Moving into multiculturalism. Multicultural attitudes of socially mobile individuals without a migration background

ABSTRACT

In various Western European cities, international migration has transformed the former ethnic majority into a numerical ethnic minority. We study people without a migration background to shed light on the former majority’s attitudes towards multiculturalism in these majority-minority contexts. Among those without a migration background, we specifically focus on socially mobile individuals in order to disentangle the influence of primary and secondary socialization on attitudes towards multiculturalism. Using survey data on Amsterdam, Antwerp, Malmo, Rotterdam and Vienna (n = 2,457), we found that, whilst controlling for the effects of primary and secondary socialization, both upward and downward mobility associates to more optimistic multicultural attitudes. We argue that the experience of social mobility equips people with a reflexivity which allows them to have a more optimistic perspective on the multi-ethnic city. In this way, this article improves our understanding of why some people are more willing than others to adapt to multi-ethnic contexts.

KEYWORDS:

• European cities
• majority-minority
• migration
• reflexive habitus
• social mobility
• socialization

International migration has substantially changed the socio-demographic make-up of several major Western European cities and increased the local ethnic diversity. Various reactions to increasing ethnic diversity can be observed among people without a migration background.¹ Some may insist on maintaining their ‘own’ culture and rejecting foreign influences (Alba and Duyvendak 2019; Mepschen 2019). Others might adapt to these new social circumstances, embrace the idea of the multi-ethnic city and consider it as ‘normal’ and part of everyday life (Wessendorf 2014). The acceptance and support of the culturally diverse society, as opposed to the wish for a culturally homogenous society, is referred to as multiculturalism (Berry and Kalin 1995: 306). Existing studies have investigated the reasons majorities use to justify or criticize multiculturalism (Verkuyten 2004) or examined the endorsement of multiculturalism by the majority in relation to ingroup identification and outgroup friendships (Verkuyten and Martinovic 2006), level of education (Breugelmans and van de Vijver 2004) and ethnic ingroup identification, perceived group essentialism and protestant ethic ideology (Verkuyten and Brug 2004).

Extant studies, however, neglect two vital aspects. Firstly, they do not shed light on how attitudes toward multiculturalism play out in settings where the former ethnic majority has become a numerical minority. In Amsterdam, for instance, the former ethnic majority group, i.e. ethnic Dutch without a migration background, is now an ethnic minority, numerically speaking.² In the field of ethnic and migration studies, various scholars refer to this as a majority-minority context – a context in which the national ethnic majority is a numerical minority at the local level (Craig and Richeson 2014a, 2014b; Crul 2016). This development can be observed in several large cities in Europe and has been described as ‘one of the most significant urban transformations of our time’ (Crul and Lelie 2019: 191).

This demographic development is an important aspect of inter-group relations as previous research finds that the loss of numerical group status consequently leads to more negative attitudes towards increased ethnic diversity. For instance, researchers primed Whites in experimental designs with being a future minority in the United States. After the exposure to this demographic projection, Whites endorse more conservative political policies (Craig and Richeson 2014a) and feel more angry towards and fearful of ethnic minorities (Outten et al. 2012). Based on the outcomes of these experiments, the authors suggest that ‘increasing diversity of the nation may engender a widening partisan divide’ (Craig and Richeson 2014b: 1189) and ‘rather than ushering in a more tolerant future, the increasing diversity of the nation may actually yield more intergroup hostility’ (Craig and Richeson 2014a: 758).

Secondly, despite empirical evidence that a person’s life trajectory affects inter-ethnic attitudes (Priest et al. 2014), the role of the life course in the formation of multicultural attitudes has received less attention in extant studies. To address this gap in the field of ethnic and migration studies, we draw from research on social mobility. We do so for two reasons. Firstly, as socially mobile individuals have been socialized in different social strata, a focus on social mobility allows us to examine the importance of childhood socialization experiences relative to socialization experiences later on in life. As previous research finds that attitudes towards multiculturalism are particularly stratified along educational lines (Breugelmans and van de Vijver 2004; van de Vijver et al. 2008), we focus on educational social mobility. Educationally socially mobile people are people who hold a higher or lower educational degree than their parents. Secondly, the experience of social mobility has proven to be an influential factor that affects attitudes and behaviour in various domains of social life (e.g. Daenekindt and Roose 2013, 2014; Van Eijck 1999; Friedman 2014, 2016; Gugushvili et al. 2020; Tolsma et al. 2009; Zhao et al. 2017).

In this article, we focus on people without a migration background within majority-minority context in Western European cities. This is important because inter-ethnic relations in these contexts cannot be fully understood without taking into account this group’s perspective on the demographic hierarchy upheavals. Research on people without a migration background in majority-minority contexts is scarce (Alba and Duyvendak 2019). We use unique survey data collected in five European majority-minority cities: Amsterdam, Antwerp, Malmo, Rotterdam and Vienna. In this way, we avoid priming respondents with the fact that they have become a numerical minority and take into consideration effects such as the normalization of diversity (Wessendorf 2014). We study whether and how social mobility shapes the multicultural attitudes of people without a migration background. Our study adds to existing explanations of multicultural attitudes as we identify social mobility as an important determinant of multicultural attitudes. In this way, this article improves our understanding of why some people are more willing than others to embrace multi-ethnic contexts.

Multiculturalism and the acquisition of inter-ethnic attitudes

The concept of multiculturalism has found applications in philosophies, institutional frameworks and political interventions. Due to the application in a variety of fields, numerous definitions of the concept exist. For this research, we follow the definition that understands multiculturalism as an individual’s attitude towards the acceptance and support of a culturally diverse society (Berry and Kalin 1995: 306). To improve our understanding of how people’s life course affects multicultural attitudes, we draw from previous research which investigates how people acquire other attitudes. Intergenerational transmission of attitudes from parent to child has received ample attention in this regard. One stream of literature stresses that attitudes towards ethnic diversity are formed to a large extent during socialization experiences. In general, we can make a distinction in people’s life course between primary and secondary socialization experiences (Berger and Luckmann 1966; Perey-Felkner 2013). As defined by Berger and Luckmann (1966), primary socialization refers to ‘the first socialization an individual undergoes in childhood, through which he becomes a member of society’ (p. 150). Secondary socialization is ‘any subsequent process that inducts an already socialized individual into new sectors of the objective world of his society’ (ibid.). For some individuals, these two socialization experiences align, but for others they may diverge in different directions. Whereas some individuals remain in the social position they were born and socialized into, others experience intergenerational social mobility. Social mobility is the process of moving between social strata within a society. This experience can play out in two directions: upward or downward. Upward mobility refers to an intergenerational move from a lower level on the societal ladder to a higher level; downward mobility refers to an intergenerational move from a higher level on the societal ladder to a lower level. Hence, a focus on social mobility allows us to scrutinize the role of socialization for multicultural attitudes.

Various studies stress the role of primary socialization. The idea that children acquire racial and ethnic attitudes from their parents during primary socialization has a long tradition in migration research, and can be traced back to Allport (1954). In their meta-analysis, Degner and Dalege (2013) reviewed empirical evidence from 131 publications on the parental socialization effects on intergroup attitudes of children and found a ‘significant medium-sized average effect size for the correlation between parental and child intergroup attitudes’ (p. 1270). These results suggest that parents largely transmit their attitudes towards ethnic diversity to their children.

Yet, parents are not the only socialization agent for attitudes towards ethnic diversity (cf. Corsaro 2015). Other socialization agents, such as other family members, peers, teachers or community mentors, can be influential in the ethnic-racial socialization process (for an overview see Priest et al. 2014). The educational system is regarded as the most significant secondary socialization agent when it comes to the transmission of liberal values (Hello, Scheepers, Vermulst, and Gerris 2004: 253). For instance, Hello, Scheepers, Vermulst and Gerris (2004) compare the effects of different types of parental influences with the effects of educational attainment on the attitude towards social contact with ethnic minorities. They find that, while parents do have socializing effects on young adults, education is the more influential agent for the socialization of attitudes towards other ethnic groups.

Previous research demonstrates that attitudes towards migration and ethnic diversity, as well as moral and political attitudes, are stratified along educational lines. In general, citizens with higher levels of education have higher degrees of openness and tolerance (Stouffer 1956). As Jacobsen (2001) summarizes: ‘[t]he higher the education, the more tolerant, the less authoritarian, the more leftist, the more modern and idealist, etc. a person becomes’ (p. 354, and cf. Jenssen and Engesbak 1994; Hellevik 1996 in Jacobsen 2001). More highly-educated people are also thought to hold comparably fewer stereotypical beliefs about ethnic minorities (Schuman et al. 1997), show less support for ethnic discrimination (Coenders and Scheepers 1998) and display a less ethnocentric worldview (Scheepers et al. 1989). Higher education is also positively related to ethnic tolerance and a preference for cultural diversity (Citrin et al. 1997; Fetzer 2000; Hainmueller and Hiscox 2010; van der Waal et al. 2010). In particular, the higher educated are more optimistic towards multiculturalism (Breugelmans and van de Vijver 2004).

Because of the paramount role of education for attitudes towards migration and ethnic diversity, we focus on educational mobility as a form of social mobility. Due to the educationally stratified nature of attitudes towards migration and ethnic diversity, educationally mobile individuals have been exposed to both the multicultural attitudes characteristic of their social origin (primary socialization) and the multicultural attitudes characteristic of their social destination (secondary socialization). Therefore, studying socially mobile people allows us to scrutinize the social roots of multicultural attitudes. In the case of educational mobility, social origin refers to the educational level of the respondent’s parents and social destination refers to the educational level of the respondent.

The experience of social mobility as an engine for multicultural attitudes

In addition to the above discussed effects of social origin and destination, the experience of mobility in itself can have consequences. For example, Sorokin (1927; see also Daenekindt 2017) argues that socially mobile individuals are more likely to experience psychological distress. This distress, according to Sorokin’s ‘dissociative thesis’, results from the incongruence between the social position of destination and the primary socialization context. Sorokin argues that the socially mobile do not feel at ease in their current social milieu. This feeling of discomfort can generate feelings of deprivation and result in negative attitudes towards current migration developments.

We argue that such mobility effects can matter for support of multiculturalism. We do so, because both social mobility and the encounter of increased ethnic diversity in everyday life entail adaptation processes that may be similar. The socially mobile already have experience with encountering new social contexts due to their mobility experience. Their previous experience with the adaption to norms and values typical of the social position of destination, has provided them with the experience of other forms of diversity. This experience with other forms of diversity allows the socially mobile to more easily embrace multiculturalism with the increased diversity it entails. This argument is further supported by the theory of the reflexive habitus (Sweetman 2003). Sweetman argues that for some individuals, reflexivity has become habitual and ‘for those who display a flexible or reflexive habitus, processes of refashioning – whether emancipatory or otherwise – may be second nature rather than difficult to achieve’ (p. 537). Inspired by Giddens (1991), Sweetman holds changes in contemporary society, such as globalization, accountable for the growing reflexivity in society. However, Sweetman stresses that the reflexive habitus is not equally distributed in society. Some groups, such as younger people, are more prone to inhabit reflexivity compared to older people. Similarly, the socially mobiles’ movement between social positions makes it easier for them to adapt to new social circumstances, such as the multi-ethnic city, compared to their immobile counterparts. Extant research demonstrated that social mobility generates increased reflexivity (Abrahams and Ingram 2013; Ingram and Abrahams 2015; Ingram 2018; Reay et al. 2009) and Ingram (2018) shows ‘that reflexivity generated by awareness of different influences of different fields can help to generate flexibility’ (p. 212). Based on this, we argue that social mobility – both upward and downward mobility – generates an attitudinal and psychological make-up that enables a greater openness towards multiculturalism. Mobile individuals have, so to speak, embodied the ability to adapt to difference, and this will be reflected in their attitudes towards the multicultural city. Hence, our first hypothesis is as follows:

H1: Both upwardly and downwardly mobile individuals hold more optimistic attitudes towards multiculturalism compared to their immobile counterparts.

The effects of upward and downward social mobility on multicultural attitudes

Hypothesis 1 considers the consequences of upward and downward mobility for multicultural attitudes to be equal. It is, however, also possible that the two different mobility trajectories, i.e. upwards and downwards, have different implications for multicultural attitudes. For instance, upward mobility is a more positive experience than downward mobility as it entails increased social status. Furthermore, as Alba (2009) claims, the absence of competition – for instance, on the labour market – enables the blurring of ethnic boundaries:

[E]thno-racial boundaries are most likely to fade in relevance for those on the disadvantaged side when the perceived threat to the privileges of those on the advantaged side is reduced, and the latter are consequently less tempted to resort to the devices of ethno-racial exclusion. (p. 90)

This implies that people are more likely to hold positive multicultural attitudes once the perceived threat from the Other is reduced. Upward mobility could have such an effect on an individual. If a person has experienced upward mobility and perceives that they are better off than their parents, they might also have more optimistic multicultural attitudes as they do not perceive a threat from people with a migration background.

Alba's (2009) argument also applies to downward mobility. Downward educational mobility might lead to greater perceived threat causing the downwardly mobile to be less favourable of multiculturalism. At the same time, based on relative deprivation theory, the experience of downward mobility can result in frustration and insecurity. These negative feelings are ‘caused by disappointing comparisons with one’s own past (that is, when the trajectory of one’s life suddenly deviates from the expected) […]’ (Rydgren 2007: 247). Due to the drop in social status, the downward experience allegedly leads to possible frustration, which in turn leads to greater ethnic antagonism. Therefore, people who have experienced downward social mobility may be less likely to embrace multiculturalism:

H2: Upward social mobility results in more optimistic attitudes towards multiculturalism, while downward social mobility results in more pessimistic attitudes towards multiculturalism.

Data and methodology

To investigate the consequences of social mobility on the multicultural attitudes of Europeans without a migration background in multi-ethnic cities, we used data from the Becoming a Minority (BaM) project. BaM is a comparative project which investigates attitudes and behaviour of people without a migration background toward ethnic diversity in majority-minority contexts in Europe. The data were collected in 165 neighbourhoods in the majority-minority³ cities Amsterdam, Antwerp, Rotterdam, Malmo and Vienna (n = 2,457) in 2019. The survey respondents were randomly selected from the population of people aged between 25 and 45 years-old. All respondents and both of their parents were born in the country of survey and are hence considered as people without migration background. As respondents are nested in neighbourhoods as well as in cities, we had to account for the structure of the data. Therefore, we included city-dummies in the model and requested robust standard errors to account for the clustering of errors at the neighbourhood level.

Measures

We measure attitudes towards multiculturalism using the items presented in Table 1. The answer categories for each item range from (1) strongly disagree to (5) strongly agree. To capture the attitudes towards multiculturalism, we took the mean of the seven items (Cronbach’s Alpha = .889, M = 3.05, SD = .85). High scores on this measure associate to the willingness to adapt to a multicultural society while low scores associate to the wish to maintain one’s own culture.⁴

Table 1. Items for multiculturalism scale.

[Country] schools should adapt more to the diverse cultural backgrounds of their pupils (e.g. taking into account pupils’ culturally related food restrictions).

[Country] should learn more about the cultural way of life of people with a migration background (e.g. learning about religious or cultural festivities).

[Country] politicians should think more about the interests of people with a migration background.

[Country] people would benefit from having more social contact with their neighbours with a migration background.

[Country] politicians should think more about the interests of people with a [country] background.*

[Country] schools should uphold the cultural traditions of [country].*

[Country] people should maintain the [country] cultural way of life.*

Note: Items with asterisk correspond to reverse-keyed items.

Educational level and educational mobility. Based on the International Standard Classification of Education 2011 (ISCED), we classified educational level into three categories: (1) low (categories 1–3), (2) medium (categories 4-5), and (3) high (categories 6–8). Social mobility is measured as the difference in highest educational degree between father and respondent (e.g. Tolsma et al. 2009).⁵ Table 2 presents the mobility table.

Table 2. Overview of intergenerational educational mobility among survey respondents, relative frequencies (n = 2009).

		Destination: Education respondent
		Lower	Medium	Higher	Total
Origin: Education father	Lower	12.6	5.0	21.8	39.4
	Medium	3.3	1.4	11.2	15.9
	Higher	4.4	2.5	37.8	44.7
	Total	20.3	8.9	70.8	100

Control variables. To account for spuriousness, we controlled for variables that are related to both multicultural attitudes and educational level. We controlled for gender (54.1 per cent female) and age (M = 34.4, SD = 6.0) as these variables were found to have an effect on multicultural attitudes and educational attainment (Arends-Tóth and Vijver 2003; Tolsma et al. 2009; van de Vijver et al. 2008). We further included net household income ranging from ‘low’ to ‘high’ on a 5-point scale (M = 3.15, SD = 1.2) in our analysis. We do so as both income and educational attainment have effects on inter-ethnic attitudes (e.g. Semyonov et al. 2004) but at the same time, educational attainment can have repercussions on the level of income. Hence, to estimate the effects of education independently of the effects of income, we included net household income as a control variable.

Analysis

To test our hypotheses we applied Diagonal Reference Models (DRMs) (Sobel 1981) in Stata (Kaiser 2018). DRMs are particularly suitable for estimating effects of movements across levels of categorical variables, such as education, and are therefore widely applied in social mobility research (e.g. Daenekindt and Roose 2013, 2014; Houle and Martin 2011; Tolsma et al. 2009). In analysing socially mobile individuals, the model considers three effects: (1) the origin effect, which is the influence of primary socialization, (2) the destination effect, which is the influence of secondary socialization, and (3) the mobility effect, which is the influence of the discrepancy between primary and secondary socialization. DRMs were developed to disentangle these three effects. Therefore, for the purpose of this study DRMs are superior to conventional methods which do not allow to estimate effects of social mobility net of the effect of social origin and destination (van der Waal et al. 2017).

Theoretically, DRMs are based on the idea that immobile individuals represent the core of a specific social position. This because immobile individuals have been socialized by only one social position and have therefore embodied the ‘core’ characteristics of that social position (De Graaf et al. 1995; Houle 2011; Sorokin 1959). DRMs therefore estimate the effects of (1) social origin and (2) destination for mobile individuals based on immobile individuals. Consider, for example, the top right cell in the mobility table (Table 2). Individuals in this cell grew up in the lowest social position and ended up in the highest social position. To estimate the influence of the social position of origin (1) for this cell, the model will look at the immobile individuals in the cell ‘low-low’, while it will look at the immobile individuals in the cell ‘high-high’ to estimate the influence of the social position of destination (2). The (3) mobility effect is then estimated as the effect which can neither be attributed to the origin nor the destination and can therefore be interpreted as the effect of social mobility net of the effect of origin and destination.

Using DRMs, we modelled the multicultural attitudes of socially immobile individuals as a function of the multicultural attitudes characteristic of the social origin and the multicultural attitudes characteristic of social destination. We specified the baseline model as follows: $Y_{ijk}=p{\mu }_{ii}+\left(1-p\right){\mu }_{jj}+\sum {\text{β}}_b{x}_{ijb}+{ϵ}_{ijk}\left(M0/\mathrm{b}\mathrm{a}\mathrm{s}\mathrm{e}\mathrm{l}\mathrm{i}\mathrm{n}\mathrm{e}\mathrm{m}\mathrm{o}\mathrm{d}\mathrm{e}\mathrm{l}\right)$Y is the value of the dependent variable, multicultural attitudes, and the subscripts i and j represent the social positions of origin and destination of actor k. µ_ii (origin) and µ_jj (destination) are the estimated means for multicultural attitudes of immobile respondents. The model then predicts the scores for socially mobile based on µ_ii and µ_jj. The p-parameter represents the effect of the position of social origin relative to the social destination and is constrained to the interval [0; 1]. If p is significantly higher than .5, social origin has a stronger effect on the dependent variable than the social destination (and vice versa). For example, an origin weight p of .8, indicates that multicultural attitudes of socially mobile individuals are more similar to the immobile in their origin than to the immobile in their destination. The covariates x_b can be interpreted as in ordinary least square regression analysis with the subscript b indexing the different covariates. These covariates also include the effects of social mobility.

Results

To test our hypotheses, we specified different models. M0 is the baseline model and all the other models are extensions of it. M1 includes the effect of social mobility in general (no distinction between upward and downward mobility) (Hypothesis 1). M2 includes the effects of upward and downward mobility separately (Hypothesis 2). Table 3 shows the parameter estimates of the three models.

Table 3. Parameter estimates from the diagonal reference models predicting multicultural attitudes (n = 1756, 163 neighbourhood clusters).

	M0		M1		M2
	b	95% CI	b	95% CI	b	95% CI
Diagonal intercepts
µ11: lower education	−.24 (.05)***	[−.34; −.13]	−.22 (.05)***	[−.33; −.12]	−.23 (.06)***	[−.34; −.12]
µ22: medium education	−.09 (.06)	[−.21; .03]	−.18 (.07)*	[−.31; −.04]	−.17 (.08)*	[−.33; −.01]
µ33: higher education	.33 (.04)***	[.25; .41]	.40 (.05)***	[.30; .50]	.39 (.06)***	[.28; .50]
Weight parameters
p: weight of origin	.23 (.06)^a	[.10; .35]	.36 (.08)	[.21; .52]	.32 (.21)	[−.09; .73]
q: weight of destination (1−p)	.77 (.06)^b	[.65; .90]	.64 (.08)	[.48; .79]	.68 (.21)	[.27; 1.09]
Mobility effects
General mobility			.16 (.06)* ^c	[.03; .28]
Upward mobility					.13 (.14)	[−.14; .39]
Downward mobility					−.18 (.10)	[−.01; .37]
Control variables
Gender (male)	−.10 (.04)**	[−.17; −.03]	−.11 (.04)**	[−.18; −.04]	−.11 (.04)**	[−.18; −.04]
Age	−.01 (.00)***	[−.02; −.01]	−.01 (.00)***	[−.02; −.01]	−.01 (.00)***	[−.02; −.01]
Income	−.03 (.02)	[−.07; .01]	−.02 (.02)	[−.06; .01]	−.03 (.02)	[−.06; .01]
City dummies
Amsterdam	.38 (.09)***	[.20; .56]	.35 (.09)***	[.18; .53]	.35 (.09)***	[.17; .53]
Antwerp	.45 (.09)***	[.29; .62]	.43 (.08)***	[.27; .60]	.43 (.08)***	[.27; .60]
Malmo	.34 (.10)**	[.15; .54]	.32 (.10)**	[.12; .51]	.32 (.10)**	[.13; .51]
Rotterdam	.34 (.09)***	[.16; .52]	.31 (.09)**	[.13; .49]	.31 (.09)**	[.13; .48]
Vienna	ref. cat.		ref. cat.		ref. cat.
Log pseudolikelihood	−2006.93		−2001.85		−2001.80

* p < .05; ** p < .01; *** p < .001.

^a < .5 (p < .05); ^b > .5 (p < .05); ^c p = .014.

Robust standard errors in parentheses.

For model selection, we examined the parameter estimates added to the baseline model. M1 extends M0 with the effect for general mobility. M1 is an improvement compared to M0 because this added predictor is significant. M2 includes variables for upward and downward mobility. M2 does not meaningfully improve model fit as the added effects are not significant. Therefore, we selected M1.⁶

As we can see in M1, immobile respondents who have attained higher education express the greatest openness toward multiculturalism and significantly so (µ₃₃ = .40). Immobile people with lower levels of education hold more negative multicultural attitudes (µ₁₁ = −.22). This is in line with previous research that finds that higher education is associated with greater openness towards issues on migration and ethnic diversity and shows that multicultural attitudes in majority-minority settings are stratified as we expected. Furthermore, we can see that, in general, men are less open to multiculturalism than women (b = −.11). Moreover, despite the sample’s limited age range, we found that older people are significantly less open towards multiculturalism than younger people (b = −.01). Income has no effect on multicultural attitudes.

The p-parameter represents the relative influence of primary socialization (p) on multicultural attitudes relative to the influence of secondary socialization (q). A value of .5 indicates that the socially mobile resemble the immobile in both their origin and destination equally. The superscript a in Table 3 indicates whether the influence of primary socialization on the multicultural attitude is significantly higher or lower than .5. Superscript b indicates the same but for the influence of secondary socialization. Whereas the literature review has shown that scholars put great emphasis on the importance of primary socialization for the formation of inter-ethnic attitudes, we find no evidence that either socialization context has a predominant effect on multicultural attitudes. The p-parameter does not differ significantly from .5, which suggest that multicultural attitudes are shaped by both socialization experiences.

We then turn to the estimate of the mobility parameter general mobility to test Hypothesis 1. Hypothesis 1 states that the socially mobile are more open to multiculturalism regardless of whether their social mobility was upward or downward. M1 shows that social mobility has a positive effect on the multicultural attitude (b = .16), while controlling for the effects of social origin and social destination. This provides evidence for Hypothesis 1 and indicates that, among Europeans without a migration background, socially mobile people are more open to multiculturalism than socially immobile people.

Discussion and conclusion

In this article, we studied the attitudes towards multiculturalism of people without a migration background in majority-minority cities. Our findings provide evidence of an association between multicultural attitudes of Europeans without a migration background and the experience of social mobility (while controlling for the effects of origin and destination). As our results demonstrate, both upward and downward mobility are related to a greater openness to multiculturalism in majority-minority contexts. This suggests that the socially mobile are more flexible to adapting to on-going migration developments such as increasing ethnic diversity.

By focussing on socially mobile individuals, we were able to get an analytical understanding of the influence of primary and secondary socialization experiences on multicultural attitudes. Whereas other researchers have found that parents pass on their inter-ethnic attitudes to their children to a large extent (Degner and Dalege 2013), our results show that the primary socialization context does not have a predominant effect on respondents’ multicultural attitudes when socialization experience is conceptualized as educational level of father and educational level of respondent. Rather, multicultural attitudes are shaped by both socialization experiences during childhood and socialization experiences later in life.

Whereas sociological theory often associates the experience of social mobility with undesirable psychological consequences, our analysis highlights positive consequences. As Giddens (1991) argues, late-modern developments, such as economic and cultural globalization, demand a certain level of reflexivity from people in contemporary societies. Alterations to the traditional structures of the labour market, the changing nature of social relations and growing diversity in all aspects of society require flexibility and an openness to adaption. Adapting to the multi-ethnic city is no exception to this development. Our findings align with reflexive habitus theory (Sweetman 2003), which posits that this habitus is not distributed equally in contemporary society. We added to this theory by identifying individuals in society who are more likely than others to inhabit such a reflexive habitus and we demonstrate the implications that such a habitus entails. The socially mobile seem to have an advantage over their immobile counterparts in that they can react to this demand for flexibility more easily as they have already embodied it through their mobility trajectory (cf. Sorokin 1927). In line with previous empirical research on social mobility and reflexivity (Abrahams and Ingram 2013; Ingram and Abrahams 2015; Ingram 2018; Reay et al. 2009), our findings suggest that the mobility experience equips people with a reflexivity which allows for the openness to embrace cultural diversity. In addition to the experience of passing through different social contexts during their mobility trajectory, socially mobile individuals are likely to continue navigating between different social worlds or social circles – and the associated norms and values – after their mobility experience. This allows the socially mobile to embrace majority-minority realities more easily than their non-mobile counterparts who have not had to negotiate different contexts as the consequence of their lack of educational mobility, either in their past or present.

Remarkably, we found that also downwardly mobile individuals demonstrate a greater openness to multiculturalism compared to their immobile counterparts. While the experience of downward mobility has generally been associated with feelings of frustration and the expression of more negative inter-ethnic attitudes, our study finds no evidence for this. On the contrary, we found that the downwardly mobile are more open to multiculturalism than their immobile peers. This suggests that relative deprivation theory might not be helpful for understanding the consequences of downward mobility on inter-ethnic attitudes.

We argue that the openness of the downwardly mobile is the result of the reflexivity which they develop throughout the social mobility process. At the same time it is possible that the downwardly mobile find themselves in social destinations which might facilitate them with greater structural meeting opportunities (Blau 1960; Martinović 2013) with people with a migration background and that personal inter-ethnic contact might contribute to their greater tolerance (Allport 1954; Paluck et al. 2019; Pettigrew and Tropp 2006). Nevertheless, our research adds to an already existing body of empirical work showing that the experience of downward social mobility is multi-faceted and can play out in different ways in different domains of social life, more precisely when it comes to positive or negative attitudes towards immigrants. While downward mobility might lead to more negative attitudes towards highly specific aspects of inter-ethnic attitudes (for instance, inter-ethnic marriage, as Tolsma et al. show [2009]) others have not found any negative effects of downward mobility on migration issues overall (Paskov et al. 2020; Stawarz and Müller 2020; Tolsma et al. 2009).

A novelty of this study is that we focussed on individuals without a migration background who became an ethnic minority as a result of international migration. We are aware that the majority-minority perspective is highly dependent on who is classified as a majority and who a minority (Alba 2018). We focussed on majority-minority cities, with ‘people without a migration background’ defined as people who were born in the country of residence with both parents born in the same country (cf. Arends-Tóth and Vijver 2003; Crul and Lelie 2021; Martinović 2013). We believe that our conclusions provide insights into the way multicultural attitudes are developed in general and, hence, can be extended to attitudes towards multiculturalism in other cities. Yet, there are limits to which cities our findings can be extended. Previous research shows that context influences both multicultural attitudes (Leong and Ward 2006) and the consequences of social mobility on, for instance, attitudes towards immigration (Paskov et al. 2020). It is for future research to investigate how macro effects shape the association between the experience of social mobility and multicultural attitudes.

For this research, we relied on the concept of multiculturalism. The concept of multiculturalism is not without critique. For example, Vertovec and Wessendorf (2010) argue that multiculturalism fosters separateness in form of ‘parallel societies’ (p. 8) through its denial of common values. While we acknowledge this critique, we also believe that the concept of multiculturalism is very valuable to study cities characterized by growing ethnic diversity as the integration of people with a migration background partially depends on multicultural attitudes of people without a migration background (Breugelmans et al. 2009). A second reflection on the concept stresses that multiculturalism has an attitudinal, as well as behavioural dimension. Both dimensions do not coincide as, for example, previous research found that optimistic attitudes towards ethnic diversity do not necessarily translate into more ethnically diverse social networks (Blokland and van Eijk 2010). In this article, we focussed on the way social mobility shapes the attitudinal dimension of multiculturalism. In a similar way, future research should address the way experiences of social mobility relate to the behavioural dimension of multiculturalism, which can provide additional insights into the ramifications of international migration on Western societies.

We hope that, in addition to relying on attitudes to understand ethnically diverse contexts, the concept of ‘reflexive habitus’ inspires future research to study other aspects of social life in ethnically diverse cities. For instance, future research could investigate whether the reflexive habitus can be transferred to other aspects of social life. Instead of departing from a perspective where attitudes predict inter-ethnic contact, future research could explore whether embodied reflexivity makes it easier to maintain comparably more inter-ethnic contact. Similarly, research on ethnic diversity of social networks could benefit from including a social mobility perspective. In this article, we contributed to research on inter-ethnic attitudes by introducing such a social mobility perspective. More generally, we hope that our study inspires future research to further improve our understanding of why some people are more willing than others to adapt to multi-ethnic contexts.

Acknowledgements

We thank Elif Keskiner, Ismintha Waldring, Lore van Praag and Maurice Crul for their valuable comments and suggestions on earlier versions of this manuscript. We also very much appreciated Lise Woensdregt’s title idea.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the European Research Council [grant agreement ID: 741532].

Notes on contributors

Lisa-Marie Kraus

Lisa-Marie Kraus is a PhD candidate at the Vrije Universiteit Amsterdam. As part of the Becoming a Minority (BaM) project, which is funded by the European Research Council with Maurice Crul as the Principal Investigator, her research interests concern the experiences of people without a migration background in European majority-minority cities. For more information on the BaM project see www.bamproject.eu.

Stijn Daenekindt

Stijn Daenekindt holds an MSc in Statistics (2012, KU Leuven) and a PhD in Sociology (2015, Ghent University). In his research, he mainly focusses on (aspects of) social inequality.

Notes

1 Here understood as people who were born in the country of residence with both parents born in the same country.

2 When migration background is defined as born in the Netherlands with both parents born in the Netherlands (cf. CBS 2018)

3 Based on Statistics Netherlands (CBS 2018) for Amsterdam and Rotterdam; Stad Antwerpen (2019) for Antwerp; Statistics Sweden (2018) for Malmo; Stadt Wien MA 17 (2017) for Vienna and when ‘migration background’ is defined as born in the country of residence with both parents born in the same country. When the sources use a different definition of ‘migration background’, the figures were calculated manually so that they fit the definition of ‘migration background’ for this research as closely as possible.

4 We performed an exploratory principal component analysis to test whether we could find more distinct, underlying concepts. The analysis did not yield a clean multi-factor structure and therefore, following van de Vijver et al. (2008), we proceeded with a unifactorial multiculturalism scale.

5 Because some studies suggest that fathers have more impact on the attitudes of sons and mothers on the attitudes of daughters (e.g. Nieuwbeerta and Wittebrood 1995), we conducted sensitivity analyses in which we measured social mobility as the difference in highest educational degree between father and son and mother and daughter. These sensitivity analyses yielded the same substantive conclusions as the analyses reported here.

6 We also fitted the following model: $Y_{ijk}=\left(p+m{x}_{ijm}\right){\mu }_{ii}+\left(1-\left(p+m{x}_{ijm}\right)\right){\mu }_{jj}+\sum {\text{β}}_b{x}_{ijb}+{ϵ}_{ijk}$ to test what is often referred to in the literature as the maximization/maximalization hypothesis. The maximization hypothesis tests whether the socialization context associated with the highest social status guides attitudes: in other words, the upwardly mobile adapt to the attitudes characteristic of their social destination, while the downwardly mobile retain the attitudes characteristic of their social origin. The model, however, does not improve the model fit.