Abstract
Immigration and the ways in which host societies receive newcomers pose challenges for modern civil societies. This article contributes to the ongoing discussion about how ethnic diversity influences tolerance towards immigrants. Compared to previous studies, we analyse tolerance as a sequential concept in order to uncover the effects of contextual diversity on attitudes towards immigrants and the granting of certain rights to this group. Moreover, we distinguish different shares of ethnic groups based on their ethnic and cultural origins both on the independent and dependent variable. The analysis relies on a subnational survey of sixty municipalities in Switzerland, revealing that only certain ethnic groups are seen as an economic and cultural threat.
Disclosure statement
No potential conflict of interest was reported by the author.
Notes
1. It has to be noted, however, that we only test the mere exposure effect and not the actual individual interaction between members of diverse ethnic groups (see Laurence Citation2014).
2. 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 by respondents: immigrants, juveniles, Muslims, marginal groups and extremists.
3. 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. With regard to the scale reliability for these three items, a Cronbach's alpha of 0.73 indicates a very decent value.
4. For the sixty Swiss municipalities, the diversity measure based on the Herfindahl index is highly correlated with the measure for the share of immigrants (ρ = 0.95). For this reason, only the share of immigrants is integrated in the analyses. Some may further argue that the change in the share of immigrants is a more important predictor than the actual share. Further analyses (available on request) demonstrate that the change in the share of the overall share of immigrants has no significant influence on the rejection and acceptance of immigrants in Swiss municipalities.
5. Regarding our conceptualization of tolerance in a rejection and acceptance component, some may argue that we are confronted with a methodological challenge: our tolerance measurement leads to a non-randomly selected sample, as only those individuals who state immigrants as a disliked group are included in the outcome model (depth of tolerance). We estimated additional Heckman selection models (available on request) to test for this selection bias. These models account for the possible omitted variable bias as we only observe the outcome variable (depth of tolerance) if the criterion of group selection (immigrants as objectionable) is met. However, the estimated values for the inverse Mills ratio, which accounts for sample selection on the first stage, shows no significant selection bias in our models. The values for the three tolerance measures are: b = 0.49 z-value = 0.76 for the general immigrants model; b = −2.32 z-value = 1.17 for the model with immigrants from the Balkan states; b = 1.17 z-value = 1.20 for the model with Muslims. As all three values are non-significant, we cannot make any conclusions from their direction and their general meaning.
6. The question may arise of whether contextual diversity truly implies a potential threat as mentioned in the theory section. To test this assumption, we regressed the perceived cultural and economic threat, respectively, on the different contextual diversities. The estimates are depicted in . The results show that the share of immigrants has a significant negative effect on economic as well on cultural threat, underscoring the considerations of the contact theory. In sharp contrast, the shares of immigrants from the Balkan States as well as the share of Muslims exhibit a positive and significant effect on cultural threat perceptions, but have no effect on economic threat perceptions.
7. In further analyses (available on request) we tested some possible interactions between the different diversity measures and socio-demographic factors (sex, age, education) as some of these effects may only be applicable to individuals with specific characteristics. The results, however, revealed no significant or relevant interaction between these factors.
Additional information
Notes on contributors
Carolin Rapp
CAROLIN RAPP is Post-Doctoral Researcher at the Institute of Political Science at the University of Bern.