Mechanisms of exclusion: attitudes toward allocation of social rights to out-group population

This study examines the ways in which perceived socioeconomic threat, perceived threat to national identity, and prejudice (as reflected in negative stereotypes and desire for social distance) prompt objection to allocation of rights to out-groups. The paper presents a simultaneous test of three theoretical explanations and delineates the complex interrelationships among all concepts, using data from a representative sample of Israelis in structural equation models. This study finds that: (1) perception of socioeconomic threat is positively associated with perception of threat to national identity; (2) the two types of threat do not directly affect attitudes toward allocation of rights, but instead tend to increase negative stereotyping of the out-group; (3) negative stereotypes are likely to contribute to a desire for social distance from the out-group; and (4) the attribution of negative characteristics and a sense of social distance tend to increase objection to allocation of social rights to the out-group. Keywords: Exclusionary attitudesimmigrationperceived threatprejudicelabour migrantsIsrael Acknowledgements The author thanks Moshe Semyonov, Noah Lewin-Epstein, Peter Schmidt and anonymous reviewers for helpful comments and suggestions. Notes 1. One notable exception is the study by Halperin, Canetti-Nisim and Pedahzur (2007); however, the research was carried out using purposive sampling procedure and was not based on a representative sample of Israeli adults. 2. The survey was conducted by the B. I. Cohen Institute, Tel Aviv University. Information was obtained through face-to-face interviews in respondents’ homes, lasting on average forty minutes. The response rate for the survey was 57.5 per cent. 3. Due to data limitations, I use a single-item measure to capture the perceived threat to national identity while measuring all other key variables in the study by latent variables with multiple indicators. It is important to note that a single-item measure is less capable of capturing complex social phenomena than multi-item measures and may suffer from possible reliability issues. 4. Although fully determining the causal relations using non-longitudinal data is somewhat problematical, I estimate the models under the premises that have been developed from well-established theoretical arguments on causal relations between: (1) perceived threat and either exclusionary attitudes or prejudice; and (2) prejudice and exclusionary attitudes. It should also be noted that Schlueter, Schmidt and Wagner's (2008) study provides longitudinal evidence that perceptions of threatened group interests are causally antecedent to out-group derogation as suggested by the conventional ‘competitive threat’ theoretical model. No support is found for the assumptions of the reverse or of the reciprocal model. Specifically, the study reveals that heightened levels of perceived group threat lead to heightened levels of dislike, ethnic distance and intended negative behaviour toward foreigners. 5. Neither threat exerts a significant effect on exclusionary attitudes, even without introducing controls for socio-demographic variables in the models. 6. The effects of socio-demographic variables on each one of the attitudinal variables have been introduced into the final general model only if these effects were found significant in previous models that examined each one of the attitudinal variables separately. In order to reach this final general model, I started with the preliminary model that included only perceived SETHREAT and socio-demographic characteristics. At each step I added one additional variable to the model and the relevant correlations between this variable and those already included in the model. Throughout the process I only introduced into the new models paths (causal correlations) that were found significant in the previous steps of the research. I used this technique to avoid statistical artifact as a result of a large number of variables and causal correlations. 7. Whereas the full model includes all possible paths, all non-significant effects are eliminated from the trimmed models. Thus, the model includes only paths that represent significant effect. 8. SETHREAT and NTHREAT are interrelated. The limitation of cross-sectional data does not allow me to implement an appropriate test for causal relationship between two concepts. With the data in hand, I estimated a non-recursive model (Kline Citation2005) including SETHREAT and NTHREAT; the results reveal a statistically significant path from SETHREAT to NTHREAT, but a statistically insignificant path from NTHREAT to SETHREAT. This finding is in line with Raijman and Semyonov's (2004) results. Following that, I decided to introduce only one path from SETHREAT to NTHREAT to the general model. However, since I do not have longitudinal data and there is no strong theoretical assumption about the relationship between the two types of threat, it is important to be cautious with the interpretation of the path from SETHREAT to NTHREAT. As to the relationship between two components of prejudice (STEREOTYPES and SDISTANCE), the assumption of previous studies on the topic is that attributing negative characteristics to the out-group takes place before a desire to maintain social distance from them, and therefore negative stereotyping contributes to the desire for social distance (e.g. Verberk, Scheepers and Felling Citation2002; Bobo and Charles Citation2009). I estimated the general model under this premise, while introducing a path from STEREOTYPES to SDISTANCE. In addition, I estimated a non-recursive model including STEREOTYPES and SDISTANCE; the results reveal a statistically significant path from STEREOTYPES to SDISTANCE, but a statistically insignificant path from SDISTANCE to STEREOTYPES.