Edited by: Leonardo Carlucci, University of Studies G. d’Annunzio Chieti and Pescara, Italy
Reviewed by: Gary Bouma, Monash University, Australia; Renee Zahnow, The University of Queensland, Australia
This article was submitted to Personality and Social Psychology, a section of the journal Frontiers in Psychology
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
High levels of social cohesion have been shown to be beneficial both for social entities and for their residents. It is therefore not surprising that scholars from several disciplines investigate which factors contribute to or hamper social cohesion at various societal levels. In recent years, the question of how individuals deal with the increasing diversity of their neighborhoods and society as a whole has become of particular interest when examining cohesion. The present study takes this a step further by combining sociological and psychological approaches in investigating whether the
Growing diversity is a fact in many societies today. In public discourse, migration is usually identified as the main driver for diversification. Continuous global migration movements since the end of the Cold War, and even more so since the beginning of the 21st century (
It is therefore not surprising that the number of studies investigating the impact of diversity on social cohesion, particularly from the fields of political science and sociology, has increased significantly in recent years. The starting point for much of this research is the seminal work of
In examining the reasons for these inconclusive results, many studies bear at least one of three major issues: an oversimplified operationalization of social cohesion, a focus on objective characteristics of diversity (e.g., the degree of diversity in a certain community) instead of investigating individual (subjective) perceptions of it, and finally a narrow understanding of diversity as referring to ethnic diversity only. We discuss these points in the following, as they constitute the gaps in the literature addressed by the present paper, and then go on to mention individual characteristics that have proven to be important for understanding acceptance of diversity, namely people’s intergroup anxiety (IGA), their level of empathy, and their political orientation (PO). However, the literature that we discuss investigates this link on the individual level only. How such individual attributes impact the acceptance of diversity on the societal level is still largely unclear. To understand the relationship between emotions, attitudes, and behaviors of the individual and its social environment is, however, of utmost importance when one wants to understand processes of change on the macro level.
The first issue identified above refers to the operationalization of social cohesion. Putnam and many others use trust in various groups as a proxy for social capital and cohesion (
One of the more influential attempts to address this issue has been the
Instead of focusing on the objective level of diversity as a factor influencing social cohesion – as it is broad practice in the literature – the SCR incorporates acceptance of diversity as a dimension of social cohesion. Thereby, the individual’s handling of societal heterogeneity as “negotiated difference” (
Most research that focuses on the impact diversity can have on the individual and which individual preconditions influence this relationship stems from social psychology. Those studies are based on variations of conflict theory (e.g.,
Although the overall association between increasing levels of diversity and decreasing social cohesion seems to be rather robust, various studies found individual characteristics to affect this relationship.
The third point of critcism addresses the fact that in public discourse there is often a narrow understanding of diversity. The work by Putnam and others on factors affecting social cohesion mirrors the public debate as it predominantly focuses on ethnic diversity as well as linguistic and religious diversification (c.f.
Sociology calls this development the individualization and pluralization of life forms (
In order to address these gaps in the literature, we intend to investigate whether the acceptance of diversity (according to a multidimensional measure) in a given entity is primarily a consequence of the social structure or a matter of mentalities prevailing among residents of that entity. Intergroup anxiety and individual empathy will receive special attention, as will political orientation. As mentioned above, this paper is interested in the impact of individual level characteristics on the group-level acceptance of diversity. However, research usually focuses on one level only. Particularly, the link between individual-level intergroup anxiety and acceptance of diversity is rather established. Based on the aforementioned work by
While intergroup anxiety can be seen as a motor of the rejection of others, empathy has been found to function as a safeguard against the rejection of otherness. Being considered a relatively stable personality trait related to agreeableness (
Furthermore, there is research on the relationship of people’s political orientation and outgroup attitudes and behavior. Right-wing orientations have been found to correlate with a greater need for stability and security, as well as the wish to maintain the status quo, particularly in Western Europe (
In conclusion, the current research proposes that the acceptance of diversity in a community is central for the level of social cohesion in a given entity. The pressing question that arises from this observation is which factors contribute to acceptance (or rejection) of diversity in a given social entity. Psychological research in this area usually focuses on individual preconditions that affect individual-level acceptance of otherness, but does not address their impact on community-level acceptance of diversity. The present study aims at overcoming this gap. While taking a clearly social psychological stance, its goal is to investigate if the acceptance of diversity in a social entity is related to prevailing mentalities of individuals once the social structure of a community is accounted for. By combining macro and micro level approaches, this paper not only adds to social psychological literature on the acceptance of diversity, but also addresses pressing sociological questions about the influence of citizens’ emotions, attitudes, and behaviors on life in diverse societies. Thereby, the paper aims at making a contribution relevant also to policy makers and actors in civil society. Unlike most studies in psychology, the reported research relies on representative survey data from the 16 federal states of Germany, thereby offering higher levels of generalizability of its findings than the usual convenience sample approach dominating much of the work on stereotypes and prejudice (
The data used for our analyses stem from the research project
The goal of the research project
In order to ensure a representative sample at both the national and federal state level, weighting was applied in the analysis which took into account: (1) the probability of being selected into the sample according to the dual frame approach and the disproportionate sampling; (2) scaling so that the sum of the weights is equal to the number of realized interviews (
The level of intergroup anxiety was measured using a modified version of an instrument suggested by
Descriptive information on variables used.
Mean | SD | |||
---|---|---|---|---|
ln(per capita GDP) | 10.44 | 0.22 | 10.06 | 10.98 |
Acceptance of diversity | 67.29 | 3.76 | 61.49 | 72.30 |
Intergroup anxiety | 4.29 | 1.61 | 0 | 10 |
Empathy | 4.15 | 0.66 | 1 | 5 |
Political orientation | 4.41 | 2.11 | 0 | 10 |
Age (years) | 51.34 | 17.81 | 16 | 97 |
Female | 0.51 | 0.01 | 0 | 1 |
Marital status: married or cohabitating | 0.63 | 0.01 | 0 | 1 |
Marital status: divorced, separated or widowed | 0.18 | 0.01 | 0 | 1 |
Marital status: single | 0.21 | 0.01 | 0 | 1 |
Education: University degree | 0.34 | 0.01 | 0 | 1 |
Education: vocational training | 0.51 | 0.01 | 0 | 1 |
Education: no completed secondary education | 0.13 | 0.01 | 0 | 1 |
Employed | 0.61 | 0.01 | 0 | 1 |
Migration background | 0.14 | 0.01 | 0 | 1 |
SD, standard deviation
Example indicators used for measuring the seven dimensions of acceptance of diversity.
Dimension of cohesion | Example indicator |
---|---|
Age | I do not get along very well with people who are clearly younger or older than me. |
Disability | I find many demands from disabled people to be exaggerated. |
Gender | I am against quotas for women. |
Sexual orientation/gender identity | Transgender people should stay to themselves. |
Cultural otherness/ethnicity | If I had a choice, I would rather have nothing to do with people from other countries. |
Religion | Islam also fits into the Western world (−). |
Economic disadvantage | We take too much care of people who are failures in our society. |
A complete list of the 23 indicators can be found in Table 18 of
Modified versions of two subscales of the Interpersonal Reactivity Index (
Political orientation was measured by asking participants to locate themselves on a spectrum of political beliefs from 0 (“Left”) to 10 (“Right”;
Our analyses use the Bosch Diversity Barometer (BDB) score calculated by the
The BDB was calculated using the data as described in “The Data Set.” The first step involved exploratory factor analysis and internal consistency checks of all survey items related to the proposed seven dimensions of acceptance of diversity. The items included statements such as “Islam also fits into the Western world.” And “If I had a choice, I would rather have nothing to do with people from other countries.” Participants were asked for their agreement on a scale from 1 (“Do not agree at all”) to 4 (“Fully agree”). In order to simplify further analyses and allow for easier interpretation of the BDB by the general public, all items were transformed to a 100-point scale. For an item to be selected as a measure of a particular dimension, it had to meet an absolute factor loading of 0.40 or greater. This resulted in the selection of a total of 23 items, with three to four items per dimension
Sociodemographic characteristics of participants were controlled for in the models. These included age (
Several structural variables were initially selected for our analysis based on their significant partial correlation with federal state scores on the BDB after controlling for per capita gross domestic product (GDP), according to initial results by
We are interested in the relationship between individual levels of intergroup anxiety, empathy, and political orientation and federal state levels of acceptance of diversity, while also accounting for the structural properties of the federal states individuals live in. While multilevel models regressing individual-level outcomes on both individual- and group-level predictors (i.e., macro-micro models) have become quite standard in the social sciences for analyzing processes that occur in hierarchical systems (e.g., individuals nested within groups), there is much less of a consensus on the best analysis method for regressing group-level outcomes on individual-level predictors (i.e., micro-macro models;
We therefore choose to follow the two-step approach outlined by
In step one, we fit a series of three multilevel linear regression models each for intergroup anxiety, empathy, and political orientation using the
In step two, we fit another series of three group-level linear regressions for each micro-level predictor of interest with Stata 16’s
In step one, the null models (M01) for intergroup anxiety, empathy, and political orientation are intercepts-only models (see
Step one: multilevel empty model regressions of intergroup anxiety, empathy, and political orientation: Model 01.
Predictor | IGA | Empathy | PO |
---|---|---|---|
Intercept | 4.33 |
4.15 |
4.38 |
Federal state variance | 0.05 | 0.00 | 0.11 |
Individual variance | 2.56 | 0.43 | 4.38 |
AIC | 10,780 | 5,688 | 12,301 |
Random intercept model. IGA, intergroup anxiety; and PO, political orientation.
All regression coefficients are unstandardized. Standard errors in parentheses.
In the next model (M11), the group-level covariate of per capita GDP was added, indicating only a significant (negative) effect of per capita GDP on individual levels of intergroup anxiety (see
Step one: multilevel regressions of intergroup anxiety, empathy, and political orientation on macro-level covariate: Model 11.
Predictor | IGA | Empathy | PO |
---|---|---|---|
Intercept | 11.35 |
3.36 |
5.68 (4.32) |
(ln)per capita GDP | −0.67 |
0.08 (0.08) | −0.12 (0.42) |
Federal state variance | 0.02 | 0.00 | 0.11 |
Individual variance | 2.57 | 0.43 | 4.38 |
AIC | 10,774 | 5,689 | 12,302 |
Random intercept model. IGA, intergroup anxiety; PO, political orientation; and AIC, Akaike information criterion.
All regression coefficients are unstandardized. Standard errors in parentheses.
Step one: multilevel regressions of intergroup anxiety, empathy, and political orientation on macro- and micro-level covariates: Model 21.
Predictor | IGA | Empathy | PO |
---|---|---|---|
Intercept | 10.31 |
3.33 |
4.91 (4.15) |
(ln)per capita GDP | −0.59 |
0.07 (0.07) | −0.05 (0.41) |
Intergroup anxiety | −0.04 |
0.31 |
|
Empathy | −0.22 |
−0.35 |
|
Political orientation | 0.18 |
−0.03 |
|
Age (years) | 0.06 |
0.02 |
−0.03 (0.02) |
Age (squared) | −0.00 |
−0.00 (0.00) | 0.00 (0.00) |
Female | 0.14 (0.08) | 0.23 |
−0.15 (0.17) |
Marital status: married or cohabitating | Ref. | Ref. | Ref. |
Marital status: divorced, separated or widowed | −0.00 (0.09) | −0.08 |
−0.28 |
Marital status: single | −0.08 (0.12) | −0.04 (0.05) | −0.37 |
Education: University degree | Ref. | Ref. | Ref. |
Education: vocational training | 0.52 |
−0.02 (0.03) | 0.18 (0.12) |
Education: no completed secondary education | 0.35 |
−0.04 (0.07) | 0.06 (0.16) |
Employed | −0.31 |
0.02 (0.03) | 0.16 (0.15) |
Migration background | 0.20 |
0.07 (0.04) | −0.08 (0.10) |
Federal state variance | 0.03 | 0.00 | 0.09 |
Individual variance | 2.26 | 0.40 | 3.99 |
AIC | 10,437 | 5,497 | 12,060 |
Random intercept model. IGA, intergroup anxiety; PO, political orientation; AIC, Akaike information criterion; and Ref., Reference group.
All regression coefficients are unstandardized. Standard errors in parentheses.
In step two, the first models (M12) predict acceptance of diversity using the group-level residuals from the multilevel null models. This accounts for the variance of the individual levels of intergroup anxiety, empathy, and political orientation, leaving only a negative significant relationship between intergroup anxiety and acceptance of diversity (see
Step two: linear regression of acceptance of diversity on residuals of null model (M01) on intergroup anxiety, empathy, and political orientation: Model 12.
Predictor | IGA | Empathy | PO |
---|---|---|---|
Intercept | 67.29 |
67.29 |
67.29 |
Residuals (null model, M01) | −15.72 |
3.42 (24.76) | −3.15 (3.10) |
AIC | 72.54 | 92.69 | 91.71 |
BIC | 74.86 | 95.01 | 94.03 |
IGA, intergroup anxiety; PO, political orientation; AIC, Akaike information criterion; and BIC, Bayesian information criterion.
All regression coefficients are unstandardized. Standard errors in parentheses.
Step two: linear regression of acceptance of diversity on residuals of M11 on intergroup anxiety, empathy, and political orientation: Model 22.
Predictor | IGA | Empathy | PO |
---|---|---|---|
Intercept | 67.29 |
67.29 |
67.29 |
Residuals (Model 11) | −16.70 |
−13.37 (26.24) | −2.41 (3.14) |
AIC | 87.66 | 92.46 | 92.14 |
BIC | 89.97 | 94.78 | 94.45 |
IGA, intergroup anxiety; PO, political orientation; AIC, Akaike information criterion; and BIC, Bayesian information criterion.
All regression coefficients are unstandardized. Standard errors in parentheses.
Step two: linear regression of acceptance of diversity on residuals of M12 on intergroup anxiety, empathy, and political orientation: Model 32.
Predictor | IGA | Empathy | PO |
---|---|---|---|
Intercept | 67.29 |
67.29 |
67.29 |
Residuals (Model 12) | −12.88 |
−12.94 (27.14) | −2.60 (3.34) |
AIC | 88.62 | 92.49 | 92.12 |
BIC | 90.94 | 94.80 | 94.44 |
IGA, intergroup anxiety; PO, political orientation; AIC, Akaike information criterion; and BIC, Bayesian information criterion.
All regression coefficients are unstandardized. Standard errors in parentheses.
Growing diversity has become a reality in the majority of Western societies since World War II. Research interested in its impact on the individual as well as on the society mostly focuses on ethnic, cultural, and religious variables. However, this narrow conception does not reflect the actual breadth of diversity in most communities. While many see increasing diversity as a chance for growth, development, and innovation, others are worried about the erosion of social cohesion and the loss of defining elements of their society, such as established customs and social rules. It is indicative therefore to go beyond ethnic origin and religion when we speak about diversity in the proverbial WEIRD societies, and include other relevant aspects. Only if we take this multidimensionality into account, is it possible to understand the mechanisms that foster or hamper the acceptance of diversity in a given community. Do individual factors play a role alongside structural variables? By building a bridge between psychological and sociological approaches, we sought to answer the question of whether
In designing our study, we took into account some of the shortcomings of prior research. First, instead of operationalizing diversity only in ethnic or cultural-religious terms, we used a multidimensional conceptualization to measure acceptance of diversity in a broader understanding, including the following seven dimensions: age differences, disability, gender, sexual orientation and gender identity, ethnicity, religion, and economic disadvantage (c.f.
How do the results of the present study contribute to understanding which factors play a vital role in the acceptance of diversity today? Firstly, as hypothesized, we found that when study participants express higher levels of intergroup anxiety, their German federal states of residence exhibit lower levels of diversity acceptance. This fits with prior psychological research (
Finally, we obtained two unexpected results. Contrary to our hypothesis, we could not confirm the expected positive effect of individual empathy on acceptance of diversity on the aggregate level. At first glance, this result is puzzling since research consistently shows that empathy does have a positive effect on individual attitudes and behavior toward others (
The second unexpected finding refers to the political orientation of our study participants. Although previous research is scarce and provides inconclusive results, studies – particularly for the German context – have indicated that right-leaning political orientations go hand-in-hand with higher intergroup anxiety (
The results of our analyses make an important contribution to the literature and help to frame the work of practitioners on a community level. First and foremost, our study underscores the central role intergroup anxiety plays for the acceptance of diversity. Although this finding is generally not new, investigating this link from a mixed individual-structural perspective puts another emphasis on the relevance anxiety plays in intergroup relations as well as on its strongest antidote: intergroup contact. Research consistently shows that contact helps to lower perceived intergroup threat (
However, promoting intergroup contact is a challenge in many communities because resources – both monetary and in terms of personnel – are usually scarce. As our analyses show, this is particularly problematic for communities that are already disadvantaged, since lower economic status of a community increases intergroup anxiety and therefore exacerbates the problematic link between intergroup anxiety and diversity acceptance. At the same time, empathy training, as for example offered by Roots of Empathy, an originally Canadian civil society organization (
Besides its conceptual and methodological strength, the present study comes with various limitations. First, in being a case study of Germany its generalizability is limited, particularly since previous research has found the German context to differ from other European settings or the United States (
Ensuring and improving social cohesion have become important goals in public discourse, in political decision-making, and also in academic work. In the search for relevant factors that influence social cohesion, diversity has crystallized to be one of the central variables both on a structural as well as on an individual level. Since increasing diversity is a fact in Western societies, this paper argued that identifying factors that contribute to high vs. low levels of subjective acceptance of diversity is much more fruitful than focusing on the objective level of diversity in social entities. By taking a multilevel approach, our study showed that individuals’ intergroup anxiety is the key to diversity acceptance in the community. Our approach also takes into account the level of economic prosperity in the community. Surprisingly, neither individual levels of empathy nor political orientation played a role. Although at first sight, the partial “non-findings” of our study may seem disappointing, our research carries a clear message that cannot be emphasized enough: fear is the biggest enemy of the acceptance of others. Whereas empathy and political orientation are often seen as rather stable personality attributes that are difficult to change, there is a well-proven remedy for intergroup anxiety: “Intergroup
The multilevel dataset used for the analysis, as well as the data dictionary are accessible for the scientific community through the Open Science Framework and can be downloaded using following link:
Ethical review and approval was not required for the study on human participants in accordance with the local legislation and institutional requirements. Written informed consent from the participants’ legal guardian/next of kin was not required to participate in this study in accordance with the national legislation and the institutional requirements.
KB, RA, and ML contributed to the conception and design of the study. RA and ML conducted the literature search and wrote the first draft of the manuscript. ML performed the statistical analysis. KB contributed with his expertise in the research field. RA and KB designed the data collection instruments. RA coordinated and supervised data collection. All authors contributed to the article and approved the submitted version.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The article was prepared within the framework of the HSE University Basic Research Program. We thank Caroline Schnelle for her tireless assistance in digging up the pertinent social science literature, and Dr. Georgi Dragolov for his invaluable help with data handling.
1Example items for the seven dimensions of the BDB can be found in
2Cronbach’s alpha of internal consistency ranged from 0.37 for “disability” with three items, to 0.79 for “sexual orientation and gender identity” with four items. By making use of the application of the Spearman-Brown formula (pp. 223–226) of
3The multilevel structural equation modeling approach, which replaces manifest with latent aggregation for micro-level predictors, is another possible approach. However, due to the small number of groups in our analysis (
4No weights were applied at the group level because all federal states were included in the data collection and no sampling of federal states was necessary (e.g.,