Content uploaded by Marion Fischer-Neumann
Author content
All content in this area was uploaded by Marion Fischer-Neumann on Apr 23, 2018
Content may be subject to copyright.
Special Issue: New Migrants’ Socio-Cultural Integration
Between ethnic options
and ethnic boundaries –
Recent Polish and
Turkish migrants’
identification with
Germany*
Claudia Diehl
University of Konstanz, Konstanz, Germany
Marion Fischer-Neumann
University of Bamberg, Bamberg, Germany
Peter Mu
¨hlau
Trinity College Dublin, Dublin, Ireland
Abstract
We describe migrants’ early patterns of identification with the receiving society and
explain why these differ by migrants’ origins. Using longitudinal data from a novel survey
among recent immigrants from Poland and Turkey in Germany enables us to analyse the
nexus between social assimilation, ethnic boundaries and identification more directly
than previous studies.
Theoretically, we start out from assimilation theory and its assumption that migrants’
identification with the receiving country is a consequence of their preceding social and
cognitive assimilation and from the literature on ethnic boundaries.
Results suggest that Turkish new migrants start out with higher levels of identification
with Germany than Poles. Over time, however, their national identification decreases
while it increases for Poles. This is partly explained by the fact that Turkish migrants’
social assimilation stagnates; more importantly, only Turks perceive more rather than
less discrimination and value incompatibility over time.
Ethnicities
2016, Vol. 16(2) 236–260
!The Author(s) 2016
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1468796815616156
etn.sagepub.com
Corresponding author:
Claudia Diehl, University of Konstanz, Universitaetsstrasse 10, Konstanz 78465, Germany
Email: claudia.diehl@uni-konstanz.de
*
This work uses data from the international project ‘Socio-Cultural Integration Processes of New Immigrants
in Europe’ (SCIP) that was generously funded by the NORFACE Research Programme on Migration.
All authors have contributed equally to this article.
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
While both groups’ identificational integration with the receiving country thus starts
out from different conditions, they do not show a fundamentally dissimilar pattern with
respect to the consequences of assimilation and discrimination for their national iden-
tification. Yet, the negative impact of the latter is stronger for Turks than for Poles,
reflecting the greater salience of ethnic boundaries for this group.
Keywords
Assimilation, discrimination, recent migrants, identification, assimilation theory
Introduction
There is a renewed interest in migrants’ ‘emotional’ or ‘identificational’ integration
in the receiving societies (Joppke, 2007). This is not only reflected in the public
debate on some, mostly Muslim migrant groups’ alleged unwillingness to become
full and loyal members of their receiving societies. There is also a substantial body
of research in sociology, psychology and social psychology on migrants’ identity
patterns, on the factors influencing them and on their consequences, e.g. on psy-
chological well-being, out-group attitudes or political engagement (Fischer-
Neumann, 2014; Verkuyten and Martinovic, 2012). Some studies are based on
experimental research (Gaertner and Dovidio, 2000: 103ff.), others rely on
survey data, for example from the International Social Survey Programme
(ISSP). These studies mostly analyse the correlates of identity types or profiles
using cross-sectional data (Schwartz, 2005: 299). They evince considerable inter-
individual variation in ethnic and national identities. Correlates on the individual
level include education (Zimmermann et al., 2007), proficiency in both the receiv-
ing and sending countries’ languages (Esser, 2009), contact to majority and minor-
ity members (De Vroome et al., 2014; Leszczensky, 2013), citizenship status
(Ersanilli and Koopmans, 2010) and experiences of discrimination (Verkuyten
and Yildiz, 2007). Intergroup variation seems to be substantial as well, with
some groups identifying more strongly with the receiving or sending country
than others (Diehl and Schnell, 2006; Zimmermann et al., 2007). On the group
level, ethnic group size, the degree of ethnic replenishment, residential segregation
(Esser, 2004; Jime
´nez, 2008), as well as the salience and the nature of ethnic group
boundaries play an important role (Verkuyten, 2005: 159).
In this paper, we will describe and explain early changes in recently arrived
migrants’ identification with the receiving country. Doing so will provide insight
into a very dynamic phase in migrants’ identities that has so far remained a black
box in integration research. Migrating to a new country is a typical change in social
context that social identity theorists have described as a trigger for changes in
identity and their meanings (Howard, 2000: 379; Owens et al., 2010: 488): New
– alternative or overarching – ‘ethnic options’ (Waters, 1990) are opening up
as assimilated (e.g. ‘German’ or ‘American’), hyphenated (e.g. ‘Turkish-
German’) or pan ethnic (e.g. ‘European’) identities. Depending on their experiences
Diehl et al. 237
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
in the host society, e.g. discrimination due to their ethnic background, migrants
may decide to distance themselves from or to embrace these new – or their old –
identities.
We will focus on migrants’ identification with the receiving country – i.e. their
‘national’ identification (see Verkuyten and Martinovic, 2012) or their ‘identifica-
tional assimilation’ (Gordon, 1964: 71) – rather than with their country of origin
for theoretical, empirical and policy reasons. Theoretically, it can be assumed that
being exposed to a new context affects migrants’ identification with the receiving
country more than their identification with the country of origin. After all, it
seems unlikely that migrants abandon their homeland identity right away, even
though this may happen in the long run. In turn, empirical evidence for reactive
ethnic identities, although a prominent concept, is yet very scarce (Diehl and
Schnell, 2006), and if it happens at all, it will probably not happen right after
migrating to a new context but over time or in the next generation (Hansen,
1962; Portes and Rumbaut, 2001). Moreover, from a policy perspective, it seems
important that new members of a society develop some sense of belonging to
their new homeland and to understand which factors can hamper this process.
Compared with this, the question of whether they maintain their ethnic identifi-
cations and thus develop some sort of ‘dual’ identity or abandon their old ties
and identifications and become ‘assimilated’ seems less important. However, we
will take into account findings from previous studies showing that there is an
empirical correlation between these two – analytically separate – dimensions of
identification (Skrobanek, 2009; Verkuyten and Martinovic, 2012; Verkuyten and
Yildiz, 2007) by briefly presenting some preliminary findings on this issue in our
conclusion.
Our analyses are based on novel data from a two-wave survey among recently
arrived immigrants in Germany. We will assess how and why these patterns differ
inter-individually and between migrants from various origins. In particular, we will
analyse to which extent differences in migrants’ identity trajectories reflect their
ongoing cognitive (referring mostly to language acquisition) and social (referring
mostly to making friends with majority members) assimilation processes as well as
their group-specific reception contexts. Notably, differences in the nature of ethnic
boundaries will be more salient for some immigrant groups than for others. By
analysing longitudinal data collected among newly arrived migrants, our study
moves beyond existing research in several respects. To our knowledge, no study
on ethnic identity has so far focused on migrants who have moved to Europe only
recently.
In the following, we will describe in further detail the theoretical arguments that
guide our research and present existing empirical findings. These relate to changes in
migrants’ identification with the receiving country in general and to the role of their
assimilation in other spheres and their experiences and perceptions of discrimination
in particular. This section will be followed by a description of the ethnic groups
under consideration here and the corresponding reception contexts and climates they
face in Germany. Along with this, we will present our theory-driven expectations
238 Ethnicities 16(2)
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
about both groups’ early identity trajectories. Based on this we will present our data,
empirical findings, and finally sum up and discuss our results.
Theoretical background and existing findings
Theoretically, migrants’ ethnic identities are examples of ‘collective’ or ‘social
identities’, two terms that are often used interchangeably (Owens et al., 2010:
490). According to Social Identity Theory (SIT), social identity is ‘that part of a
person’s self-concept which derives from his knowledge of his membership of a
social group (or groups) together with the value and emotional significance
attached to that membership’ (Tajfel, 1981: 255). There are many ways to measure
immigrants’ identities, most importantly, ethnic labelling and self-categorisation,
the sense of belonging to a group, ethnic behaviour and in-group attitudes (Phinney
and Ong, 2007). Many empirical studies on immigrant identities focus on their
‘ethnic’ and ‘national’ identifications, the former referring to migrants’ identity
as a member of the country of origin and the latter referring to their identification
with the country of destination (Verkuyten and Martinovic, 2012).
Migrants’ identification with the reception context: theoretical arguments...
When it comes to explaining if and why migrants identify with their new receiving
country a canonical and simple answer is that this ‘just happens’ over time. Inter-
individual and intergroup variations in the pace of this process are due to migrants’
differential exposure to contexts and contacts that promote or hamper their iden-
tification with both contexts. This was the idea behind Milton M. Gordon’s famous
dictum that migrants’ identification with the receiving country is the last and final
stage of their integration process that follows more or less ‘naturally’ ‘[...] once
structural integration has occurred [...]’ (1964: 81). Migrants who have acquired
the necessary language skills and left the ethnic niches of the receiving country’s
labour market meet majority members and enter the majority society’s primary
groups (‘structural assimilation’ in Gordon’s terms). Once this step has been com-
pleted, their identification with the receiving country will automatically increase.
Gordon argues that this process might take several immigrant generations to be
completed, implying that he had a rather long-time span in mind. But the idea that
acculturation and social ties with majority members promote migrants’ identifica-
tion with the majority country can also be applied to first-generation migrants’
national identification.
1
In order to come to testable conclusions about differ-
ent groups’ identity trajectories it is necessary, although, to identify the general
mechanisms that are at work behind Gordon’s assumption (see also De Vroome
et al., 2014).
Migrants’ acculturation can be considered to be the first step in their assimilation
process. It has been argued that skills in the majority language influence migrants’
identification with the receiving context (Walters et al., 2007) because they are a
precondition for contacts with natives. Another important mechanism is that
Diehl et al. 239
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
speaking the receiving country’s language comes along with a greater exposure to
its values, norms and practices. Furthermore, language is an important means to
indicate belonging and solidarity and to demarcate identities (Miller, 2000). It has
also been argued that speaking the language of the receiving country increases
minority members’ similarity with majority members and that this enhances their
identification with the latter (Hochman and Davidov, 2014: 346).
The relationship between social assimilation and national identification has also
received some attention. One mechanism described in the sociological literature is
based on the assumption that contacts with natives signal to minority members that
ethnic boundaries are permeable and that belonging to the majority is a feasible
strategy (Leszczensky, 2013, for a similar argument on the role of majority contacts
in the naturalisation decision see Diehl and Blohm, 2003). According to SIT, per-
meability of intergroup boundaries is the main precondition for low status group
members to abandon devalued group memberships/identities and to become mem-
bers of higher status groups, which, in turn may foster positive social identities and
psychological well-being (Tajfel, 1978). Verkuyten and Martinovic thus argue that in
an ‘[...] intergroup group structure with permeable group boundaries, ethnic minor-
ity group members tend not to use strategies of ethnic identification and social
competition, but rather national identification and individual mobility’ (2012: 93).
Signalling permeability of ethnic boundaries is, however, not the only mechan-
ism that links social and identificational assimilation. Interaction with majority
members may again come along with increasing exposure to receiving society’s
norms, values and social practices and – for some groups more than for others –
with the pressure to adopt these, including a national identification (Lubbers et al.,
2007; Schulz and Leszczensky, 2015). Furthermore, contacts with natives offer
opportunities to obtain social approval for declaring or showing loyalty to the
receiving country (Esser, 2009: 360).
More recent theoretical approaches to migrants’ adaptation emphasise the role
of salient ethnic boundaries – i.e. socially relevant ethnic distinctions that matter in
a given reception context (Wimmer, 2008: 975) – in migrants’ assimilation process.
Salient boundaries come along with some degree of social closure so that minority
members’ access to resources is limited, in other words: they are discriminated
against by majority members (Wimmer, 2008: 980). In line with this argument,
proponents of classical and neo assimilation theory have conceded that discrimin-
ation can slow down the integration process (Alba, 2005; Gordon, 1964: 78). It can
affect both migrants’ motivation to identify with the receiving country and their
opportunities for doing so (Esser, 2009: 360). Feeling as an integral part of the
receiving country may not only become less attractive if it is perceived as being
exclusionary and discriminatory but will also reduce perceived opportunities to
receive social approval for visible signs of loyalty. As a consequence, ‘assimilation
[...] as a strateg[y] for individuals to ‘shift sides’ and escape a minority stigma [...]’
(Wimmer, 2008: 19) may be perceived as not being an option by individuals who
feel discriminated against. This could negatively affect their readiness to identify
with the majority.
240 Ethnicities 16(2)
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
While salient ethnic boundaries often come along with ethnic discrimination
that may in turn hamper migrants’ identification with the receiving society, another
consequence of ethnic boundaries is also noteworthy. According to Alba, salient
or, in his terminology, ‘bright’ boundaries involve social and cultural distinctions
that are ‘unambiguous, so that individuals know at all times which side of the
boundary they are’ (Alba, 2005: 22). Minority members who perceive substantial
cultural distances between majority and minority may find it difficult to identify
with the majority population unless they have abandoned old loyalties and iden-
tifications (Schulz and Leszczensky, 2015; see also Fleischmann et al., 2013: 215).
Such a ‘crossing’ of group boundaries is, however, unlikely to happen among
recent first generation migrants. Things look different for minority members who
face ‘blurry’ boundaries (Alba, 2005: 22). For these groups, the characteristics of
majority and minority group do not seem to be mutually exclusive. Thus, it should
be considerably easier for them to identify with the receiving country rather soon
after their arrival. Salient ethnic boundaries may manifest themselves in discrim-
inatory attitudes and behaviour – and migrants’ perceptions hereof – as well as in
alleged incompatibilities between the characteristics of minority and majority. Both
mechanisms can thus be expected to link bright ethnic boundaries directly to lower
levels of identification with the receiving country.
...empirical evidence and open questions
Several studies have analysed the relationship between migrants’ cognitive and
social assimilation, discrimination and perceived incompatibilities between major-
ity and minority culture on the one hand and their identification with the latter on
the other hand. Previous studies based on cross-sectional data show that migrants’
cognitive and social assimilation is associated with higher national identification
especially for first generation migrants (De Vroome et al., 2014: 21). Hochman and
Davidov (2014) show that migrants’ cognitive assimilation, i.e. their proficiency in
German, has a positive effect on their identification with Germany.
A recent longitudinal examination of the relationship between migrants’ social
assimilation and their identification with the receiving society among German born
adolescents with Turkish background revealed that social assimilation and identi-
fication are unrelated once unobserved heterogeneity and reverse causality are
taken into account (Leszczensky, 2013). The author concedes, however, that
causal effects might occur at earlier stages in life. Esser (2009) did not find an
unambiguous relationship between contacts with natives and migrants’ identifica-
tion with Germany in his longitudinal study based on data from the German Socio-
Economic Panel (GSOEP) either.
Jasinskaja-Lahti et al. (2009: 121) show in their longitudinal study that there is
in fact a negative relationship between experiences of discrimination and national
identification among FSU migrants in Finland but not between discrimination and
ethnic identification. De Vroome et al. (2014: 21) come to a similar conclusion and
argue that perceived acceptance or rejection by the majority strongly influences
Diehl et al. 241
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
immigrants’ sense of national belonging. Schulz and Leszczensky (2015) show that
salient ethnic boundaries counteract the positive effect of native friends on
migrants’ identification with the majority.
The studies reviewed so far have in common that they study the relationship
between migrants’ cognitive and social assimilation, discrimination and national
identification not at the beginning of the assimilation process but give a snapshot of
this relationship at a later – and necessarily rather arbitrary – stage of their stay in
the receiving country. Others look at this relationship among second-generation
migrants which makes a lot of sense because they were born in the receiving coun-
try and an ‘identification gap’ between this group and majority members is more
puzzling than between immigrants in a narrower sense of the word (i.e. those who
immigrated themselves) and the latter. However, these studies cannot answer the
question of whether some migrants’ lower (or higher) levels of identification with
the receiving country existed already right after or even before immigration or
evolved during their course of their stay.
Furthermore, with a few exceptions (Esser, 2009; Jasinskaja-Lahti et al., 2009;
Leszczensky, 2013), the studies mentioned so far have looked into the relationship
between assimilation, discrimination and identification based on cross-sectional
data. Even though they shed much light on the phenomenon under consideration
here, they cannot test any assumptions about causal relationships between different
dimensions of the assimilation process.
And finally, the arguments and findings presented so far refer to the general
mechanisms triggering or hampering migrants’ identification with the receiving
country and focus on migrants in general or on specific groups but they do not
study group differences systematically (for a recent study on these differences see
Schulz and Leszczensky, 2015). We argue, however, that it needs to be taken into
account that new immigrants’ adaptation process starts out from group-specific
circumstances. In this respect, an ethnic group’s degree of residential and labour
market segregation, linguistic and cultural distance, overall educational level as
well as the nature and strength of ethnic group boundaries appear to matter most.
In order to come to testable conclusions about the identity trajectories for the
groups under consideration here, we will now turn to describing the migration
history, immigrant population and the reception context for recent migrants
from Turkey and Poland to Germany.
2
Based on this, we can formulate specific
expectations that guide our empirical analysis of both groups’ early integration
trajectories.
Poles and Turks in Germany: expected results
Contemporary newcomers from Turkey or Poland enter into rather distinctive
trajectories of Germany’s post-war history of immigration. There are now about
2.8 million German inhabitants with Turkish migration background, thus consti-
tuting the second largest single immigrant group in Germany (Ethnic Germans
being the largest). The pioneer migrants were predominantly male low skill
242 Ethnicities 16(2)
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
labour migrants who came to fill the German economy’s labour demand in the
1950s and 1960s. After a recruitment stop in 1973, family members joined them and
settled permanently in the Federal Republic. Family reunion is still the most
important migration motive among Turks (Bundesamt fu
¨r Migration und
Flu
¨chtlinge, 2010: 206). There are also a considerable number of Turkish migrants
coming to Germany to pursue post-secondary education, and a few Turkish skilled
migrants have arrived under the new governmental policy of attracting high-skilled
immigrants.
In contrast, while it is true that large numbers of Polish speakers had migrated
from the former Eastern Prussian provinces to the industrial centres in the West
Germany before the first World War, today’s 640,000 or so persons with Polish
migration background have mainly arrived in the post-communist period – either
as ‘Ethnic Germans’ (Aussiedler) or as workers or students (Bundesamt fu
¨r
Migration und Flu
¨chtlinge, 2010: 37ff.). Since Poland’s accession to the
European Union in 2004, Poles have received new rights to freedom of movement
within the wider European Union, though Germany restricted labour migration
from Poland until May 2011.
Poles and Turks differ with regard to their ethnic group’s size and ethnic insti-
tutional completeness. Given their numbers, especially in some larger German cities
such as Berlin or Cologne, new coming Muslim Turks enter institutionally more
complete ethnic communities than Poles. Germany’s Turks also have been found to
have comparatively few interethnic friendships, even in the second generation and
even compared with Turks in other European cities (Crul and Schneider, 2012: 389).
In a similar vein, both groups differ with respect to the nature and strength of ethnic
boundaries they face upon arrival. Social distances on behalf of natives are particu-
larly strong for Turks (Blohm and Wasmer, 2008) who also more often experience
discrimination than non-Muslim immigrants (Hans, 2010) such as Poles. After all,
ethnic boundaries tend to be defined religiously in Europe (Foner and Alba, 2008;
Zolberg and Woon, 1999) and stereotypes about groups’ alleged unwillingness to
adapt and contribute to German society and culture are quite widespread (for an
example see Sarrazin, 2010). New arrivals with a Muslim background such as Turks
will thus soon be confronted with the vivid debate about the incompatibility
between Islam and Western culture that has clearly gained momentum during the
last decade. This is not the case for Poles in Germany. The critical public discourse
on immigration from Eastern Europe to Germany focuses mostly on Romanians
and much less so on Poles.
In sum, Poles join a rather well-integrated group of co-ethnics in Germany
whereas Turkish migrants’ assimilation proceeds slower than for other groups.
Furthermore, salient ethnic boundaries exist mostly between Turkish migrants
and the majority population in Germany but not between Polish migrants and
Germans. Starting out from the theoretical arguments outlined above and our
description of group specific reception contexts we can now formulate specific
expectations about both groups’ patterns of early changes in ethnic and national
identities (for a summary of our expected results see Figure 1).
Diehl et al. 243
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
Given the different natures of ethnic boundaries that Turkish and Polish
migrants face in Germany, we assume that Poles show a rather ‘classical’ pattern
of social and cognitive assimilation and low perceptions of discrimination and
group differences. We expect Turks, in turn, to assimilate socially and cognitively
slower than Poles and to perceive more discrimination and group differences as a
consequence of more salient ethnic boundaries than Poles.
Notwithstanding these initial differences, we further expect to find support for
the assumption from assimilation theory that identification with the receiving con-
text is generally stronger among those individuals who speak the language and
interact with natives. Based on the boundary approach, we also expect that experi-
ences of discrimination and perceptions of strong cultural differences between
majority and minority are associated with lower levels of identification with the
residence country – for both Poles and for Turks.
Data and analytical strategy
In our empirical analyses, we draw on data from a unique dataset produced in the
international survey project on Socio-cultural Integration Processes among New
Immigrants in Europe (SCIP) that was funded by the NORFACE Research
Programme on Migration (Diehl et al., 2015). The SCIP project is a two-wave-
panel study of selected migrant groups in which about 7000 recent migrants aged
between 18 and 60 were surveyed in four European destination countries –
Germany, Netherlands, United Kingdom and Ireland. Migrants with a maximum
stay of 1.5 years were interviewed soon after their arrival and as many as possible
were re-interviewed again another 1.5 years later.
3
To analyse group differences,
Poles as a rather recent immigrant group to these destinations, and Turks/
Pakistanis/Moroccans as groups representing the classical labour/colonial
Poles:
Time in Germany + Assimilaon
a + Idenficaon with Germany
- Boundaries
b -
Turks:
Time in Germany < P. Assimilaon a + Idenficaon with Germany
> P. Boundaries b -
(a)
(a) (b)
(b)
Figure 1. Relationship between time in and identification with Germany for Polish and
Turkish new immigrants: expected results.
a
Assimilation refers to cognitive and social assimilation.
b
Boundaries refer to perceived discrim-
ination and value incompatibilities between majority and minority.
244 Ethnicities 16(2)
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
migration to Western Europe, were included in the SCIP survey. These groups
contribute greatly to the share of migrant population in the four countries (for a
detailed description of the methodological setup of the project see Gresser and
Schacht, 2015). In Germany, immigrants from Turkey and Poland having
stayed in Germany up to 1.5 years were interviewed in Turkish and Polish
CAPI-interviews. Initially, a random sample was drawn from population registers
in five large cities.
Migrants’ identificational integration is measured using the ISSP identity ques-
tions (full questionnaire available at: http://www.scip-info.org): How important is
the following to your sense of who you are/ how proud are you of [...]? Answer
options included, among others [...] the country where you were born?, [...] your
current country of residence? These items are measured using a 4-point scale that
were combined into an additive index ranging from 2 to 8 (very proud/important-
not proud/important at all).
4
As mentioned in the introduction, we will concentrate
on migrants’ identification with their receiving country.
The independent variables are measured as follows: How well would you say you
understand/speak/read /write German when someone is speaking to you? (1 ¼not at
all, 4 ¼very well) (cognitive assimilation); Let us talk a little more about the people
who are important to you personally and who you feel close to that live in Germany.
Please do NOT include your parents, your husband/wife or your children, but you
CAN include other relatives. For up to four persons mentioned it was asked (among
others): Is the background of this person Polish/Turkish, German, or some other
group? As a second indicator migrants were asked: How often do you spend time
with Germans? (1 ¼never, 6 ¼daily) (social assimilation).
Levels of discrimination are operationalised by perceived group discrimination
rather than by individual experiences of discrimination since it can be assumed that
discrimination can hamper migrants’ identification if they think that members of
their group are discriminated against – even if they personally haven’t experienced
any discrimination: Some say that people from Poland/Turkey are being discrimi-
nated against in Germany. How often do you think Polish/Turkish people are discri-
minated against in Germany? (1¼never, 5¼very often). In the SCIP survey, migrants’
subjective perceptions of salient group differences were asked directly. Their
approval of the item: The values of Poles/Turks and Germans are irreconcilable/
totally different (1¼strongly disagree, 5¼strongly agree) thus serves as a second
indicator for salient ethnic boundaries.
Time spent in Germany is measured in months. Every migrant was interviewed
twice but for reasons related to the practicalities of sampling and fieldwork the time
migrants’ had already spent in the country at the time of their first interview ranged
from 1 to 15 months.
5
This enables us to analyse the relationship between migrants’
assimilation, experiences of discrimination and identification in more detail than by
just comparing wave 1 and wave 2 interviews. Independent of that, the time span
between the first and the second interview was a minimum of 15 months and varies
little between respondents.
Diehl et al. 245
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
We start out by examining if changes in migrants’ identification with Germany
differ between Poles and Turks over time. Based on this, we estimate a set of
random effect regressions to utilise the between and within variation of time
spent in Germany.
6
Doing so, we analyse if group-specific trajectories reflect
early experiences in those factors theoretically expected to trigger migrants’ iden-
tification with the receiving context, namely their social and cognitive assimilation
and their early perceptions of exclusion. These regressions are conducted separately
for the two groups to allow that both groups react differently to assimilation and
exclusion. By running both, regressions of assimilation and exclusion on time, in a
first step, and of identification on time, assimilation and exclusion in a second step
we can ‘decompose’ the total time effect into a direct (time!identification) and an
indirect (time!assimilation/exclusion and assimilation/exclusion!identification)
effect. Finally, we illustrate our findings by simulating how identity trajectories of
Turks would look like if they experienced Poles’ levels of assimilation and exclusion
or reacted to these experiences as Poles do.
Findings
In order to get a grasp of the general patterns of identificational integration for
Poles and Turks, we display Polish and Turkish migrants’ identification with the
receiving country in Figure 1 as a function of the time spent in Germany.
7
As
expected, both groups show different patterns of adaptation during their first 3
years in Germany (Figure 2).
In the beginning of their stay, Turkish migrants identify more with Germany
than Poles’ but this changes over time: Turkish migrants’ identification with the
reception country stagnates later on and eventually decreases and after about 24
months, they show slightly lower levels of identification with Germany than Poles.
The latter group’s identification with the reception context continuously increases
over time.
8
While this basic pattern does indeed provide at least some support for the
expectation that Turks’ identity trajectories deviate from the pattern pre-
dicted by assimilation theory, it is yet unclear which processes underlie the declin-
ing identification of Turks with Germany. A closer look at the group averages of
the model variables (see Table 1) yields several salient differences: only Poles show
rising levels of identification with Germany between the two waves. To provide a
full picture of both groups’ identity patterns, we also display mean values
for identification with Poland/Turkey. These reveal that while Turks identify some-
what stronger with Turkey than Poles with Poland, both groups show a
slight though non-significant increase in their identification with the country of
origin over time. This may point to a heightened salience of the homeland
identity after migration. Fewer Turks than Poles are working, a finding that is
likely to reflect the different migration histories of Poles and Turks to Germany:
Poles come mostly to Germany to work and study whereas Turks come mainly to
246 Ethnicities 16(2)
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
join their families already living in Germany. Accordingly, many immigrated as
spouses.
With respect to the indicators for migrants’ early social and cognitive assimila-
tion, results show that Poles and Turks have similar German skills in wave 1 but
that Turks improve their proficiency in German more between waves 1 and 2,
possibly related to the fact that they know more German speaking co-ethnics
when they arrive. The picture is mixed with respect to both groups’ social assimi-
lation: Poles spend more time with Germans than Turks and this gap widens over
time. However, more Turks than Poles have Germans among their best friends.
The indicators for the ethnic boundaries suggest a group specific pattern: Turks
and Poles perceive similar degrees of group discrimination at the beginning of their
stay but these perceptions increase significantly over time only for Turks, whereas
they remain stable for Poles.
9
Perceived value incompatibility is substantially
higher for Turks than for Poles in wave 1 and tends to increase over time for
Turks but remains stable – at comparatively low levels – for Poles.
We will now turn to our multivariate analyses in order to study the relationship
between migrants’ early cognitive and social assimilation, their experiences of dis-
crimination and their identity trajectories. We ran separate random effects regres-
sions on both, the indicators of assimilation and discrimination and on migrants’
identification using multiply imputed datasets.
10
Results including calculated
5.6
5.7
5.8
5.9
6
6.1
6.2
Identification with Germany
0 3 6 9 12 15 18 21 24 27 30 33 36
Time in Germany (months)
Poles Turks
Figure 2. Identification with Germany for Poles and Turks by time in Germany in months
(means).
Diehl et al. 247
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
Table 1. Descriptives and means of variables in wave 1 and wave 2 (balanced panel).
Poles (n¼590) Turks (n¼432) Total (n¼1022)
Mean Mean Min Max N
Wave 1 Wave 2 Wave 1 Wave 2
Identification Identification with Germany
(add. Index)
5.85 6.04
b
6.02 5.94 2 8 2060
Identification with country
of origin (add. Index)
6.34 6.48 6.73
a
6.80
a
2 8 2173
Time in Germany in months 8.52
a
28.38
ab
6.73 26.7
b
0 41 2198
socio-demographics Age at immigration 32.4
a
32.4
a
28.5 28.5 16 60 2222
Ethnic Turk 0 0 0.76 0.76 0 1 2168
Female 0.56
a
0.56
a
0.45 0.45 0 1 2254
Education
Primary or less 0.032 0.033 0.17
a
0.16
a
0 1 2211
Secondary 0.47
a
0.46
a
0.38 0.38 0 1 2211
Tertiary 0.50 0.51 0.45 0.46 0 1 2211
Employment status
Working 0.57
a
0.66
ab
0.21 0.39
b
0 1 2238
In education 0.19
b
0.15 0.30
ab
0.22
a
0 1 2238
Other (e.g. unemployed.
retired)
0.81
a
0.85
ab
0.70 0.78
b
0 1 2238
Cognitive and social
assimilation
German language skills
(mean Index)
2.40
a
2.68
b
2.35 2.74
b
1 4 2221
Time spent with Germans 4.96
a
5.10
a
4.26 4.31 1 6 2223
Number of German friends 0.41 0.50 0.58
a
0.67
a
0 4 2254
Discrimination/ Perceived value incompatibility 2.66 2.62 3.30
a
3.37
a
1 5 2107
value incompatibility Experiences of group discrimination 2.74 2.72 2.75 2.92
ab
1 5 2069
a
Significantly larger for Poles (Turks) than Turks(Poles) (p<0.05).
b
Significantly increase/decrease between Wave 1 and Wave 2 (p<0.05).
248 Ethnicities 16(2)
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
Table 2. Identification with Germany of Polish and Turkish new immigrants: Direct, indirect and total effects of time in Germany (unstandar-
dised effects. for full models see Appendix A1, M1-5 and Appendix A2).
Time in Germany !
Assimilation and
perceptions of discrimination
Assimilation and
perceptions of discrimination
!Identification (see A2)
Time in Germany !
Identification
Poles Turks Poles Turks Poles Turks
B SE b SE b SE b SE b
%of
total
effect b
%of
total
effect
Indirect effects... (a) (b) (a*b)
... German language
skills (see A1. M1)
0.114***(0.010) 0.178*** (0.015) 0.029 (0.074) 0.107 (0.119)0.003 3.43% 0.019 –21.15%
...Time spent with
Germans (see A1. M2)
0.074* (0.034)0.065 (0.050) 0.087**(0.032) 0.112** (0.037)0.006 6.68% 0.007 8.08%
...Number of German
friends (see A1. M3)
0.044* (0.021) 0.030 (0.029) 0.039 (0.038) 0.158** (0.057)0.002 1.78% 0.005 5.26%
...Perceived value
incompatibility (see A1. M4)
0.014 (0.024) 0.039 (0.036)0.001 (0.046) 0.132** (0.048)0.001 0.89%0.018 20.46%
...Experiences of group
discrimination (see A1. M5)
0.011 (0.022) 0.098** (0.033)0.078 (0.050)0.188***(0.051) 0.000 0.01% 0.005 5.72%
Time in Germany in months /10 (c‘)
Direct effects 0.084** 87.20%0.083+ 92.15%
(c)
Total effects
1
0.096***100% 0.090** 100%
Note:+p<0.10.*p<0.05.**p<0.01.***p<0.001; coefficients in bold indicate that absolute value of coefficient is statistically significantly (p<0.05) larger for Poles
(Turks) than for Turks (Poles).
Diehl et al. 249
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
indirect and total effects
11
are summarised in Table 2 (for full models see Appendix
A1, M1–5 and Appendix A2).
Overall, the total effect of Time in Germany confirms that over time, Turkish
migrants’ identification with Germany decreases significantly, while it increases
substantially for Poles (see last row in columns 10 and 12 in Table 2). These effects
are significantly different between both groups. Furthermore, Table 2 reveals that
over time, Turks and Poles cognitive assimilation increases, for Turks even stronger
than for Poles. However, social assimilation increases only for Poles but not for
Turks. Multivariate results confirm that both groups also show a different pattern
in terms of their experiences of discrimination: over time, Turks perceive more
group discrimination whereas this is not the case for Poles (see columns 2 and
4 in Table 2).
With respect to the impact of migrants’ early assimilation and their experiences
of discrimination on their identification, the patterns looks again similar for both
groups (see columns 6–9 in Table 2): Migrants’ early cognitive assimilation is unre-
lated (Poles) or weakly related (Turks) to their identification with Germany,
whereas social assimilation enhances both groups’ identification with the receiving
context. Discrimination diminishes their identification with Germany but this effect
is substantially larger for Turks than for Poles. This finding could be related to the
greater salience of ethno-religious boundaries for Turks than for other migrants
that turn perceptions of discrimination into something more threatening for Turks
than for Poles – possibly because they point to a larger societal problem and more
severe social exclusion. The finding that the perception of value incompatibilities
diminishes only Turks’ identification with Germany but not Poles’ sense of national
belonging points in the same direction. Obviously, these incompatibilities are more
fundamental in nature for Turks than for Poles.
The direct effects of time in Germany on identification are somewhat attenuated
when taking the indicators of cognitive and social assimilation as well as ethnic
boundaries into account (see second last row in columns 10 and 12 in Table 2).
However, the relative size of the direct effect of time in Germany on identification
with Germany remains quite large, especially for Turks (see columns 10 and 12 in
Table 2). This is partly related to the fact that the indirect effects of cognitive
assimilation and discrimination point in opposite directions and counterbalance
each other. Obviously, the variables under consideration here (and the measure-
ments used!) can only account for a small proportion of the overall trend in
Turkish and Polish migrants’ identification with Germany over time. We will
come back to this in our conclusion.
In Figure 3, a simulation is presented that shows how the decline in Turkish
migrants’ national identification (not identification itself!) with Germany would
decrease (or increase) if they experienced similar degrees of assimilation and dis-
crimination and showed similar reactions to these processes than Poles.
12
Turkish
migrants’ decline would be by about 30% smaller if their perceptions of group
discrimination and value incompatibilities were as low as Poles’ perceptions. Their
decline in identification with Germany would be about 18% smaller if they were as
250 Ethnicities 16(2)
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
resilient to discrimination and perceived value incompatibilities as Poles. Turkish
migrants’ identification with Germany would also be higher if their social assimi-
lation progressed as fast as Polish migrants’ assimilation. Since Turkish migrants’
assimilate faster cognitively, i.e. learn German quicker, and since cognitive assimi-
lation tends to enhance Turkish migrants’ identificational assimilation, the decline
in their identification with Germany would be even more pronounced if they
learned German as slowly as Poles do. To put it differently: the fact that
Turkish new migrants learn the language faster than Poles attenuates the decrease
in Turkish migrants’ identification.
Conclusion
Our analyses of Polish and Turkish migrants’ early patterns of identification with
the receiving country have revealed some interesting and significant differences
between these groups. Most importantly, only new migrants from Poland show
an increasing identification with Germany over time, whereas this is not the case
for Turks. In fact, their national identification decreases.
Theoretically, we have argued that both groups’ identificational integration has
started out from rather different conditions: As non-EU immigrants, Turks arriv-
ing in Germany join an ethnic group that lags behind other minority groups with
respect to their cognitive, structural and social assimilation. Since immigration
Figure 3. How would it affect Turkish migrants’ decline in identification with Germany if they
showed Polish migrants’ patterns of assimilation and discrimination? (Simulation based on
table 2).
Diehl et al. 251
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
from Turkey is still heavily network based, it can be expected that newcomers show
a similar pattern of comparatively slow integration into the status systems and
social networks of the majority. Moreover, Turks arriving in Germany join an
ethnic group that is in the centre of a heated societal debate on integration, that
is met by natives with comparatively high levels of social distance and that faces
discrimination more often than other groups.
Based on assimilation theory and the literature on ethnic boundaries, we
expected that both groups’ identity trajectories start out from very different con-
ditions but there is no reason to assume that the basic mechanisms leading to
migrants’ identification with the majority differ a great deal for newcomers from
Poland and Turkey: On the one hand, their assimilation in other spheres should
enhance both groups identificational integration with the receiving country, on the
other hand, discrimination and perceived value incompatibilities – reflecting bright
ethnic boundaries – should hamper it.
Overall, our analyses confirm these expectations: over time, both groups show a
clear pattern of cognitive, i.e. language, assimilation. While Poles assimilate
socially during their first months in Germany this process stagnates for Turks.
However, the most pronounced difference between the groups is that the share
of individuals who feel that their group is discriminated against increases among
Turkish immigrants. In addition, Turks comparatively strong perceptions of value
incompatibilities between Germans and Turks stagnate over time.
While both groups’ identity trajectories thus start out from rather different
conditions, they do not show fundamentally dissimilar patterns with respect to
the consequences of assimilation and discrimination. The former increases
migrants’ identificational assimilation and the latter hampers it even though the
negative impact of discrimination is stronger for Turks than for Poles.
By looking at new migrants, we have been able to study the nexus among social
assimilation, discrimination and identification more directly than previous studies.
Most importantly, we could demonstrate that Turkish migrants’ comparatively low
level of identification with Germany does not exist from the very beginning but
evolves over time and reflects rising levels of discrimination and a stagnating pro-
cess of social assimilation. However, our results show that the indirect effect of
discrimination and social assimilation is quite small. That is, even if Turks experi-
enced less discrimination and established more contacts with natives they would
not show a pattern of rising national identification as Poles do – but their decline in
identification with Germany would be less pronounced.
There are several possible answers to the question of which factors could
account for this. First of all, it is possible that our model is misspecified, i.e.
other factors than the ones under investigation here explain this finding. While
we think that no alternative theoretical approaches are at hand – the arguments of
the proponents of the Theory of Segmented Assimilation (e.g. Portes and
Rumbaut, 2001) mainly refer to migrants’ stable or even reactive identification
with the sending context – we conducted some additional analyses not presented
here in order to look into several alternative ideas.
252 Ethnicities 16(2)
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
First of all, we included a very rough indicator for migrants’ structural integra-
tion in our models (in the more current sense of migrants integration the educa-
tional system and the labour market). We have not done this in our original
analysis since structural integration is a tricky concept for first-generation
migrants. What does it mean – finding a job, finding a ‘good’ job or finding a
job with many German colleagues? We decided to go for finding a job which is an
indicator for migrants’ labour market participation that may imply greater expos-
ure to receiving society’s values, norms and practices. However, results show that
Turks lower rates of inclusion in the labour market do not explain their declining
levels of identification with Germany. This is in line with previous findings that
labour market integration influences migrants’ identification only indirectly via an
increase in migrants’ social assimilation (De Vroome et al., 2014).
We also included migrants’ religiosity (religious attendance) in our analyses.
Even though we do not see any reason to argue that being religious, especially
an attachment to Islam, needs to hamper migrants’ identification with the receiving
society directly (and not indirectly by evoking discrimination), this is a popular
argument. However, we found migrants’ religiosity to be unrelated to their identi-
fication with the receiving country. And finally, we included migrants’ identifica-
tion with the country of origin in our analyses. If both identities were incompatible
as some authors suggest (see Berry et al., 2006 for the European context), a slight
increase in newcomers’ identification with Poland and Turkey (compare Table 1)
would be reflected in a slight decrease in their identification with Germany. Unlike
several other studies (e.g. Yagmur and Van De Vijver, 2012) we found, however,
that both identity types are positively correlated for both groups of recent
migrants. In this respect, our findings confirm the argument by Jasinskaja-Lahti
et al. (2009: 108) who state that ‘it seems that among minority groups negative
attitudes towards the national out-group may be related to other factors than in-
group identification’. Last but not least, we replicated our analyses by conducting
fixed (with time measured by wave) rather than random effect models (results
available upon request) and this leads to the same conclusions.
Despite all these checks, the picture we have outlined here is preliminary and it is
quite possible that Turkish migrants’ identification with Germany gains momen-
tum later on. However, our analyses show that Turkish migrants’ early patterns of
identification with the receiving country are influenced by their early experiences in
Germany. Since these are quite different from the experiences of Poles, both
groups’ identificational integration trajectories become more dissimilar over the
time period considered here. Even though the differences are not large, our analyses
have been able to link these initial differences in migrants’ identification trajectories
to those factors that are described in the literature as hampering the process of
adopting new identities, namely perceiving discrimination and value incompatibil-
ities between the home- and the receiving culture and staying apart from the social
cliques and networks of majority members. One could argue that migrants’ iden-
tifications can be considered to belong to the realm of private choices that may
have little relevance for their eventual integration in the status systems of the
Diehl et al. 253
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
receiving country. Things look different, of course, when these choices reflect feel-
ings of exclusion and salient ethnic boundaries rather than what Mary Waters once
called ‘ethnic options’.
Acknowledgements
The authors thank Matthias Koenig, Lars Leszczensky and Mieke Maliepaard for helpful
comments on earlier drafts of this manuscript.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, author-
ship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication
of this article.
Notes
1. We argue that for new migrants, their labour market integration is less important and
difficult to capture empirically, for more details see the discussion in the conclusion.
2. This following paragraph describing recent migration patterns of Poles and Turks in
Germany is partly taken from Diehl and Koenig (2013) and their analyses of new
migrants’ religious adaptation that is also based on SCIP data.
3. Panel mortality was high due to the high mobility of this group not only within Germany
but also between Germany and the sending countries. Even though large efforts were
undertaken to re-interview migrants after 1.5 years only about 45% could be re-inter-
viewed. Lasinskaja-Lahti et al. (2009) report similar rates; in the US New Immigrant
Survey panel mortality was equally high (according to an email exchange with the NIS
project director in March 2014).
4. Cronbach’s afor identification with the country of origin is 0.72 for both samples and for
identification with Germany is 0.74 (Poles) and 0.71 (Turks).
5. During fieldwork, addresses were stepwise assigned to interviewers and new addresses
were only issued when the target persons that were issued first were either interviewed
identified as ‘‘not-at-homes’’ or refusals or their addresses were found to be non-existent.
There is thus little reason to assume that migrants interviewed earlier differed systemat-
ically from those interviewed later on (see Gresser and Schacht, 2015).
6. Since there is little variation in the time span between the two interviews, fixed effect
regressions only allow us to estimate the difference in national identification between the
two waves and not as a function of time in the residence country. The estimated effects
for social and cognitive assimilation and discrimination and value compatibility are con-
sistent in the random effect model as are the group differences between the effects.
7. The graph depicts the cross-sectional relationship between time in Germany and national
identification based on the pooled data of respondents partaking at both waves. The plots
have been somewhat smoothed using locally weighted regressions.
8. This holds regardless of whether the respondent has spent time in Germany before
migration.
254 Ethnicities 16(2)
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
9. Results not displayed here show that personal experiences of discrimination are even
lower for Turks than for Poles in wave 1 (with 29 versus 36% having felt discriminated
against) but they increase significantly for Turks but decrease for Poles over time (to 32
versus 41% resp. in wave 2).
10. Missing values due to item non-response were multiply imputed using chained equa-
tions. Results refer to the analysis of five imputed datasets using Rubin’s (1987) com-
bination rules. In comparison to other missing data techniques, multiple imputation is
more efficient, reduces potential bias in the estimation of coefficients (if missing values
are not completely random but to some degree correlated with the vector of observed
variables), and estimates standard errors that reflect the uncertainty of the missing
information correctly (Allison, 2002).
11. Total effects are calculated as sum of the direct and indirect effects.
12. By recalculating the effects for Turks using the ‘Polish’ coefficients of the regressions of
the relevant mediation variables on time (‘experience’) or the ‘Polish’ coefficients of the
regressions of identification on the mediator variables of interest (‘sensitivity’). The
predicted change of the identification of the Turks between t0 und t1 (It1(T)-It0(T))
is the product of the coefficients for the mediation variables and the change in the
mediation variables over time: It1(T)It0(T)¼B(T)(M t1(T) Mt0(T)). The same
holds ceteris paribus for the Poles: It1(P) It0(P) ¼B(P)(M t1(P) Mt0(P)). The simu-
lated change for Turks ‘with the experiences of Poles’ is then: B(T)(M t1(P) Mt0(P)),
the simulated change for Turks ‘with the sensitivities of Poles’ is then: B(P)(M t1(T)
Mt0(T)), expressed as a percentage of the observed change in identification of the
Turks.
References
Alba R (2005) Bright vs. blurred boundaries: Second-generation assimilation and exclusion
in France, Germany, and the United States. Ethnic and Racial Studies 28(1): 20–49.
Allison PD (2002) Missing Data. Thousand Oaks California: Sage.
Berry JW, Phinney JS, Sam DL, et al. (2006) Immigrant youth: Acculturation, identity, and
adaptation. Applied Psychology 55(3): 303–332.
Blohm M and Wasmer M (2008) Einstellungen und Kontakte zu Ausla
¨ndern [Attitudes and
contacts with foreigners]. In: Noll H, Habich R (eds) Datenreport 2008: ein Sozialbericht
fu
¨r die Bundesrepublik Deutschland. Bonn: Bundeszentrale fu
¨r politische Bildung,
pp.206–214.
Bundesamt fu
¨r Migration und Flu
¨chtlinge (2010) Migrationsbericht 2008.Nu
¨rnberg:
Bundesministerium des Innern Referat O
¨ffentlichkeitsarbeit.
Crul M and Schneider J (2012) Conclusions and implications: The integration context mat-
ters. In: Crul M, Schneider J, Lelie F (eds) The European Second Generation Compared:
Does the Integration Context Matter? Amsterdam: Amsterdam University Press,
pp.375–404.
De Vroome T, Verkuyten M and Martinovic B (2014) Host national identification of immi-
grants in the Netherlands. International Migration Review 48(1): 76–102.
Diehl C and Blohm M (2003) Rights or identity? Naturalization processes among ‘‘Labor
migrants’’ in Germany. International Migration Review 37(1): 133–162.
Diehl C, Gijsberts M, Gu
¨veli A, et al. (2015) Socio-Cultural Integration Processes among
New Immigrants in Europe (SCIP) – Data file for download. GESIS Data Archive,
Cologne.
Diehl et al. 255
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
Diehl C and Koenig M (2013) God can wait – New migrants in Germany between early
adaptation and religious reorganization. International Migration Review 51(3): 8–22.
Diehl C and Schnell R (2006) Reactive ethnicity or assimilation? Statements, arguments, and
first empirical evidence for labor migrants in Germany. International Migration Review
40(4): 786–816.
Ersanilli E and Koopmans R (2010) Rewarding integration? Citizenship regulations and the
socio-cultural integration of immigrants in the Netherlands, France and Germany.
Journal of Ethnic and Migration Studies 36(5): 773–791.
Esser H (2004) Does the ‘‘New’’ immigration require a ‘‘New’’ theory of intergenerational
integration? International Migration Review 38(3): 1126–1159.
Esser H (2009) Pluralisierung oder assimilation? Effekte der multiplen inklusion auf die
integration von migranten. Zeitschrift fu
¨r Soziologie 38(5): 358–378.
Fischer-Neumann M (2014) Immigrants ethnic identification and political involvement in
the face of discrimination: A longitudinal study of the German case. Journal of Ethnic and
Migration Studies 40(3): 339–362.
Fleischmann F, Phalet K and Swyngedouw M (2013) Dual identity under threat: When and
how do Turkish and Moroccan minorities engage in politics? Zeitschrift fu
¨r Psychologie
221(4): 214–222.
Foner N and Alba R (2008) Immigrant religion in the US and Western Europe: Bridge or
barrier to inclusion? International Migration Review 42(2): 360–392.
Gaertner SL and Dovidio JF (2000) Reducing Intergroup Bias: The Common Ingroup Identity
Model. Philadelphia, PA: Taylor & Francis.
Gordon MM (1964) Assimilation in American Life. New York: Oxford University Press.
Gresser A and Schacht D (2015) SCIP Survey-Methodological Report. Konstanz. Available
at: www.scip-info.org (accessed 2 November 2015).
Hans S (2010) Assimilation oder Segregation? Anpassungsprozesse von Einwanderern in
Deutschland [Assimilation or segregation? Processes of adaptation of migrants in
Germany]. Wiesbaden: VS Verlag fu
¨r Sozialwissenschaften.
Hansen ML (1962) The third generation in America. Commentary 14: 492–500.
Hochman O and Davidov E (2014) Relations between second-language proficiency and
national identification: The case of immigrants in Germany. European Sociological
Review 30(3): 344–359.
Howard JA (2000) Social psychology of identities. Annual Review of Sociology 26:
367–393.
Jasinskaja-Lahti I, Liebkind K and Solheim E (2009) To identify or not to identify? National
disidentification as an alternative reaction to perceived ethnic discrimination. Applied
Psychology 58(1): 105–128.
Jime
´nez TR (2008) Mexican immigrant replenishment and the continuing significance of
ethnicity and race. American Journal of Sociology 113(6): 1527–1567.
Joppke C (2007) Beyond national models: Civic integration policies for immigrants in
Western Europe. West European Politics 30(1): 1–22.
Leszczensky L (2013) Do national identification and interethnic friendships affect one
another? A longitudinal test with adolescents of Turkish origin in Germany. Social
Science Research 42(3): 775–788.
Lubbers MJ, Molina JL and McCarty C (2007) Personal networks and ethnic identifica-
tions: The case of migrants in Spain. International Sociology 22(6): 721–741.
Miller JM (2000) Language use, identity, and social interaction: Migrant students in
Australia. Research on Language and Social Interaction 33(1): 69–100.
256 Ethnicities 16(2)
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
Owens TJ, Robinson DT and Smith-Lovin L (2010) Three faces of identity. Sociology 36(1):
477.
Phinney JS and Ong AD (2007) Conceptualization and measurement of ethnic identity:
Current status and future directions. Journal of Counseling Psychology 54(3): 271–281.
Portes A and Rumbaut RG (2001) Legacies: The Story of the Immigrant Second Generation.
Berkeley, CA: University of California Press.
Rubin DB (1987) Multiple Imputation for Nonresponse in Surveys. New York: Wiley.
Sarrazin T (2010) Deutschland schafft sich ab: Wie wir unser Land aufs Spiel setzen.
Deutsche: Verlags-Anstalt.
Schulz B and Leszczensky L (2015) Native friends and host country identification among
adolescent immigrants in Germany: The role of ethnic boundaries. International
Migration Review. Epub ahead of print n/a. DOI: 10.1111/imre.12163.
Schwartz SJ (2005) A new identity for identity research: Recommendations for expanding
and refocusing the identity literature. Journal of Adolescent Research 20(3): 293–308.
Skrobanek J (2009) Perceived discrimination, ethnic identity and the (re-) ethnicisation of
youth with a Turkish ethnic background in Germany. Journal of Ethnic and Migration
Studies 35(4): 535–554.
Tajfel H (1978) Differentiation Between Social Groups: Studies in the Social Psychology of
Intergroup Relations. London: Academic Press.
Tajfel H (1981) Human Groups and Social Categories: Studies in Social Psychology.
Cambridge [Cambridgeshire], New York: Cambridge University Press.
Verkuyten M (2005) The Social Psychology of Ethnic Identity. Hove: Psychology Press.
Verkuyten M and Martinovic B (2012) Immigrants national identification: Meanings, deter-
minants, and consequences. Social Issues and Policy Review 6(1): 82–112.
Verkuyten M and Yildiz AA (2007) National (dis) identification and ethnic and religious
identity: A study among Turkish–Dutch Muslims. Personality and Social Psychology
Bulletin 33(10): 1448–1462.
Walters D, Phythian K and Anisef A (2007) The acculturation of Canadian immigrants:
Determinants of ethnic identification with the host society. Canadian Review of Sociology
and Anthropology 44(1): 37–64.
Waters MC (1990) Ethnic Options: Choosing Identities in America. CA: Berkeley.
Wimmer A (2008) The making and unmaking of ethnic boundaries: A multilevel process
theory. American Journal of Sociology 113(4): 970–1022.
Yagmur K and Van De Vijver FJR (2012) Acculturation and language orientations of
Turkish immigrants in Australia, France, Germany, and the Netherlands. Journal of
Cross-Cultural Psychology 43(7): 1110–1130.
Zimmermann L, Zimmermann KF and Constant A (2007) Ethnic self-identification of first-
generation immigrants. International Migration Review 41(3): 769–781.
Zolberg AR and Woon LL (1999) Why Islam is like Spanish: Cultural incorporation in
Europe and the United States. Politics & Society 27(1): 5–38.
Diehl et al. 257
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
Appendix
Table A1. Turkish and Polish assimilation and perceptions of discrimination and value incompatibilities (random effect panel regression).
German language
skills
Time spent
with Germans
Number of
German Friends
Perceived value
incompatibility
Experiences of
group discrimination
Poles Turks Poles Turks Poles Turks Poles Turks Poles Turks
Time in Germany
in months/10
0.114*** 0.178*** 0.074* 0.065 0.044* 0.030 0.014 0.039 0.011 0.098**
(0.010) (0.015) (0.034) (0.050) (0.021) (0.029) (0.024) (0.036) (0.022) (0.033)
Ethnic Turk – 0.092+ – 0.303 – 0.057 – 0.051 – 0.036
– (0.053) – (0.192) – (0.095) – (0.109) – (0.114)
Female 0.374*** 0.040 0.030 0.615*** 0.245*** 0.146+ 0.054 0.002 0.097 0.232*
(0.053) (0.054) (0.095) (0.158) (0.073) (0.088) (0.070) (0.092) (0.065) (0.103)
Age at immigration/10 0.135*** 0.149*** 0.041 0.136 0.035 0.032 0.014 0.060 0.052+ 0.121+
(0.024) (0.042) (0.047) (0.116) (0.034) (0.057) (0.034) (0.067) (0.031) (0.073)
Education (ref.: secondary)
Primary or less 0.077 0.192** 0.426** 0.246 0.064 0.008 0.120 0.117 0.129 0.201
(0.121) (0.072) (0.130) (0.234) (0.173) (0.112) (0.204) (0.125) (0.148) (0.141)
Tertiary 0.155*** 0.003 0.082 0.330* 0.195** 0.198* 0.141* 0.220* 0.049 0.203*
(0.044) (0.051) (0.092) (0.154) (0.062) (0.083) (0.063) (0.093) (0.061) (0.096)
Employment status
(ref.: working)
In education 0.326*** 0.066 0.000 0.009 0.082 0.022 0.216* 0.110 0.211** 0.232+
(0.052) (0.057) (0.114) (0.185) (0.090) (0.107) (0.094) (0.113) (0.075) (0.120)
(continued)
258 Ethnicities 16(2)
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
Appendix
Table A1. Continued
German language
skills
Time spent
with Germans
Number of
German Friends
Perceived value
incompatibility
Experiences of
group discrimination
Poles Turks Poles Turks Poles Turks Poles Turks Poles Turks
Other (e.g.
unemployed.
retired)
0.038 0.066 0.958*** 0.632*** 0.128 0.017 0.059 0.021 0.086 0.028
(0.048) (0.052) (0.138) (0.172) (0.089) (0.091) (0.082) (0.106) (0.073) (0.109)
Constant 2.473*** 2.640*** 5.300*** 4.961*** 0.272* 0.484* 2.734*** 3.541*** 2.872*** 2.047***
(0.094) (0.140) (0.195) (0.407) (0.132) (0.213) (0.137) (0.245) (0.131) (0.252)
Sigma (e) 0.304 0.380 1.142 1.407 0.687 0.792 0.817 1.012 0.725 0.880
Sigma (u) 0.560 0.410 0.722 1.047 0.616 0.571 0.442 0.460 0.474 0.713
N 1041 748 1041 748 1041 748 1041 748 1041 748
n 590 432 590 432 590 432 590 432 590 432
Source: SCIP 2010–2013. Note: Standard errors in parentheses; unstandardised effects; + p<0.10.*p<0.05.**p<0.01.***p<0.001.
Diehl et al. 259
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
Table A2. Panel regression models: Turkish and Polish migrants’ identification with Germany.
Random effects
Poles Turks
Time in Germany in months/10 0.084** (0.030) 0.083+ (0.049)
Ethnic Turk – – 0.210 (0.140)
Female 0.241* (0.100) 0.101 (0.134)
Age at immigration/10 0.176*** (0.044) 0.031 (0.094)
Education (ref.: secondary)
Primary or less 0.250 (0.170) 0.415* (0.173)
Tertiary 0.269** (0.090) 0.381* (0.151)
Employment status (ref.: working)
In education 0.004 (0.120) 0.098 (0.169)
Other 0.121 (0.110) 0.207 (0.143)
German language skills 0.029 (0.074) 0.107 (0.119)
Time spent with Germans 0.087** (0.032) 0.112** (0.037)
Number of German Friends 0.039 (0.038) 0.158** (0.057)
Perceived value incompatibility 0.001 (0.046) 0.132** (0.048)
Experiences of group discrimination 0.078 (0.050) 0.188*** (0.051)
Constant 4.833*** (0.358) 6.268*** (0.518)
R
2
overall 0.063 0.122
R
2
within 0.029 0.069
R
2
between 0.082 0.131
Sigma (e) 0.874 1.146
Sigma (u) 0.892 0.888
N977 709
n576 427
Source: SCIP 2010–2013. Note: Standard errors in parentheses; unstandardised effects. +p<0.10,*p<0.05,
**p<0.01,***p<0.001.
260 Ethnicities 16(2)
at Universitaetsbibliothek Bamberg on April 4, 2016etn.sagepub.comDownloaded from
