat posed by the group. In most cases, respondents might be expected to be tolerant of a target group no matter how much they object to it, provided it does not pose any serious threat to them or their values.”

Within the so-called conflict theory there are two accounts referring to mechanisms through which threat fosters prejudice and undermines tolerance (Rudolph and Popp, 2010). The dissimilarity thesis is rooted in psychological understandings of categorization and belief congruency. It is based on the observation that people prefer to associate with and are more comfortable with those believed to share similar belief systems (Rokeach et al., 1960). The perception of otherness plays a crucial role here, as observable ethnic, linguistic, and religious differences stemming from diverse immigration backgrounds are a primary means of social categorization and shape subsequent perceptions of belief congruency. Otherness due to different immigration backgrounds is often connected to the belief that immigrants hold different interests, values, or different social expectations and beliefs than the majority of the population, consequently resulting in a sense of wariness (Rahn et al., 2003). This wariness manifests itself through less positive evaluations of out-group members. Individuals are expected to have less tolerance for “others”, especially for people of a different ethnicity or culture, i.e. immigrants. As an alternative explanation, the economic threat thesis argues that members of the ethnic majority in a given setting will become more hostile toward members of a minority group as the size of that minority group increases (Blalock, 1967). This account would predict, for example, that the Swiss in a given locality will exhibit greater animosity toward immigrants as the number of immigrants in that locality increases. The presumed mechanism underlying this relationship is the threat—real or perceived—generated by inter-ethnical competition for scarce resources and the fear of losing one's social status (Blalock, 1967; Côté and Erickson, 2011; Forbes, 1997; Oliver and Wong, 2003). As the size of a minority group increases, members of the majority are more likely to feel that their social, economic, and political privileges are at risk, thus becoming more intolerant toward non-Swiss. This discussion leads to the following hypothesis:5

Hypothesis 1. The more an individual feels economically or culturally threatened by immigrants, the more intolerant he or she will be toward non-Swiss.

It has to be noted, however, that greater exposure to diverse groupings as well as their values and lifestyles might also bridge the gap between in-groups and out-groups. The contact hypothesis states that interaction between members of different groups could overcome existing prejudice and could thus aid in developing positive attitudes toward groups and values that differ from one's own (Allport, 1954).6 Following Ikeda and Richey (2009, 657), “tolerance requires ‘perspective-taking’, which is the psychological ability to understand another person's feelings and desires.” This ability to take a perspective, again, is usually developed through social interaction (Mutz, 2002; Pettigrew, 1998). In other words, social interaction with people of conflicting viewpoints allows us to see where the other person is coming from. It is this open mindedness and learning about diverse perspectives, ideas, and lifestyles that precedes the dismantling of prejudice (Allport, 1954; Dovidio et al., 2003, 10; Iglic, 2010, 718). As a consequence, positive attitudes toward the formerly rejected group may evolve and more intense contact may be developed among diverse group members.

Allport (1954), however, was well aware that certain conditions must be fulfilled before contact between individuals from different groups can lead to diminished prejudice. Contact must be of such a frequency, closeness, and duration that it has the potential to lead to meaningful relationships between individuals. In addition, contacts have to be symmetrical such that interacting parties are close to having identical status in the encounter situation (see also Dovidio et al., 2003; Pettigrew, 1998; Pettigrew and Tropp, 2011; Rydgren and Sofi, 2011). We must therefore distinguish between casual contacts and true personal relations (Rydgren and Sofi, 2011). Casual contacts will not provide relevant new information about others different from oneself and do not encourage increased familiarity. Altogether, casual contacts are unlikely to reduce prejudices and are even likely to increase stereotypes and prejudices by providing highly biased information about people with conflicting viewpoints (Pettigrew and Tropp, 2011; Rydgren and Sofi, 2011; Wagner, 2002). We therewith expect that casual and non-regular contacts do not promote enhanced cultural understanding and may even intensify existing prejudice and threat perceptions. In contrast, Pettigrew (1998) argues that particularly institutionalized contact, which takes place on a regular basis, offers the possibility of personal relationships between individuals, independent of their ethnic, political, or religious background, and may even provide the foundation for possible friendships across diverse groups. Especially these true contacts may enforce perspective-taking and enhanced cultural understanding thereby diminishing prejudice and threat perceptions.

Such social interactions mainly occur within social contexts like the workplace, neighborhood, circle of friends, or a civil society organization in which individuals are embedded. As people spend most of their everyday lives within such substructures on the meso-level, it is there where social encounters take place and relationships are formed. The vast majority of people devote a lot of time to their friends, neighborhood, or workplace. The organizational setting within voluntary associations guarantees that individuals interact on a continual and enduring basis. Doubtlessly, some of these social substructures do not guarantee social interaction on a regular basis, but they offer an arena to get in touch with diverse groupings. Within these social contexts thus the effects of casual and regular contacts have to be distinguished. According to these considerations our second research hypothesis is formulated as following:

Hypothesis 2. The more regular contact an individual has with immigrants among his or her friends, neighbors, at the workplace, or in an association the lower the effect of perceived economic or cultural threat on intolerance toward immigrants. Casual contacts, however, are thought to maintain suspicion and threat perceptions consequently holding intolerance.

Data, Measurement and Method

To test the hypotheses about how threat and contact with immigrants in different social substructures shapes interethnic tolerance, we draw on a survey funded by the Schweizerische Gemeinnützige Gesellschaft. This survey contains 4,955 respondents from 60 municipalities ranging in size between 2,000 and 20,000 inhabitants. The individuals in the communes were randomly chosen and questioned by means of CATI. The communes also meet certain criteria (size, cultural-linguistic region, urban vs. rural) so as to represent the variety among Swiss communes. Due to missing values, the restriction to Swiss citizens, and the thorough tolerance measurement, our final sample comprises around 600 respondents.7

Our dependent variable is the degree of individual intolerance measured according to the rejection and acceptance components of intolerance. This conceptualization bears certain consequences for the measurement: to measure the degree of tolerance or intolerance one has to first ask the respondent about the group he or she likes the least in his or her country.8 Our data set “Schweizer Freiwilligen-Monitor Gemeinden” (Schweizerische Gemeinnützige Gesellschaft, 2011; Traunmüller et al., 2012) shows that around 42 percent of the respondents did not mention a single group that they found objectionable, whereas 58 percent expressed an objection toward a group.9 As we further restrict our analysis to intolerance toward immigrants we will solemnly focus on the around 30 percent of respondents that stated immigrants as their most objectionable group. In a second step, the respondent is asked to indicate which right or activity he or she would or would not extend to immigrants (see also Gibson, 1992). To capture the depth of intolerance toward immigrants we created an additive index covering three questions on the social rights the respondent would be willing to guarantee immigrants. These three questions ask whether an immigrant should be allowed to be a teacher at a local school or hold public office, and if the respondent would like to have a person of the objectionable group as neighbor or not (see appendix for the exact wording of the questions).10 Our additive index ranges from “0” to “3”; whereby “0” indicates that the respondent would grant an immigrant all three rights. This is identified as true tolerance: Although the respondent objects to a group, he or she nevertheless thinks they should be granted equal rights (see Gibson, 2005; Mondak and Sanders, 2003, 2005). In contrast, a value of “3” indicates denial of all three activities. In other words, higher values of the index imply deeper intolerance. Table 1 displays the distribution in our sample of the depth of intolerance toward immigrants in the Swiss municipalities. 31 percent of the respondents that mentioned immigrants as objectionable would bestow all three rights to them; 25 percent reject all three privileges. Most of the respondents deny one or two rights.

Table 1. Depth of Intolerance Toward Immigrants in Switzerland
depth of intolerancefrequencypercentage
all activities denied 26625.36
two activities denied 23922.78
one activity denied 21620.59
all activities permitted 32831.27

To capture the idea of the cultural dissimilarity thesis we asked respondents whether immigrants enrich or erode the Swiss culture (item ranges from 0 =  “erode Swiss Culture” to 10 =  “enrich Swiss culture”); we recoded this item so that higher values identify higher levels of perceived cultural threat. In order to measure the perceived economic threat, respondents were asked if they think immigrants threaten their job (items range from 0 =  “no threat at all” to 10 =  “high perceived threat”). To measure the strength of interethnic contact in terms of regular (true) or casual contacts, we use a battery of questions designed to gauge a person's interaction space regarding the contact with immigrants within one's group of friends, the neighborhood, the workplace, and civil associations. Social interaction with immigrants in each of these four social substructures is measured according to the intensity of contact (casual vs. regular). This thorough contact measurement enables us to adequately measure the intensity of contact and distinguish the social context in which contact takes place. Our models further control for other possible determinants of intolerance in the Swiss context (sex, age, education, religiosity, immigration background, ideology, and social trust). All these variables are commonly used in previous research on (in)tolerance (Cigler and Joslyn, 2002; Côté and Erickson, 2009; Gouws and Gibson, 1992; Ikeda and Richey, 2009; Kirchner et al., 2011; Mutz, 2002; Petersen et al., 2011). A more detailed description of all variables and their operationalization can be found in the appendix.

Regarding our conceptualization of tolerance in a rejection and acceptance component, we are confronted with a methodological challenge: our intolerance measurement leads to a non-randomly selected sample, as only those individuals who state immigrants as disliked group, are included in the outcome model (depth of intolerance) (see Mondak and Sanders, 2003). Following Sullivan et al. (1993) one solution to solve this problem would be to regress both the group selection and the depth of intolerance on the independent variables.11 However, two separate regressions cannot account for the possible omitted variable bias: We only observe the outcome variable (depth of intolerance toward immigrants) if the criterion of group selection (immigrants as objectionable) is met. Consequently, with this kind of procedure we would not control for factors that lead to this group selection (Vella, 1998). A powerful method to simultaneously overcome the selection problem as well as the omitted variable bias is to implement a two-step Heckman procedure (Heckman, 1979) as this procedure follows the assumption that the first step (the selection of immigrants) indirectly influences the results on the outcome stage (depth of intolerance toward immigrants). The obvious advantage is that although we estimate two separate models like Sullivan et al. (1993) we account for the results of the first stage in the outcome stage. Here, the first step compromises a probit model where the group selection (rejection component) is regressed on the socio-economic control variables, namely sex, age, education, and political ideology. Based on this probit model, the inverse Mill's ratio (IMR) is computed for each observation of the selected sample.12 This estimates the selection probability and will be integrated in the second step as an instrument to control for sample selection.13 In the second OLS-regressions only the inverse Mill's ratio and our main explanatory variables on threat and contact are integrated in the models.

Empirical Results

What drives intolerance toward immigrants in Switzerland? We ran several models to address our research question and to test our hypotheses on the influence of dissimilarity, threat, and contact on the formation of intolerance. All estimations are derived from two-step Heckman selection models due to the selection process underlying our research question (immigrants mentioned as objectionable group yes/no). We first ran a model estimating the selection probability; therefrom we calculated the inverse Mill's ratio. This ratio was then integrated in the outcome models to control for sample selection. Model 1 in Table 2 shows the results for the selection stage. In order to further test our intolerance conceptualization, we ran additional models on the selection stage including our threat and contact measures as well as their interaction effects (model 2 to model 5).

Table 2. Immigrants as an Objectionable Group (Selection Equation)
 model 1model 2model 3model 4model 5
 Est (s.e.)Est (s.e.)Est (s.e.)Est (s.e.)Est (s.e.)
  1. * p < 0.1, ** p < 0.05, *** p < 0.01

second level 0.324***0.329***0.331***0.1640.257**
third level 0.224**0.371**0.365***0.1570.295**
sex -0.023-0.010-0.0020.047-0.038
age 0.0020.0030.00*0.120***0.003
political ideology 0.118***0.077***0.077***0.095***0.070***
perceived economic  0.060***0.063***0.055**0.054***
threat  (0.01)(0.02)(0.02)(0.02)
perceived cultural  0.134***0.134***0.164***0.138***
threat  (0.02)(0.02)(0.02)(0.02)
contact with immigrants…      
…among friends  -0.066   
…among neighbors   0.012  
…at work    -0.023 
…within associations     -0.051
Intercept -1.266***-1.891***-1.963***-2.344***-1.727***
N 24352335233013111890

We see in model 1 that a person who leans more to the right politically as well as respondents with a medium or high level of education are more likely to perceive immigrants as objectionable.14 The results further reveal that age and gender do not have a relevant effect on group selection. In sum, a person who supports the political right and has a rather high level of education will be very likely to select immigrants as an objectionable group as opposed to any other group (e.g., unemployed, drug addicts, homosexuals) or no group at all. Concerning our main predictor variables (threat and contact) we can conclude that both economic and cultural threat have a relevant and significant influence on selecting immigrants as an objectionable group. In contrast, none of the four contact measures significantly influences the objection toward immigrants, indicating that contact with immigrants does not change the probability to state immigrants as a disliked group.

Our main interest, however, lies on the question in how far threat perceptions and social contacts relate to the depth of intolerance. Accordingly, we further estimate models testing the effect of perceived economic and cultural threat as well as their interaction with different contact measures on the level of intolerance (model 6 to model 13, table 3). In these outcome models all estimations contain further political ideology as control variable, and the inverse Mill's ratio controlling for the possible sample selection. The IMR was estimated on the basis of model 1 in Table 2. The results are presented in Table 3, whereby models 6 to 9 include economic threat and models 10 to 13 compromise the cultural threat measure. Furthermore our findings are visualized in figure 1 and 2.15 Turning to the variables of our main interest, the threat measures, the four contact variables as well as their interaction, we can conclude the following:

Table 3. Threat Perceptions, Social Contacts and Depth of Intolerance Toward Immigrants (Outcome Models)
  model 6 model 7 model 8model 9model 10model 11model 12model 13
  1. * p < 0.1, ** p < 0.05, *** p < 0.01

perceived economic 0.18***0.14***0.32***0.15***    
threat (0.04)(0.04)(0.06)(0.03)    
perceived cultural     0.17***0.16***0.25***0.19***
threat     (0.04)(0.03)(0.05)(0.03)
contact with immigrants…         
…among friends 0.068   0.07   
 (0.25)   (0.21)   
threat * contact among -0.06   -0.06   
friends (0.04)   (0.04)   
…among neighbors  -0.22   -0.23  
  (0.25)   (0.22)  
threat*contact among  -0.02   -0.02  
neighbors  (0.05)   (0.04)  
…at work   0.91**   0.53 
   (0.40)   (0.40) 
threat*contact at   -0.17***   -0.09 
work   (0.06)   (0.06) 
…within associations    -0.22   -0.01
    (0.23)   (0.24)
threat*contact within    -0.00   -0.04
associations    (0.05)   (0.05)
political ideology 0.03-**0.00-0.05-0.02-0.02
Inverse Mill's Ratio -0.75-1.16**-0.28-0.80-0.79-1.24**-0.84-0.77
Intercept 1.222.18**-0.211.621.492.39***1.091.52
N 614569478397622576483402
Figure 1.

Marginal Effects of Perceived Economic And Cultural Threat On Depth of Intolerance toward Immigrants

Figure 2.

Marginal Effects: Perceived Threat and Contact

To begin with, the economic and the cultural threat perceptions have a positive effect on intolerance if the respective contact measure equals zero (see table 3). Figure 1 depicts the marginal probability of economic and cultural threat on being completely intolerant versus being truly tolerant (grey and black solid line). The more respondents perceive immigrants as eroding the Swiss culture and threatening the economic situation, the more they tend to exhibit intolerance toward immigrants. Obviously, there is no difference regarding the effect of these two threat measures: both with higher levels of economic as well as cultural threat the probability of denying all three rights to immigrants rises up to 60 percent; whereas the probability of being truly tolerant declines to 10 percent for individuals feeling highly threatened. Against this background, it is undeniable that threat is one of the strongest factors fostering individual intolerance (Gibson, 1992; Sullivan et al., 1993).16

But how does social contact in distinct substructures affect the relationship between misgivings toward immigrants and the degree of intolerance? Recall that we expect more intense contact with immigrants to reduce the effect of perceived threat on intolerance, while casual contacts with immigrants are supposed not to change this relation in a relevant manner. Following these predictions, we estimated four interaction effects for each threat measure and our measures of contact intensity in the respective social substructures. For a better and more detailed interpretation of the results in Table 3, the marginal effects of contact on the relation between threat and intolerance are displayed in Figure 2. For our contact measures being a dichotomous variable the effects are presented for the casual contact as well as the regular contact category (hollow circles illustrate the marginal effects for casual contact, solid circles show regular contact).17

At first glance, we can see that in six out of eight instances regular contact reduces the strong effect of economic as well as cultural threat on intolerance more than casual contact as the black circles lie more left on the scale. More specifically, with regard to the results in Figure 2a, we see that particularly at the workplace regular social interaction may overcome existing prejudice and therewith reduce the effect of economic threat on intolerance (shown by the largest difference between the marginal effect of regular and casual contact).18 In contrast, having only monthly contact with immigrants seems to support existing prejudice resulting in a stronger effect of economic threat perceptions on intolerance. A similar but less strong and distinguishable effect can be found for social interactions in the neighborhood as well as among friends. For the case of associational contact and economic threat, however, the effects of casual and regular contact are reversed (but nearly identical).19

Turning to Figure 2b we observe analogous results: in general, compared to regular contacts casual social interactions result in higher probabilities of intolerance based on cultural threat perceptions. Most obviously, amongst one's friends the contact hypothesis comes into effect as enhanced regular social interaction may diminish the prejudice and hence lead to lower levels of intolerance. Most outstandingly, more regular contact with immigrants at the workplace are not able to encourage cultural perspective-taking. This finding is striking for three reasons: On the one hand, the workplace is the setting in which people spend most of their time (as compared to other contexts – one's circle of friends, neighborhood, and associations). On the other hand, the workplace is also the social context in which self-selection is at its lowest. Moreover, because working is not a voluntary activity for most people, the possibilities of avoiding exposure to dissonant views—to change the subject, to speak with someone else, or to just walk away—are quite limited. Finally, this finding is notable given that cultural differences clearly emerge in this era of increasing international cooperation. Increasing interaction within culturally heterogeneous teams weakens the tolerance toward immigrants due to the fears of losing one's own cultural identity. It is also worth pointing out that our results are independent of a respondent's profession: there is no pattern indicating that respondents with specific types of professions who have a lot of contact with immigrants at the workplace are deeply intolerant.

Finally, referring to our results, the question arises to what extent the differences between casual and regular contact are significant. Although all estimated marginal effects of the two contact categories are significant according to their standard errors (not shown here), we may not draw on any conclusion concerning the significance between these two effects.20 Here only the difference between regular and casual contact regarding the contact among friends and perceived erosion of Swiss culture reaches a value larger than 1.96; all other differences may not be considered as statistically significant. It has to be noted, however, that this significance is based on the sample size which is quite low for the case of the workplace or association in our analyses. Nevertheless, our results provide some tendency that regular contact with immigrants helps to overcome the negative impact that a perceived economic threat and cultural dissimilarity has on intolerance toward immigrants.


Looking at the latest political developments, we have many reasons to believe that the Swiss society has become increasingly intolerant toward immigrants in recent years. In this study we examined the role social substructures play in driving intolerance toward immigrants in Switzerland. While previous studies suggest that intolerance is rooted either in the institutional setting of a country (Peffley and Rohrschneider, 2003; Kirchner et al., 2011; Dunn et al., 2009) or altered by education, economic prosperity, and post-modern attitudes (Andersen and Fetner, 2008; Nevitte, 1996), we have argued that important sources of intolerance in Switzerland are to be found in the Swiss civil society. With respect to the core insights of the conflict and contact hypotheses, we can present the following findings:

  • – First, threat is traditionally seen as a necessary condition for intolerance (especially Gibson, 1992; Stouffer, 1955; Sullivan et al., 1993): the more threatened an individual feels by a certain group, the more intolerant he or she will be toward this group. In this respect, our findings provide further evidence for this key insight of the literature on intolerance formation. We find that the fear of the erosion of Swiss cultural values that accompanies increasing immigration is tied to intolerance toward immigrants. The same can be said regarding the threat that immigrants pose for the job market: Swiss who view rising immigration to mean a loss of economic privileges are less tolerant toward immigrants.
  • – Second, our results provide some evidence that in contrast to casual contacts regular interaction with immigrants helps to overcome the negative impact that a perceived economic threat and cultural dissimilarity has on tolerance toward immigrants. It has to be noted, however, that this effect varies depending on the interaction space and with regard to the kind of danger one relates to increasing immigration. In particular, compared to casual contacts, regular exchanges with immigrants among one's friends seems to show the strongest moderating effect on the relationship between perceived cultural erosion and intolerance. In contrast, increased contact with immigrants at the workplace is unable to mitigate the effect of perceived cultural erosion on intolerance. In this interaction space more contact with immigrants even strengthens the influence the perception of a vanishing Swiss culture has on intolerance.

We wish to underscore that our findings are only suggestive and explorative. Although a step in the right direction, we need more investigations that empirically scrutinize the role of different kinds contact in means of quantity as well as quality and competition in diverse contexts with different aspects of tolerance. According to insights from social psychology research by Aberson and Haag (2007) it does not suffice to distinguish only contact intensities (quantity): one should further discriminate the quality of contacts in means of negative and positive experiences an individual relates to contact situations in social substructures. By distinguishing contact quantity and quality they, for example, could find out that the quantity in means of contact intensity only exhibits a positive direct effect on perspective-taking, whereas it has no relevant effect on reducing anxieties and threat perceptions, respectively. In contrast, quality measures of contact may diminish perceived anxieties and therewith foster tolerance by reducing existing prejudice. Moreover, since there are salient differences between different groups of immigrants, in terms of ethnicity, religion, and class, for example, future research should strive to account for these differences. We are unfortunately unable to differentiate between immigrant groups due to the restricted availability of micro-level data in the present survey.

Finally, the problem of how to approach the arguments presented in a comparative perspective remains. On the one hand, one might argue that we cannot make any strong statements regarding our findings for cases other than Switzerland. On the other hand, the late Stein Rokkan once called Switzerland a microcosm of Europe due to its cultural, linguistic, religious, and regional diversity. Rokkan recommended that anyone wishing to study the dynamics of European politics should immerse him or herself in the study of Switzerland (Linder, 2010). In addition, Switzerland has been described as composed of three groups that “stand with their backs to each other” (Steiner, 2001, 145). Here, studies have shown that the three Swiss cultural-linguistic regions have more in common with their respective neighboring countries than with each other in terms of specific aspects of civil society and cultural life (Freitag and Stadelmann-Steffen, 2008; Kymlicka, 2003; Meier-Dallach et al., 1991). In other words, conclusions drawn from empirical analysis in Switzerland are likely to be valid for other countries or cultural contexts in Europe as well.


  1. 1

    Although one certainly cannot speak of one single Swiss culture (Boltanski and Chiva, 1996), we follow Wimmer (2002, 239) and refer to this term for reasons of simplicity.

  2. 2

    The aim of this initiative (Ausschaffungsinitiative) was to make it easier to deport criminal immigrants from Switzerland, even if they had permanent residency.

  3. 3

    Traditionally, tolerance is understood as political tolerance. Political tolerance thereby signifies permitting certain groups to be actively involved in political life, for example by taking part in elections or peaceful demonstrations (Sullivan et al., 1979). Tolerance, however, is not only related to political rights, but also to the toleration and acceptance of socio-cultural and socio-economic differences within a society (Weldon, 2006, 335). This kind of tolerance, “the willingness to live and let live, to tolerate diverse life styles and political perspectives,” is known as social tolerance (Norris, 2002, 158). It has to be noted, however, that these two concepts are difficult to separate, as both political and social tolerance are based on the idea of accepting groups and their underlying value system in form of a co-existence. Gibson (2006), for example, states that while political and social tolerance are independent of one another, he refers to the measurement of both concepts where groups are preselected according to their political or social nature. Nevertheless, in our analyses we will not differentiate between political and social tolerance, since in our underlying data set both political and social groups are included in the questionnaire (see appendix); we are thus compelled to combine both tolerance concepts.

  4. 4

    Within the psychological literature the “principle-implementation gap” also proposes a two-dimensional conceptualization of anti-immigrant attitudes, differentiating between attitudes and action (Dixon et al., 2007; Dovidio et al., 2011). In terms of our intolerance measurement this indicates that rejection is the attitude (e.g. prejudice) and acceptance is the action (e.g. granting political rights).

  5. 5

    The importance of differentiating between distinct types of threat is based on insights from both the social psychological literature as well as studies on tolerance (Gibson, 2006; Stephan and Stephan, 2000). It has to be noted that, according to these contributions, one could also distinguish further types of threat (socio-tropic, ego-tropic, symbolic, and realistic threat).

  6. 6

    Following Sullivan et al. (1993, 4), tolerance is not defined by the mere absence of prejudice. As stated above, tolerance is defined by the rejection, which could be defined as a prejudice, but also by the acceptance component, which makes our intolerance conceptualization distinguishable from xenophobia, prejudice and ethnocentrism.

  7. 7

    As we are restricting our analysis to intolerance toward immigrants, it would not make sense to include non-Swiss citizens in the sample, as they themselves are immigrants. We therefore attempt to fulfill the requirements for a content controlled measurement (Sullivan et al., 1993).

  8. 8

    The respondent was asked whether he or she objects to a group in society (open question, no groups were given or read out to the respondent). The following groups were the five most often mentioned groups by respondents: immigrants, juveniles, Muslims, marginal groups, and extremists. Referring to the philosophical branch of tolerance research, not all groups can be handled equally (e.g., Forst, 2003; Peffley and Rohrschneider, 2003; Petersen et al., 2010). Some groups, such as criminals, drug addicts, or extremists, are objectionable due to their mere nature, as they tend to be involved in illegal acts. For instance, if a person with an immigrant background is intolerant toward right-wing extremists this person may not be easily declared as intolerant, for he or she is showing intolerance toward an intolerant group. This phenomenon is called the paradox of tolerance, which identifies a lack of tolerance toward intolerant groups/persons (reasonable intolerance). Groups that express intolerance toward people and groups involved in criminal activities are thus more likely not to be tolerated by others. Similarly, Peffley and Rohrschneider (2006, 247) exclude criminals from their analysis, arguing that “criminals do not enjoy the same citizenship rights as non-criminals.” Against this background, we refrain from including respondents that mention criminals, extremists, marginal groups, or drug addicts in our analysis and have coded them as missing values.

  9. 9

    The data set is available on request on: , FORS-ID: 10905.

  10. 10

    The frequency of denying or allowing one of these activities is equally distributed, meaning that none of the rights is more frequently denied than the others. For all three questions the distributions show that about half of the respondents answered them with yes, and the other half with no. With regard to the scale reliability for these three items, a Cronbach's alpha of 0.73 indicates a very decent value.

  11. 11

    Sullivan et al. (1993, 250) show in separate analyses for the rejection and acceptance component that socio-economic variables (sex, age, education, race, class, etc.) are assumed to purely affect the probability of selecting an objectionable group, but not the assignment of societal rights to the least-liked group. According to them, the depth of (in)tolerance is particularly influenced by variables identifying certain political or psychological variables.

  12. 12

    The inverse Mill's ratio is computed according to the following equation: math formula, it estimates the ratio of the probability density function over the cumulative distribution function of a (truncated normal) distribution (Heckman, 1979).

  13. 13

    The selection problem is based on missing values on the dependent variable (depth of intolerance toward immigrants) that are conditional on certain independent variables and thus lead to a biased inference (Heckman, 1979).

  14. 14

    This is in line with Helbling (2011) and Giugni and Morariu (2010, 93) who point out that education does not systematically improve attitudes toward immigrants in Switzerland.

  15. 15

    With regard to the controls, the political ideology variable does not show any significant or relevant effect on the depth of intolerance. Moreover, the inverse Mill's ratio is unequal to zero indicating a sample selection and thus emphasizing the necessity of implementing a two-stage Heckman model.

  16. 16

    Moreover, additional analyses not shown here could show that socio-economic variables do not have any significant or relevant influence on the depth of intolerance. Excluding the interaction effects from the analyses does not change the effects of threat and contact. Their direct influences on the depth of intolerance are robust and highly significant. Including the interaction effects in the selection equation has no significant or relevant effect. Again, these results underline the necessity of a two step-differentiation with regard to the measurement of tolerance.

  17. 17

    Due to the number of cases in each contact situation and collinearity problems, all interaction terms were estimated separately.

  18. 18

    It is also worth pointing out that our results are independent of a respondent's profession: there is no pattern indicating that respondents with specific types of professions who have a lot of contact with immigrants at the workplace are deeply intolerant.

  19. 19

    Since Côté and Erickson (2009, 1677) argue that different types of associations exert diverse effects on tolerance, we should further scrutinize the impact of social contact with immigrants within associations by separately estimating the interaction effect for three kinds of voluntary organization. However, the associations covered by the survey did not have enough respondents (<100) to run a Heckman selection model.

  20. 20

    To estimate the significance of the difference between the two contact intensities we calculate a t-test according to the following formula: math formula; whereby t-values larger than +/- 1.96 indicate significance at the 95%-level.


Table A1

VariableOperationalization and sourcea
  1. a

    All variables are retrieved from “Schweizer Freiwilligen-Monitor Gemeinden”.

intolerance Intolerance is measured in two steps according to its twofold nature of the rejection and acceptance component. First, the respondent is asked if there is any group in society he or she feels objectionable to. [Sd20N6.1] There were no categories/groups mentioned to the respondent (open question). As soon as the respondent mentioned a group he/she feels objectionable to, he/she fulfills the necessary condition of rejection (selection yes=1; no=0).

In a second step, the depth intolerance is captured in an additive index from the following three questions:

What would you say: Would you like a person from the group you mentioned above to…

  • – be a teacher at your local school? [Sd20N6.2a]
  • – be your neighbor? [Sd20N6.2b]
  • – hold public office? [Sd20N6.2c]

all variables were coded so that 1 corresponds to no, and 0 corresponds to yes. The additive index ranges from 0 =  all rights permitted (true tolerance) to 3 = all rights denied (deep intolerance).

perceived economic threat

Economic threat is measured according to the question whether the respondent thinks that immigrants coming to Switzerland take away jobs from the Swiss [Q07.N4.1].

The scale ranges from 0 “immigrants take away jobs” to 10 “immigrants help creating new jobs”. The scale was reversed so that higher values indicate greater perceived economic threat.

perceived cultural erosion/enrichment

Cultural erosion/enrichment is measured according to the question whether the respondent thinks that immigrants coming to Switzerland erode/enrich the Swiss cultural way of life [Q07.N4.2].

The scale ranges from 0 “immigrants erode Swiss culture” to 10 “immigrants enrich the Swiss culture”. The scale was reversed so that higher values indicate greater perceived cultural erosion.

contact with immigrants

How often do you contact with immigrants/immigrants…?

This question was asked for the following social contexts:

  • – contact with immigrants among friends [SD20N2.1]
  • – contact with immigrants at the workplace [SD20N2.2]
  • – contact with immigrants within association [SD20N2.3]
  • – contact with immigrants in neighborhood [SD20N2.4]

Contact is measured according to these three categories: 1 = “once a month”, 2 = “several times a months”, 3 = “at least once a week”. Category 2 and 3 were combined to the general category of “regular contacts” and category 1 depicts the measure for “casual contacts”.

gender Dummy variable with 0 =  male and 1 =  female [Sd02].
age Can you tell me your year of birth, please? 19__ [Sd01]. Recoded in years old (2010-year of birth)
educational level Recode of Sd21 in three main categories: 1 = low education, 2 = medium education, 3 = high education.
immigration background Categorical variable indentifying respondents’ ethnic background [Sd05N4; Sd05N5]: 0 = father and mother Swiss, 1 = father or mother non-Swiss, 2 = father and mother non-Swiss
religiosity How often do you take part in religious services? [Sd22.2]; variable ranges from 1 = never to 8 = several times a week.
political ideology Where on the political scale from 0 = totally left to 10 = totally right would you place yourself? [Q07]


  • Markus Freitag is Professor of Political Sociology and Director of the Institute of Political Science at the University of Bern. His research interests include direct democracy, social capital, tolerance, political participation, and volunteering. Address for correspondence: Institute of Political Science, University of Bern, Fabrikstrasse 8, CH-3012 Bern, Switzerland. Phone: +41 (0)31 631 4685; Email:

  • Carolin Rapp is a PhD candidate at the Chair of Political Sociology, University of Bern. Her research interests include tolerance research, moral politics, and research on ethnic minorities. She just finished her dissertation on tolerance in Switzerland and Europe. Address for correspondence: Institute of Political Science, University of Bern, Fabrikstrasse 8, CH-3012 Bern, Switzerland. Phone: +41 (0)31 631 5362; Email: