The issue of immigrant integration has been prominent in sociology at least since the beginning of the 20th century—at that time, research typically focused on the integration of immigrants from European countries into the USA (Zhou & Bankston 2016). In classical assimilation theory, the traditional model of integration—in the sense of an empirical generalisation, or a ‘grand narrative’ as Alba & Foner (2015) call it—involved a gradual assimilation to the majority population. This developed into the ‘melting pot’ metaphor where integration was seen as a two-way process: the traditional assimilation in combination with a constantly changing ‘mainstream culture’, which gradually adapts to new immigrant groups (e.g. Alba & Nee 2003).

Subsequently, models of integration that have taken more heterogeneous outcomes into account have been proposed, such as the idea of ‘segmented assimilation’ (Portes & Zhou 1993), or the model of acculturation proposed by Berry (1997). These narratives make a claim about divergent destinies of immigrant groups and draw attention to the fact that integration in the form of assimilation to ‘the mainstream’ is not a deterministic outcome: some groups may become (or assimilate to) constantly disadvantaged and marginalised groups in society. In particular, the proponents of this idea have emphasised the role of different ‘modes of incorporation’ (Portes & Rumbaut 1996: 82ff.; 2001: 46ff.), a concept that stresses the influence of the context of reception, including the institutional arrangements and attitudes of the receiving society or the more local context. The importance of contextual factors is likewise underlined in the neo-assimilationist perspective (Alba & Nee 2003; Alba & Waters 2011).

In its simplest theoretical form, integration could be seen as a function of: (1) selective migration and initial group differences in terms of abilities, skills, values and cultural habits and so on, including the attempts and will (or, alternatively, the resistance) of immigrants to integrate; (2) the opportunities of integration in receiving societies, including things that may obstruct integration (e.g. legal and institutional barriers, residential and other forms of segregation, the majority’s prejudice and lack of will to form inter-ethnic ties, discrimination), and things that may facilitate integration (e.g. policies enabling citizenship, language courses, anti-discrimination laws); and (3) the exposure of immigrants to the destination country and its ‘culture’ (primarily, time spent living there). This leads to three general influences that are particularly interesting when studying integration: origin group influences, destination-country influences and exposure to the destination country, as well as their interactions (e.g. exposure influences may vary across origin groups). It matters, in other words, where you come from, where you end up and how much time you have spent there.

(p.9) These three nodes summarise the starting points to understand different outcomes of the integration process; they are also key to our overarching strategy of analysis in the thematic chapters of this book. Here, we want to outline why, and in which direction, influences related to origin, destination and exposure may affect integration processes. Of course, we cannot, at this point, come up with strict hypotheses. This is not least due to the fact that integration is, as outlined, a multidimensional concept. Depending on the specific aspect of integration, different subject-specific theories have to be invoked in order to gain an idea of the exact mechanisms behind integration, and to derive more elaborated and precise expectations of these processes. This will be done in the thematic chapters of this volume. Our discussion here is on a more general and abstract level. Methodologically speaking, and following the terminology of Merton (1957), it delivers ‘general orientations’, that is, overarching heuristic sketches of how and in which direction more fine-grained theoretical arguments could and should be developed when studying specific integration outcomes.

1.3.1 Origin group influences

To understand the average differences in integration aspects across origin groups in a given destination country, we usually need to consider two factors. First, the distance between this receiving country and the sending countries (or regions) when it comes to affluence, visible minority markers (e.g. skin colour), democratic tradition, religion, values and beliefs, literacy, educational provision and so on; a distance that is likely to be negatively related to the probability of integration. Secondly, there is a selection of migrants (in relation to the origin population) on characteristics that have an influence on integration, such as their educational qualifications (e.g. Ichou 2014). The combination of origin country characteristics and the selectivity of migrants produce, for each receiving country, a compositional fingerprint of each immigrant group, and on the aggregate, of the entire immigrant population.

What are the theoretically most relevant characteristics for this compositional fingerprint? A popular and intuitive account of the probability of integration is the degree and kind of human capital possessed by individuals in immigrant groups (e.g. Borjas 1989). Marketable skills and proficiency in the destination-country language facilitate the chances of getting a job; human capital is therefore often believed to be the most important vehicle to avoid marginalisation and exclusion. It is possible to extend the conventional operationalisation of human capital (in terms of qualifications, skills and health) to other individual traits of value in the labour market, such as ‘drive’ or ‘ambition’, which are traditionally assumed to characterise immigrants (e.g. Gordon 1964). Other origin country characteristics could include cultural, perhaps principally religious, traits, values and visible minority markers, the principle on which prejudice and discrimination are often based.

Because children rarely make the decision to migrate, and often have spent a limited time in the origin country, the characteristics that adhere to origin groups (p.10) pertain mostly to parents. However, they get transmitted across generations genetically and through primary socialisation (e.g. child-rearing) and role modelling. A basic idea of the segmented assimilation concept is that the wider local community of co-ethnics will also play an important part in this process (Portes & Rumbaut 2001).

These notions suggest the seemingly trivial, but very important fact that an immigrant group’s integration depends on characteristics of the origin country and the composition of the respective minority group in the receiving country in terms of individual and community characteristics that promote integration. Here, parents’ human capital and structural integration are important for children, because they facilitate social and cultural integration and determine the socioeconomic standing of the family. The latter fact is crucial for children’s structural integration (e.g. educational opportunities, residential location), and segregation is in turn likely to negatively affect social integration. Therefore, some dimensions of integration are intrinsically related to socioeconomic inequalities, something that is occasionally pointed out (e.g. Portes & Rumbaut 2001), but often understated in the integration literature.

Origin influences, as discussed here, pertain to a given country of destination, but may differ across destination countries—the way that immigrants’ human capital can be transferred into jobs, for example, is clearly dependent on characteristics of receiving countries’ labour markets (van Tubergen et al. 2004). In Section 1.6.2, we analyse origin influences for each of our four destination countries, which will enable us to identify such interaction effects, but we must also acknowledge that for some origin characteristics (e.g. religion, literacy), our survey countries are fairly similar when compared to the bulk of origin countries (cf. Chapter 2).

Against the background of these considerations on origin influences, it should not come as a surprise that children of immigrants from the Western hemisphere— with smaller socioeconomic and cultural differences to our receiving countries— often integrate quicker than other groups. In contrast, the slowest pace could be expected for those from poorer areas with few educational opportunities and highly different cultural traits, especially Africa, the Middle East and southern and eastern parts of Asia. However, because of selective migration, the composition of the immigrant groups can also break such origin country patterns. We also expect that the differences that we do find between origin groups will depend on the composition of the groups in terms of socioeconomic characteristics, segregation and other proximate causes of individual-level integration such as language skills and religiosity—mechanisms that will be tested for in the thematic chapters of this volume and summarised in Section 1.6 of this chapter.

1.3.2 Destination-country influences

Most approaches to the study of integration, be it segmented assimilation theory or neo-assimilation theory, emphasise the importance of the context of reception for the integration prospects of different minority groups (p.11) (Portes & Rumbaut 1996; 2001; Alba & Nee 2003: 274ff.; Alba & Waters 2011: 3ff.; but see Waldinger & Catron 2016). Much of the literature has focused on receiving countries’ integration policies and the prejudice or discrimination immigrants face from the majority population. The combined ‘welcoming attitude’ by the state and the majority are assumed to have a positive effect on integration. One often assumed mechanism behind this is that citizenship rights make immigrants feel like they are accepted into their new country. However, provision of housing, availability of jobs, equality of treatment and eligibility of social benefits, among other things, are likely to be important as well (cf. Migration Policy Group 2011). Thus, inclusion has many faces.

While treating ethnic minorities as equals with the majority population is generally believed to promote integration, opinions are divided when it comes to policies treating them differently, for example by accommodating for their specific culture, as in some versions of ‘multicultural’ policy. The assumption that supporting or even encouraging cultural differences would facilitate integration is mostly based on the idea that recognition and tolerance of minority groups reduces ethnic tension and exclusion, and may lead to an overarching sense of belonging. Sceptics point to the segregating tendencies of promoting cultural differences and supporting minority communities, which would weaken the exposure to the majority culture, lead to less inter-ethnic social interaction and thereby slow down integration (e.g. Sniderman & Hagendoorn 2007).

State policies relevant to integration can, of course, be directed towards other goals. Welfare state policies and policies aiming at reducing inequality will normally favour those who are least well off, and it is possible that such policies therefore benefit immigrants (especially those newly arrived). An analogous effect for youth could be assumed to follow from child-friendly policies, for similar reasons. The mechanism here would be that social and cultural integration are facilitated when immigrant groups have high living standards and social security (see Section 1.3.1). However, generous welfare provision may also reduce incentives to work or invest in human capital, and more equal societies may attract especially low-skilled immigrants with less chance to get a firm footing in the labour market (e.g. Koopmans 2010). Because paid work is crucial for integration, it is difficult to ascertain the net effect of welfare state policies, unregulated labour markets and equality.

Nation states have different configurations of policies, some which may promote integration, while others may impede it—and what consequences different policies have is mostly unknown. Therefore, it is quite difficult to derive firm hypotheses about which of our countries would witness the strongest de facto integration—and along with the policy dimension comes other institutional characteristics as well as the attitudes of the majority population. Importantly, the different composition of immigrant groups in different receiving countries further contributes to complicate outright comparisons.

Nevertheless, statistical summaries of integration policies, majorities’ attitudes to immigrants and so on (see Chapter 2), tend to identify Sweden as a (p.12) particularly favourable place for integration within our four countries of study, while England appears to be the least favourable in several dimensions—though not civic integration. The same assumption could be derived from the polar location of these two countries when it comes to welfare state policies, child friendliness and overall inequality. However, the composition of the immigrant group in England has more advantages with regard to human capital, and with a less regulated labour market and a more widespread language, immigrants’ access to jobs may nonetheless be greater in England than in our other countries, which would speed up integration.

Although an overall ranking is difficult (partly because there are so many dimensions to consider), the Netherlands is closer to Sweden when it comes to child friendliness and welfare policies, and Germany has traditionally been less generous in handing out citizenships to immigrants. But while the Netherlands and Sweden may facilitate integration through welfare state provision, it has also been argued that the combination of such policies with multicultural policies lead to immigrant marginalisation and therefore slower integration (Koopmans 2010).

1.3.3 Exposure influences and immigrant generations

The process of ‘growing into’ a new society (its culture etc.) has, in previous literature, been studied as ‘adjustment’, ‘adaptation’, ‘acculturation’ or ‘assimilation’ and has been seen as almost inescapable, although research has also questioned the inevitability (and desirability) of this process (e.g. Portes & Rumbaut 2001). The mechanisms by which those of immigrant background disengage with the origin culture and become more similar to what has been called the ‘majority mainstream’ can be assumed to be both a gradual learning process (e.g. language proficiency) and a function of secondary socialisation through structural and social integration (cf. Esser 2006). Both these processes are arguably an outcome of exposure to the destination-country characteristics, and behind is, not least, a strong incentive structure. Integration, as Alba & Foner (2015: 5) put it, is about obtaining the ‘valued stuff’ of a society. How this valued stuff can be achieved, that is, which skills and resources are necessary and worth investing in, is determined by social production functions (Kalter & Granato 2002). Accordingly, assimilation is ‘something that frequently enough happens to people while they are making other plans’ (Alba & Nee 2003: 282).

In most studies of integration, exposure has been equated with immigrant ‘generation’ (see also Chapter 3), primarily distinguishing whether a person was foreign born and emigrated him/herself (first generation) or whether the person was born in the country of destination to foreign-born parents (second generation). In the overwhelming majority of studies of integration, the second generation is more integrated than the first. This gives rise to the expectation that in our study of immigrant youth, we might witness a relatively advanced level of integration (as compared to studies that focus on adults).

(p.13) The hypothesis that ‘higher’ generations are more integrated than lower can be derived straightforwardly from the exposure idea. Newly arrived immigrant children are unlikely to have had the time to integrate, while the second generation in many instances could be presumed to be almost indistinguishable from the majority—typically, they speak the language of the receiving country fluently and are familiar with destination-country institutional characteristics, especially the school system. Note, however, that a fair share of minority students in our study who are foreign-born migrated with their parents in early childhood. Thus, they are not comparable with first-generation immigrants in the adult population, but rather somewhere in between the conventional first and second generation (Chapter 3). For certain outcomes finer distinctions with regard to the age of arrival may reveal more gradual exposure effects among the first generation, and in some of the thematic chapters we therefore follow the convention of identifying subgroups according to age of arrival.

Theoretically, it is obvious that exposure is more complex than the concept of immigrant generation, or more generally, than the time spent in the country of destination alone. Many colonial immigrants may, for example, have been exposed to the receiving-country culture and language, and many would have been acquainted with the Anglo-Saxon (popular) culture and the English language before migrating; the same goes for German immigrants from German-speaking parts of Eastern Europe. Partly, considerations like these can be covered by the origin influences discussed in Section 1.3.1.

Conversely, an important confinement of exposure over time and generation can be ethnic segregation because it delimits the opportunities for inter-ethnic contact. Such contacts are a crucial part of the exposure mechanism in two ways: by learning to know the majority population, immigrants get socialised into the destination-country culture and learn its language; and increasing inter-ethnic relations ensure this culture, as well as destination-country institutions, adapts to immigrants’ cultures. Children of immigrants may also be more or less shielded from the destination-country culture through active protection by parents. In fact, primary (parent-to-child) socialisation is perhaps the most powerful force counterbalancing the presumed gradual immersion of immigrant children into the destination-country culture. Immigrant parents often have strong incentives to preserve their origin culture, and through norm-setting and transmission of values and behaviour, they also possess effective means to do so.

Beyond the proper first, second, third and so on generations, there are also more complex configurations of ancestry (see Chapter 3) providing specific constellations of exposure. A family composition that is likely to offer almost a ‘full exposure’ is a ‘mixed’ or ‘intermarried’ family, where one parent is foreign born and the other native born with native-born parents; children from these families are therefore more likely to be integrated than the second generation (and perhaps quite similar to the majority group). However, in cases with visible minorities in particular, it is possible that they may still encounter prejudice and (p.14) discrimination, so this mixed group may not, on average, reach total similarity with the majority population.

Another theoretically interesting group is one where children have one foreign-born parent, and the other parent born in the destination country but to foreign-born parents (we call this, lacking a better term, ‘transnational’ families; see Chapter 3). We hypothesise that for this group exposure to destination-country culture is more limited than in ethnically mixed families, and that they exhibit an integration pattern more similar to the second generation. The reason is that those of the second generation who choose to form a family with an immigrant are overall less likely to be integrated themselves, and less likely to expose their children to the receiving-country culture.

What general insights about the integration of minority and majority youth in our four European societies can be derived from our empirical analyses? What can we say about the comparisons between majority and minority origin groups, between destination countries and between immigrant generations? In this final section, we draw more general conclusions from the summaries presented earlier in the chapter, and we reflect on the pattern that stands out from the thematic chapter analyses.

1.6.1 Similar, yet different

The issue of immigrant integration is hotly debated and highly emotionalised in our receiving societies. A key aim of our analytical-empirical analysis is to present (p.21) scientifically derived facts about the lives, attitudes and behaviour of children from both an immigrant and a majority background, and their interrelations. The tone of the popular debate is often fairly pessimistic: some tend to see children of immigrant origin as victims of reduced opportunities, structural discrimination and limited support, whereas others consider this group a problem, claiming that they are incapable or unwilling to integrate, maintaining a value system that clashes with that of the majority.

Does this pessimistic view hold up to empirical scrutiny? No—at least, not unequivocally. Our first overall conclusion is that children of immigrants are in many ways fairly similar to youth of majority origin. Given our measures, their own economy is not very different from the majority’s (although their parents’ economy differs) (Chapter 4), their family relations are weaker than the majority’s for some groups but stronger for others (Chapter 6), they have friends in and outside of school (Chapter 7), delinquent behaviour does not differ much to the majority (Chapter 13), and their mental well-being is even somewhat higher than that of their peers with native-born parents, suggesting that their psychological and sociocultural adaptation is, on average, unproblematic (Chapter 14). Drawing on other analyses based on our CILS4EU data, we also know that children of immigrants do not appear to be much more unfairly treated than majority youth (Jonsson et al. 2012), nor are they more subject to bullying or social exclusion (Plenty & Jonsson 2017). In addition, they have high aspirations and optimistic plans for the future (Alm & Låftman 2016; Salikutluk 2016; Dollmann 2017). Recent studies on other data show that the structural integration into the host countries’ school systems is mixed: while their overall achievement level is somewhat lower than the majority’s, their educational aspirations are higher (e.g. Heath & Brinbaum 2014).

Clearly, the dystopian image of immigrant hopelessness and exclusion finds little support in our data, and, indeed, in our own experiences of the schools we visited when conducting the surveys. While there may be pockets of despair, while socioeconomic and ethnic segregation certainly exist, and while subgroups among ethnic minorities are among the most vulnerable in our societies, the overall living conditions and well-being of children of immigrant origin are, in many important respects, on par with those of majority youth.

It is possible, of course, that some problems specific for children of immigrants may occur later during adolescence or upon entering the labour market— our study concerns 14- to 15-year-olds. For example, other studies suggest that some ethnic differences become more pronounced at older ages, such as the over-representation of immigrants in registered crime (e.g. BRÅ 2005; Blom & Jennissen 2014), and ethnic discrimination in the housing and labour market (e.g. Riach & Rich 2002). It nevertheless appears an encouraging finding that, among young people at a formative age, common youth problems do not have a systematic ethnic bias.

As similar as youth of different ethnicities are in many respects, in other ways they are remarkably different. Most striking perhaps is that their religiosity differs (p.22) sharply (Chapter 10). While in the Netherlands and Sweden religiosity among majority youth is very low—barely more than 5% in our study claim that religion means a lot to them—the religiosity among youth coming from an immigrant background is exceptionally high. Over 50% say that religion means a lot to them, and our estimates of religious practice—measured as frequency of praying and visiting places of worship—corroborate this result. Attitudes and values also reflect such differences, for example, when it comes to tolerance for sexual minorities, the right to abortion and cohabitation without marriage (Chapter 12), as well as attitudes to gender equality (Jonsson et al. 2012). In addition, and not surprisingly, children of immigrants identify less than the majority with the destination country, although they do not ‘compensate’ by identifying more strongly with their (or their parents’) country of origin (Chapter 11).

These differences are seconded (and perhaps influenced) by the physical and social segregation between children of majority and minority background. Our results clearly show that those of majority origin rarely make friends with those of a minority background, either in their neighbourhood or in schools (Chapter 7). Those of immigrant origin actually have more inter-ethnic friendship ties, and a large majority do socialise regularly with majority peers. However, this is because there are more majority children to meet: their preferences are strongly in favour of in-group members, especially co-ethnics. The resulting ethnic segregation adds another layer to the segregation that is partly created by different resources in their families, reflected in residential and school segregation (Chapter 5). These facts are hardly surprising, even if the sheer degree of physical and social segregation is notable.

Thus, the picture we get from studying integration among youth of different origins is ultimately ambiguous. Adolescents of majority and minority origin are so similar, yet so different—in some aspects quite close, in others very distant from each other. Are the differences that we find in the social and cultural dimensions, in combination with the strong segregation in friendships, so large that we should conclude that children of immigrant and majority origin live ‘parallel lives’? The answer to that question lies, we believe, in the eye of the beholder: taking the multidimensionality into account defies any simplistic interpretations of the degree of integration.

1.6.2 Origin group influences: Do some immigrant groups integrate faster?

We find, as expected, a large heterogeneity within the immigrant group—in itself a strong case for avoiding general claims about immigrants or integration. While the summaries of the chapters in Section 1.5 already hinted at some of the most striking group variations, Table 1.2 aims at a more systematic summary of the findings for origin groups (rows in the table) across all chapters (columns in the table). The entries are based on the appendix tables to the empirical chapters; they are somewhat broad in order to present an accessible overview, consisting of only (p.23) minus, zero or plus signs, where zeros represent similarity, that is, integration (blank cells mean that the origin group was not distinguished in the analysis).² Note that the underlying analyses are not necessarily identical because different models had to be applied for different problems, but as a rule they take into account compositional differences across groups when it comes to sex, age and immigrant generation.

As can be seen in Table 1.2, origin group variation is borne out by several of our chapters. For example, the difference between children of immigrants and the majority when it comes to family ties—as analysed by the relations to ‘absent’ fathers—is impossible to generalise because they differ profoundly between ethnic groups: some minority groups have stronger ties than the majority, others weaker, and yet others are very similar to the majority group (Chapter 6).

Another example of diversity within the minority population is religiosity, where some groups—those with origins in the MENA region especially—are much more religious than the majority, whereas those of Western and Eastern European as well as Asian origin share with the majority a rather low level of religiosity (Chapter 10). A similar story holds for segregation where we find that although all ‘major’ minority groups we identify live in segregated neighbourhoods and go to segregated schools, the degree of segregation differs substantially between different minorities (Chapter 5). The conclusions are virtually the same for language skills (Chapter 9). The distance to the majority is, furthermore, very different when looking at liberal attitudes and tolerance of sexual minorities, but here most non-Western minority groups are far more traditional in their attitudes (Chapter 12). However, in general, minority groups are predominantly similar to each other in the areas where there is no minority–majority gap at all.

The presence of heterogeneity across origin groups does not preclude, of course, that some immigrant origin groups are more, some less, integrated than others overall. When studying Table 1.2 row-wise, we do find some support for our expectations with regard to origin group influences. First, children of Western origin are, to a large extent, either indistinguishable to the majority population, or resembling them (though also these groups are physically segregated). Secondly, children of (East/South) Asian, Caribbean and East European origin have no clear profile in this respect: they are different from the majority in some ways, but similar to them in others (although the ex-Yugoslavian groups appear overall to be more dissimilar). Our assumptions that the South and East Asian origin groups would be characterised by relatively low levels of integration are thus not well supported.

Thirdly, the groups with the greatest distance to the majority are youth from the MENA region and from sub-Saharan Africa. However, it should be noted that even these groups are not different in all respects. They do stand out by having, overall, different religious beliefs and stronger religiosity, more closed social (p.24)

(p.27) circles, slower language acquisition, as well as different values and attitudes. However, when it comes to delinquency or general well-being, they are fairly similar to the majority groups; their well-being is even better. In addition, we know from other studies that their educational aspirations are as high as most other immigrant groups (Heath & Brinbaum 2014; Salikutluk 2016), although they suffer from poorer school performance. Again, our conclusion must be that the integration of these groups—even if relatively slow—is multifaceted.

1.6.3 Exposure effects: Integration across immigrant generations

In Section 1.6.2, we addressed the degree of integration by comparing averages in different outcomes between minority and majority youth (Table 1.2). But integration as a process is more accurately analysed as a change in these differences, following the assumption that exposure leads to similarity. A common picture of assimilation in the US literature, for example, is one in which each immigrant generation becomes more and more similar to the (constantly changing) blend of US cultures (e.g. Alba & Nee 2003). Because generational status represents both the ‘ancestral distance’ from the country of origin and exposure to and familiarity with the receiving country, there is, as outlined, much reason to expect that ‘higher’ generations would be more integrated.

In line with our assumptions, we find that some outcomes in our sample of 14- to 15-year-olds differ according to generational status, namely family ties (Chapter 6), language proficiency (Chapter 9) and national identity (Chapter 11): for these aspects, integration appears, overall, to be a matter of time—although we cannot infer from our data that there will eventually be ‘full’ integration.

However, maybe the most unexpected finding in our study is the absence of an exposure effect for other crucial aspects of integration. In two cases, delinquency (Chapter 13) and mental well-being (Chapter 14), this might simply be a result of the fact that there were no large starting gaps to be closed. However, exposure effects are also absent for aspects of integration where there are huge differences to report. There is little sign, for example, that youth of the second generation have more inter-ethnic friends than the first generation (Chapter 7); the same goes for religiosity (Chapter 10) and for liberal attitudes and tolerance towards sexual minorities (Chapter 12). In all these cases the second generation is not (or only sporadically) more integrated than the first. Because a part of our ‘first generation’ adolescents arrived relatively young (40% before school-age; see Table 3.1), and thus are closer to the second generation than what is the case in adult samples, the similarity across generations is perhaps not as surprising as the great differences between the majority and the second generation.³

(p.28) We should note that most of our analyses are based on ‘synthetic’ generations, that is, comparing generations cross-sectionally. Therefore, there could be remaining compositional differences between generations (e.g. because there is a change in immigration patterns over time), although our analyses do include control variables that should take care of some of these differences. But even where chapters can account for ‘real’ generations, that is, parent–child effects and differences, the intergenerational stability of the distance to the majority is astonishing. Chapter 10 reveals that religiosity in Muslim (but not Christian) groups is not at all weaker among children of immigrants than among their parents (see also Jacob & Kalter 2013). Likewise, Chapter 12 shows that intergenerational transmission of non-liberal sexual values is extremely strong, especially when the religiosity in a family is high.

We also hypothesised that children of intermarriages and transnational marriages would show slightly more integration than the second generation, because their exposure to the host-country institutions and culture is likely to be greater— but that they would still differ from the majority population. Our expectations here were mostly supported. Children who have one foreign-born and one host-country-born parent (‘intermarriage’) are closer to the values, ethnic identity and religiosity of the majority than those who have two foreign-born parents—but they are similar to this group in inter-ethnic friendship, mental well-being and delinquency. Those with one foreign-born and one host-country-born second-generation parent (‘transnational’) are on the whole quite similar to the second generation, with one systematic exception: they have more inter-ethnic friends. We should note that the ethnic origin group composition of these ethnically mixed groups might influence the results (intermarriage between native-born people and those from geographically and culturally remote regions is rare), although most analyses do control for immigrant origin group.

A careful overall conclusion is that exposure appears to have some effect on integration, but the more surprising result is the absence of a difference between the first and second generation in some crucial aspects of integration. This pattern still needs to be corroborated with longitudinal data that can follow individuals over time, although such data that cover a long time and are of sufficient size are hard to come by.

1.6.4 Destination influences: Comparing receiving-country integration patterns

Because we launched our survey in four countries using identical questionnaires, we have an unusually strong basis for assessing differences in the integration process across these countries. From the statistical descriptions in Chapter 2, one could expect Sweden to provide somewhat more fertile ground for integration and England to provide somewhat less, with Germany and the Netherlands somewhere in-between. This assumption could be derived from analyses of immigrant and integration policy, public attitudes towards immigrants, and from variations (p.29) in welfare state provision, inequality and child friendliness across our receiving countries—although other mechanisms may well work in the opposite direction, and previous research has not produced much evidence on what type of policies would best promote integration.

To sum up the results for destination countries from Table 1.2, we found no systematic differences in integration across our host countries—and in several respects we did not find much variation whatsoever. When it comes to children’s own economic resources, mental well-being, delinquency, religiosity and liberal values, there are few destination-country differences in integration to report at all. While this lack of variation could be seen as a drawback in identifying policy effects, it is an advantage when considering our previous conclusions: the fact that we find much the same pattern of ethnic origin group and generational influences across destination countries suggests that they could also be generalised to other Western European countries.

We need to caution against drawing firm conclusions about differences in integration across destination countries because of the variation in immigrant group composition (Chapter 2) as well as heterogeneity within nominally similar immigrant groups (Chapter 3), even though we attempt in Table 1.2 to come to grips with this by showing results for relatively similar origin groups. With these caveats in mind, the main impression from our analyses is nevertheless one of destination-country similarity in the integration of youth of immigrant background.

We do, however, find host-country differences along some dimensions. One is language acquisition, where England, as anticipated, has a substantial advantage over our other host countries (Chapter 9). Importantly, this is very likely to have positive effects on other types of integration. Our results with regard to inter-ethnic attitudes (Chapter 8) give some support for the idea that the reception of immigrants is more favourable in Sweden, as the minority groups have an unusually positive view of the majority population there. However, we do not find similar support in our analysis of ‘national’ ethnic identity (Chapter 11), where Sweden appears to lag somewhat behind the other countries, and where England shows the smallest differences in destination-country identification rates between the majority and minority groups (to some extent because of low values among the majority). Once we adjust, as far as we can, for compositional differences in immigrant populations, the destination-country differences are not great, however.

1.6.5 The mechanisms behind integration

The situation children of immigrants are faced with across destination countries is not only relatively similar overall in descriptive terms, as noted in section 1.6.4, but also when it comes to the mechanisms behind it. Table 1.3 reports some stylised results from the individual chapters. We have produced this table, still necessarily selective, mostly by judging in qualitative terms how much a theoretically informed predictor—say, socioeconomic background—changes the ethnic differences in the stepwise models of the chapters’ appendix tables. (p.30)

(p.31) Table 1.3 shows that, by and large, the contribution of major theoretically expected mechanisms behind ethnic group differences varies, if at all, only to a small extent. The most striking exception is the difference between Germany and Sweden on the one hand, and England and the Netherlands on the other, when it comes to the role of SES and cognitive skills in explaining language skill differences (Chapter 9). This might well have to do with the fact that England and the Netherlands both, due to their colonial history, have high numbers of immigrant children from origin countries where English and Dutch are spoken. Furthermore, there is a tendency, as one could also expect, for socioeconomic factors to be somewhat more important for some other outcomes (Chapters 4 and 6) in Sweden, where the comparatively large group of immigrant refugees lag behind in terms of employment and incomes. There is also a relatively weak impact of migration-relation factors on explaining father absence (Chapter 6) in England, and a relatively weak impact of diversity on inter-ethnic attitudes in Germany (Chapter 8).

1.6.6 The role of policy for integration

The fact that we do not find no systematic destination-country differences in integration, and little variation in the mechanisms behind, suggests that there is much in the integration process that is generic to the experience of immigrant families. Hence, the role of policies in integration-related debates might be somewhat overrated. However, we would certainly not subscribe to the view that the destination country and its immigrant and integration policies do not matter at all for immigrants. One interesting piece of evidence to the contrary comes from looking at our analyses from another angle. We have regularly (and reasonably) compared the situation of children of immigrants with that of the majority within receiving countries. If we shift the perspective and compare minority groups with the majority in another country, we arguably find some evidence that the receiving country (and its institutions, policy and people) exerts an influence on immigrants. For example, even though youth of non-European origins lag behind the Swedish majority in liberal attitudes, tolerance and attitudes to gender equality, they are just, or almost (depending on the indicator), as ‘liberal’ as the majority population in Germany and (often) the Netherlands (Chapter 12; Jonsson et al. 2012). It is possible that this is due to the generally more liberal sentiments in Sweden or the more emancipated position of Swedish women in the labour market, in politics and so on—at least, it is entirely unlikely that selective origin group effects should explain this pattern. On the other hand, by the same token we would have expected religiosity among children of immigrants to be lower in the Netherlands and Sweden due to the marginal role religion plays for the majority population in these countries. But although the level of religiosity is slightly lower among minorities in Sweden when compared to the other destination countries, there is no systematic difference in immigrant-group religiosity across survey countries (Chapter 10).

(p.32) So, even though our analyses suggest quite small differences between our receiving countries, this does not rule out the integration influences of policies. In fact, our results regarding underlying mechanisms indicate that it may be more effective for integration to implement more traditional policies aiming at reducing inequality and segregation and promoting upward social mobility for immigrant groups. On the other hand, it could be argued that if such policies were effective, we would already have been able to register swifter integration in Sweden than in our other three countries. However, Sweden has not, as is evident from Chapter 5, managed to combat ethnic segregation, and still has, mostly due to the relatively large representation of refugees of its immigrant group, large ethnic differences in unemployment and poverty rates (Chapter 2; Jonsson & Mood 2017). Sweden also has far-reaching multicultural policies in place, which have been argued to impede integration when combined with generous welfare state provision (Koopmans 2010).

While our study has been able to offer important new insights into destination-country similarities in integration, our underlying analyses have been of a descriptive nature. Scrutinising the role of policy would require a more advanced analysis where policies were measured with high precision and their effects carefully identified, something that is beyond our present aims.

1.6.7 Integration and inertia: The need for realistic expectations

Our results provide ample evidence that the ethnic integration of youth in diverse European societies seems well on its way in many respects. But they also point to a considerable sluggishness of integration in some dimensions and for some groups. To better understand why this might be, it is helpful to consider some of the main mechanisms behind origin group effects in the respective domains again (see Table 1.3), and ask how likely they are to change.

As we stressed in our theoretical discussion, socioeconomic resources and primary socialisation effects may well be behind many origin group variations. This proves to be empirically important in several, though not all, chapters; for some aspects the influence on integration is considerable, among them the relation between socioeconomic origin (SES) and language acquisition in some destination countries (Chapter 9) and the influence of both SES and socialisation for liberal values in all four countries (Chapter 12). In the case of liberal values, socialisation effects are even stronger when the SES of the family is low, which is the case for many minority groups. Low intergenerational social mobility will therefore not only slow down processes of structural assimilation (e.g. Kalter et al. 2007), but also cultural and social integration. Thus, there is a clear link between persistent social stratification and barriers to integration.

Next to indirect transmission via socioeconomic background, two chapters (Chapter 10 and 12) are able to show a notably more direct intergenerational transmission of cultural orientations from parents to children, such as religiosity and other values. This is arguably a primary socialisation, or role-modelling (p.33) effect, which could counterbalance the exposure to destination-country beliefs and values. There are also indirect socialisation effects. Religiosity is, for example, one important reason behind the minority–majority gap in liberal values and tolerance, and religiosity also plays a role for externalising behaviour and family relations—although it is interesting to note that in three out of four receiving countries religious affiliation does not matter with regard to identification with the host country once religiosity is taken into account (Chapter 11).

The analyses in this book also show that the opportunity structure is a major cause behind friendship choices and negative inter-ethnic attitudes—the more ethnically mixed the environment the more likely inter-ethnic friendships will emerge (Chapter 7) and the weaker ethnic prejudice (Chapter 8). Hence, we find support for the assumption that ethnic segregation hampers integration. For friendships, ethnic homophily contributes a further mechanism, which can be seen as a reflection of the basic human tendency of individuals to befriend those who are similar to themselves (McPherson et al. 2001).

From the summary above, it appears that some of the fundamental mechanisms behind existing differences between children of immigrants and majority youth—socioeconomic inequality, segregation, socialisation, and homophily— are pervasive. This suggests that we should not be entirely surprised that change is slow for some groups and differences to the majority large for some aspects of integration. The pace of change should, we believe, be viewed from the perspective of the enormous differences between many origin societies and the typical West European destination societies: several of the more recent and largest immigrant groups come from impoverished regions of the world, where politics, religious beliefs, family and kin relations, lifestyles and values could hardly be more different from the affluent, liberal-democratic, individual-oriented and predominantly secular societies to which they move. Thus, the integration processes of both immigrants and destination-country institutions and populations are likely to be both demanding and protracted.

Our conclusion from this is that we all—researchers, politicians and people in general—need to be aware of the inertia produced by basic societal mechanisms working against integration in order to get a realistic idea of how likely or easy it is to achieve integration. But evaluating the progress in this respect is also a matter of deciding how to define integration: Do we mean that ethnic minority children—or perhaps only those of the second generation—have the same attitudes, behaviour and so on, as majority children in general? Or the same as the majority children with a similar socioeconomic background? Or the same as majority children with a similar socioeconomic background, accounting for the fact that phenomena like homophily exist for all kinds of individual characteristics and in all societies? For those who think we ought to strive towards integration, the empirical results in this volume often look quite optimistic when taking such factors into account, but rather more pessimistic when not.

In conclusion, we should not be surprised to see lingering differences between children of immigrants and the majority population, as exemplified by (p.34) the weak exposure effects we could report. As some support for this ‘realistic’ stance we could also call upon our results showing no systematic or sizeable differences in integration across our four survey societies despite variation in policies and institutional settings, a finding supported by the results in Alba & Foner (2015) when comparing integration outcomes across the Atlantic, and by Berry et al. (2006) when analysing adaptation from a psychological perspective.

The conclusions of our analyses in this book, presented above, lead into questions about the future integration of our sample of 14- to 15-year-olds, questions possible to address by utilising the longitudinal feature of our data. The natural next step is to expand on the issue of structural integration, asking whether the lingering differences we find between ethnic minority groups and the majority, as well as among minority groups, will have an impact on educational and occupational careers.