Migration background – Statistical classification and the problem of implicitly ethnicising categorisation in educational contexts

Introduction

Within only a few years, ‘migration background’ has become one of the key ethnicity-related categories in German-speaking countries. ¹ It is among the standard variables used in current official statistics and the social sciences. It is also omnipresent in everyday political and public discourses. This article aims to clarify aspects of this stellar career by inquiring into the emergence and functioning of ‘migration background’ as an educational category. How is the category of ‘migration background’ in its current conventional understanding informed by educational (and other extra-statistical) discourses and as what kind of resource does it serve for social actors such as teachers for mastering their everyday professional life (Bowker and Star, 1999; Desrosières, 1998; Lamont, 2014; Wimmer, 2013)?

The central objective is to further our understanding of categorisation processes that can be characterised as implicitly ethnicising because they do not directly or openly refer to ethnic identity. Rather, they define national or ethnic belonging by implication, via discursive connotations and de-facto correlations with ethnicised social relations (Elrick and Schwartzman, 2015). Relating findings from various empirical studies conducted over the past years, I argue that the rise of ‘migration background’ in educational contexts mirrors a decisive shift from class-centred to culturalising problematisations of educational inequalities.

The following section briefly introduces the conceptual foundation for discussing ‘migration background’ as a statistical and at the same time social category linked to fields of professional practice. The subsequent sections depict ‘migration background’ as an educational category, proceeding from the category’s past to its present semiotics and finally to possible future developments. The discussion of the emergence of ‘migration background’ focuses on the transformation of political and educational discourses over the post-Second World War decades. Once established, ‘migration background’ developed a particular semiotic structure that informs its usage in current educational settings. I argue that how and why professional educational actors make use of ‘migration background’ has important consequences for the problematisation of educational phenomena and inequalities. At the same time, ‘migration background’s’ semiotic structure points to important tensions that may well lead to its replacement by other forms of categorisation. The last section discusses these tendencies and their implications.

Statistical classification and educational boundary making

It is widely acknowledged in academic debates that ethnicity should be treated as a social construct (Brubaker, 2002). Interrelated forms of classification in various fields of social practice are relevant for understanding how the statistical measurement of ethnicity and migration is informed by struggles over boundaries and belonging and, in turn, affects structures of domination and inequality as well as processes of subject formation (e.g. Kertzer and Arel, 2000; Nobles, 2000; Supik, 2014a, 2014b). The need to take practices and discourses in various social fields into account is arguably even more pressing for categories that are ethnicising, but only implicitly so, such as ‘migration background’. ‘Migration background’ was introduced as an official statistical category in Germany in 2005. Compared to other systems of classifying ethnicity and/or migration (Morning, 2008), it is noteworthy that its official understanding does not refer to any explicit ethnic marker: According to the current definition of the German Federal Statistics Office, ² whoever was born as a non-German citizen or born to at least one parent without German citizenship has a ‘migration background’. The usage of the category in media, politics and public discourses, however, clearly denotes a specific part of the population: othered migrants, including but not restricted to post-guestworker populations. ‘Migration background’ has accordingly been characterised as ‘racialised’ (Gehring, 2016) and ‘homogenised social category’ that is used to constitute ethnic belonging (Elrick and Schwartzman, 2015).

This article starts from the assumption that this very tension between different usages of the term is crucial for an in-depth understanding of the emergence, the development and the effects of ‘migration background’. It asks how the category of ‘migration background’ has been used for defining and handling educational situations, in turn lending it relevance as a statistical category and contributing to its establishment and stabilisation. This research interest is inspired by boundary making perspectives (Lamont et al., 2015; Wimmer, 2013), by investigations of social classification in Science and Technology Studies (Bowker and Star, 1999), and by studies of statistical quantification in the context of the ‘sociology of conventions’ (Desrosières, 1998; Diaz-Bone and Didier, 2016).

Recent studies on ethnic boundary making investigate the interplay of symbolic orders and everyday practices (Lamont et al., 2015), with a focus on how social actors strategically relate to explicit ethnic boundaries (Wimmer, 2013: 49). This emphasis on strategic actions intentionally directed at ethnic systems of classification is not shared in this paper for two intertwined reasons: first, because the category under consideration is not an explicitly ethnicising category; second, because educational boundary work usually does not per se qualify as ‘ethnic agency’. Of course, a lot of boundary making happens in educational contexts, some of which may also be openly ethnicising (e.g. Duemmler et al., 2010; Visser, 2016). Overwhelmingly, however, professional educational boundary making primarily implies other forms of categorisation, such as grading students, assessing behaviour or diagnosing learning disorders.

If we want to understand how processes of ethnicisation emerge and function in such contexts, we need to take the interplay between these professional forms of educational categorisation with wider societal ‘information infrastructures' into account (Bowker and Star, 1999). This interplay works in both directions. On the one hand, educational actors are (even if indirectly) involved in the formation of statistical classifications and the definition of their ‘equivalence principles’ (Desrosières, 1998). The following section discusses how, over the past decades, educational debates may have served as an entry-point for forms of implicit ethnicisation that are black-boxed in the official statistical category of ‘migration background’. On the other hand, educational actors draw on established information infrastructures for making sense of their everyday work contexts. They of course seldom invest the purely criterion-based statistical definition of ‘migration background’, but rather employ forms of ‘prototype classification' (Rosch, 1983) that, in the case of ‘migration background’, re-activate black-boxed ethnicising connotations of the statistical classification.

The arguments presented in the following draw on findings from a number of studies conducted between 2012 and 2016. The next section focuses on the level of educational discourses in their political-economic contexts. It presents a reading of the emergence of ‘migration background’ that draws on a variety of text sources, including parliamentary material and educational texts from the 1950s to 2000s (for more details see Horvath, 2017). The patterns that become obvious have left their traces on contemporary social understandings of the category, including their usage in pedagogical contexts. The discussion of these educational usages in the following section (‘The semiotics of migration background’) builds on two sources of empirical data: first, qualitative interviews with primary and secondary school teachers from Southern Germany are used to decipher current educational understandings of ‘migration background’ and how they are anchored in the structures of everyday professional situations. In these interviews, teachers had been asked to provide in-depth characterisations of their school and their pupils, as well as to outline their own subjective theories of educational success and failure. The analysis of this material concentrated first on identifying different ‘cognitive forms’ of categorisation and classification and second on how these cognitive forms are linked to educational problematisations and to strategies of defining and handling uncertain educational situations. More detailed information on these interviews can be found elsewhere (Horvath, 2019). Second, selected results from a mixed-methods study in two schools for ‘highly-gifted’ students are reported (Horvath, 2014a). This study included, among others, qualitative interviews with teachers and a quantitative survey among students. It is used in this article to illustrate how educational narratives can become fixed around an implicitly ethnicising category (Horvath, 2018), leading to important blind spots and biased forms of perceiving and problematising educational inequalities.

The argument in this article applies to the case of Germany. There are of course similarities to other national contexts, most importantly Austria and Switzerland (regarding, e.g., migration histories, citizenship regimes and political-economic contexts). Nonetheless, any generalisation should be done with care, since both the schooling systems and the official systems of statistical classification vary between these countries.

Reordering belonging: The emergence of migration background in educational contexts

The emergence of the statistical category of ‘migration background’ does not simply reflect some kind of self-evident need to identify and count a growing migrant population. Rather, it is expressive of a complex discursive and social transformation that followed the post-Second World War era of Fordist class compromises (Amelina et al., 2016). In this section, I relate aspects of these political-economic transformations to educational discourses in order to substantiate my claim that the career of ‘migration background’ reflects a shift from class-centred to culturalising problematisations in the field of education.

The extension of civic, political and social rights and the concomitant educational expansion after 1945 carried the promise of increasing participation and equal life chances for all citizens (Marshall, 1950). This development could be seen across Western Europe – Germany is not a special case in this regard. By the mid-1960s, it had become clear that reality did not live up to these egalitarian promises. Educational research reacted by investigating the subtle mechanisms through which class relations were translated into educational inequalities. Bourdieu and Passeron (1977) and Cicourel and Kitsuse (1963) are among the many prominent social scientists who contributed to these debates. These developments also clearly affected educational discourses, in German-speaking countries as in the rest of Western Europe. For example, notions of ‘giftedness’ and ‘talent’ were widely rejected during these years; instead, problematisations that centred on social class became dominant (for example Rolff, 1980; Roth, 1968). One important feature of the class-focus during this period is its strong link to national citizenship: educational inequalities were criticised with reference to members of the same nation-state and therefore not considered legitimate (e.g. Dahrendorf, 1965). The emphasis was on social class, but nationality underhandedly was a key category in academic and educational discourses, mirroring the increasing factual relevance of national citizenship in the context of the post-Second World War welfare state (Wimmer and Glick-Schiller, 2003).

The crisis tendencies of this political-economic constellation eventually led to the labour migration programmes of the 1950s to 1970s (Castles and Kosack, 1968): mass industry was in urgent need for low-paid workers, but ‘native’ workers were less and less willing to accept the dirty, dangerous and demeaning conditions these jobs offered. The recruitment of foreign workers, construed as the ethno-national Other and deprived of important civic, political and social rights, allowed to sustain the post-Second World War class compromise for a certain period (Zolberg, 1991). As a result, national labour markets became effectively segmented. The labour market segmentation was soon followed by the emergence of new ethnic minorities (Castles et al., 1984) – this was the turning point at which educational systems became involved. Across Western Europe, the problem of how to deal with the ‘second generation’ arose.

With a lag time of a few years, the establishment of guestworker regimes left important traces in both national statistics and educational discourses. During these years, the quasi-legal category of ‘Ausländer’ (foreigner) emerged as key statistical variable. In contrast to explicitly ethnicising variables (such as ‘Volk’), the distinction of citizen VS non-citizen could be considered legitimate without question. After all, it seemed devoid of any racialising underpinnings reminiscent of the dark past of National Socialism – notwithstanding its clearly ethnicising connotations in everyday discourses. Further, the ‘Ausländer’-criterion worked as valid proxy for capturing the specific group of guestworker populations, due to the restrictiveness of the German citizenship regime (hindering naturalisation) and the relative homogeneity of migration patterns at the time.

In educational contexts, the seemingly trivial category of ‘Ausländer’ actually marked a complex social and symbolic boundary. Schools soon faced the need to cope with a new group of ‘problematic’ students. The form in which these children were framed in contemporary educational texts mirrors important aspects of the wider developments of post-Second World War public and political discourses. A specific pedagogical approach was developed under the heading of ‘Ausländerpädagogik’ (‘foreigner pedagogics’) (see, for example, Ruhloff, 1982). ‘Ausländerpädagogik’ focused on the deficits of guestworker children. Cultural backwardness and social disadvantages were deeply entangled in this understanding. In principle, ‘Ausländerpädagogik’ left room for both: post-colonial othering and more emancipatory analyses of changing class relations. In practice, however, this pedagogical programme paid little attention to class-related aspects sui generis, concentrating on problematisations of insufficient language skills and, at best, rather stigmatising versions of compensatory education (Mecheril et al., 2010). Over time, ethnicising version of the category ‘Ausländer’ became salient in educational discourses, shifting the tone of thematisations of problematic student populations from class-centred to culturalising problematisations. This development eventually culminated in the replacing of ‘Ausländerpädagogik’ by diversity-oriented programmes of intercultural pedagogics (Auernheimer, 1995).

Up to the 1980s, the category of ‘Ausländer’ worked sufficiently well to capture the groups that educational researchers, school administrators and teachers had in mind. By and by, it lost this capacity. By the 1990s, a large share of former guestworkers and especially of their descendants had acquired the citizenship of their resident state; they were therefore no longer identifiable by legal citizenship as classification principle. The massive immigration of so-called ethnic Germans from the former Soviet Union (who had relatively easy access to German citizenship) after 1989 added to the resulting undercoverage problem. The decreasing usefulness of the Ausländer-category is clearly noticeable in Figure 1. The simultaneous rise of its successor is also obvious.

OPEN IN VIEWER

Figure 1.

A lexicometric analysis of German-speaking books by Google’s ngram-viewer.

School administrators were among the first who demanded the introduction of new forms of counting the presumably problematic parts of their studentship. Their interest was primarily to secure resources: new categories were needed to illustrate the financial needs of schools in disadvantaged neighbourhoods. It thus comes as no surprise that ‘migration background’ was officially mentioned for the first time in a parliamentary report on education (Perchinig and Troger, 2011; Supik, 2014b: 19). At about the same time, educational researchers started to use the category (e.g. Granato, 2000). ‘Migration background’ offers an elegant solution to the administrative, political and analytical objectives of the different groups of actors involved. In several regards it does not break with the principle logic of the ‘Ausländer’-category. As a statistical category, it seems to measure quasi-objective characteristics. As an educational category, however, ‘migration background’ actually aims to capture the ethnicising excess of the former ‘Ausländer’-category by identifying those migrant Others who can no longer be classified by their citizenship status.

These few glimpses at its social and discursive history may serve as a first illustration of the multiple and complicated trajectories that gave rise to ‘migration background’ as a very specific form of pseudo-ethnic classification. In contrast to the ‘Ausländer’-category and its long political and statistical history, ‘migration background’ was rather abruptly introduced as an element of official systems of statistical classification, only after it had emerged in fields of practice such as education. ‘Migration background’, in other words, was not introduced as a statistical category following strategic agency by ethnic actors. Practical categorisation needs in professional contexts (such as education) played the leading role.

The semiotics of migration background

From hindsight, the initial rise of the ‘Ausländer’-category and its subsequent substitution by ‘migration background’ mark important ruptures in educational discourses of the post-Second World War era. In the following, the focus shifts from the past of ‘migration background’ to its current semiotic structure. Once established, how (and why) is ‘migration background’ used in educational contexts?

An early end to a stellar career? What educational contexts tell us about possible futures of statistical classification

Since different fields of practice follow different logics and are organised in different ways, it comes as no surprise that statistical and educational usages of the concept of ‘migration background’ diverge significantly (Bowker and Star, 1999). It would be misleading to see this divergence as some kind of misunderstanding – e.g. by criticising teachers for their inappropriate use of the category. The statistical category grew out of practical contexts, reflecting a general discursive shift. Once established, it in turn has important repercussions. The category of ‘migration background’ allows teachers to define and handle uncertain situations. Its status as an official statistical category additionally gives it a strong flavour of factuality. Teachers, administrators and researchers build on this factual character. Referring to ‘migration background’ does not require any form of explanation; it is perceived as a self-evidently relevant and meaningful variable. Currently, it also seems to delineate the minority populations that actors have in mind sufficiently well.

Its complex anchoring in different fields of practice (statistics, politics, education …) renders ‘migration background’ relatively stable. Still, ‘migration background’ is not set in stone, neither as social nor as statistical category. Similar concepts have been introduced to and abolished from official statistics throughout the world (Supik, 2014b). Accordingly, the future of ‘migration background’ is unclear. Considering the field of education, there are both stabilising and de-stabilising tendencies. One of the strongest stabilising factors is that there currently is no other category that is used across different subfields of the education system and hence allows for communication between them. This might actually be one of this category’s most important features: its ambiguous-yet-factual character allows different actors from different fields to talk about different things – using the same vocabulary to vaguely refer to ‘problematic’ groups that cannot be named explicitly lest they would have to be fully acknowledged as minority part of the nation.

At the same time, there are good reasons to expect an early end to ‘migration background’s’ so far impressive career. First, ‘migration background’ is increasingly rejected by those labelled by it, mostly because of its ethnicising and stigmatising connotations. Second, recent refugee movements have important consequences for German-speaking school systems and educational discourses. Currently, one and the same category is used for very different student biographies, be it urban post-guestworker minority populations or heavily traumatised children coming from war regions. This leads to the third and most important point for the context of this article: The mismatch between the population groups ‘meant’ by the social category of ‘migration background’ and those actually identified by the statistical category will presumably increase over the coming years. The official German definition covers an unclear proportion of third generation ‘migrants’ (the definitions in other countries are restricted to the second generation). The greater the share of the third (or even fourth) generation in society, the more fuzziness and the less accuracy is to be expected. Perhaps paradoxically, the resulting under-coverage of ‘problematic’ populations by the statistical category is paralleled by a potential problem of over-coverage. As has been shown in this article, the class-related facets of ‘migration background’ play an important role for its usage as a social category. However, it seems reasonable to assume that a growing share of those formally classified as ‘having a migration background’ are not part of the population groups imagined as problematic or framed as relevant in educational problematisations. This is mainly due to new orders of borders and mobilities that have emerged against the background of the profound political-economic transformations of the past decades (Amelina et al., 2016; Horvath et al., 2017). An increasing differentiation of rights and a hierarchisation of migrant groups are among the key characteristics of this development. There is a trend towards enforced temporariness among so-called unskilled migrants, such as seasonal workers whose children almost by definition will never become part of the receiving state’s student population, while the possibility of permanent settlement is more and more reserved for more privileged groups (Horvath, 2014b).

It thus seems doubtful whether ‘migration background’ as a catch-it-all category will continue to correspond to the needs and structures of everyday educational practices. But what could come after ‘migration background’? The current trend seems to be towards the development of explicitly ethnicising classifications. Concerning wider political and economic developments, pressure to introduce ethnicising classifications may come from political institutions under the banner of anti-discrimination policies. Most importantly, however, the rise of new nationalisms and of the far-right all across Europe – from Poland to the UK and from Scandinavia to the Mediterranean – contribute to an increasing salience of culturalising and essentialising forms of othering.

Regarding current educational contexts, there are two concrete entry points for explicitly ethnicising forms of classification: language and religion. Teachers routinely single out language as the most important challenge related to ‘migration background’ for their everyday practices. Language appears to be easy to measure and has a strong flavour of neutrality and objectivity. In contrast, teachers treat classifications based on religion with more caution. Compared to language the educational relevance of religion is far less obvious. References to religious backgrounds and habits are of course widespread, but they are often qualified. Considering the strong politicisation of religion and the prevalence of anti-Islamic racism in media and politics, the more likely development in this case could be that external pressure on educational actors increases ‘to take religion seriously’. This, in turn, could lead to a growing demand for hard statistical data on the ‘religious background’ of neighbourhoods and students.

The aim of this article was to contribute to ongoing debates on the interplay of statistical and social classification by investigating the special case of ‘migration background’ in the field of education. The objective was to further our understanding of the sociology and the politics of ethnicised statistical classification by shifting attention from the effects of statistical classification to its emergence and by shedding light on mechanisms that render systems of classification stable or unstable, allowing conjectures about possible future tendencies. At the same time, this article also implicitly contributes to discussions of the consequences of the statistical measurement of ethnicity and migration. Statistical classifications may themselves hinge on processes and discourses in other fields of practice such as education – but as soon as they are established they become ‘conventional’ and thus serve to stabilise social categories across social fields (Diaz-Bone and Didier, 2016). As conventional social forms they deeply affect how we think and talk about social phenomena such as educational disadvantages.

There therefore is a persistent need to discuss about alternatives to essentialising and culturalising forms of statistical classification – which in spite of a ‘constructivist consensus’ is widely accepted as the best we can do in measuring ethnicity and migration (Morning, 2008; Horvath and Latcheva, 2019). One possible way forward is hinted at by Desrosières (2011: 52–55) who contrasts currently dominant regression analytic approaches in the field of statistics to perspectives that follow the logic of correspondence analysis. The latter basically imply to move from taxonomic variables that can be used as clear-cut dependent or independent variables in regression models to more exploratory and relational forms of generating and analysing statistical data. In this understanding, statistical analysis of ethnicised social relations should focus on representing the manifold relations between single characteristics rather than on summarising them in classifications. In the case of ‘migration background’, the statistical task would be to map configurations and correlations of migration biographies, indicators of social, political and cultural belonging, and social class, avoiding the labelling of groups for as long as possible.