Volume 7, No. 3, Art. 2 – May 2006
Comparative Research on Highly Skilled Migrants. Can Qualitative Interviews Be Used in Order to Reconstruct a Class Position?
Abstract: BOURDIEU's theoretical concept of "classes on paper" is based on the nation-state frame. In order to include the social position of migrants, who live and work in more than one nation-state, a concept of social and structural transnationalisation is required. Using the example of highly skilled migrants, the article advances such a concept and shows, using qualitative data and "grounded theory", how and where transnational class formation takes place. The project diverges from the dominant perspective of migration research in that it does not attempt to reconstruct cultural specifics but strategically selects a sample of highly skilled migrants as a "qualitative experiment". The approach is ambitious—in transnational social spaces neither factual information nor indicators of habitus can be placed in one single frame of reference—and includes diverse kinds of data and analytical strategies to comprehend and reconstruct transnational class positions. The paper examines the notion of transnational class formation and introduces the author's empirical research before examining, in the final part of the paper, theoretical insights around (trans-) national class formation. The results show that despite the differences in national origin of the highly skilled migrants, they operate within global labour markets and inhabit similar economic and social spaces within the city. However, the paper also argues that different types of highly skilled migrants in different types of political-economic contexts, whilst inhabiting similar economic positions and similar social space, move along different and unequal paths. This divergence, we suggest, can be traced to broader structural processes of global inequality.
Key words: class, highly skilled migration, transnationalisation, documentary method, BOURDIEU, transnational middle class, state, social inequality
Table of Contents
2. Concepts of Class Transnationalised
3. Highly Skilled Migrants as a Qualitative Experiment about the Hypothesis of Transnational Class Formation
4. Strategies for Data Analysis
5. Assessing the Structuring Force of the Nation-State System
Qualitative research on class tends to focus on the social aspects of class formation. A widespread and rewarding approach is the ethnographic (WILLIS, 1978) or comparative study of milieus (SCHITTENHELM, 2005). The reproduction of symbolic boundaries has been studied by Michèle LAMONT for the middle and the working classes in the United States and France (LAMONT, 1992, 2000). BERTAUX and THOMPSON argue that the family is an important agent and mediator of social mobility and therefore find biographical case studies very useful (BERTAUX & THOMPSON, 1997). Many important contributions to class analysis have built upon an undoubted strength of qualitative research methods; by examining actors, their collective practice and their interaction, they can understand the often paradoxical ways in which class is produced and reproduced. Qualitative research in this tradition adds an important dimension to the analysis of social structure, even if it does not claim to analyse social structures per se. 
Recently, theoretical doubt has been cast on the clear cut connection between social group formation and social structure (BECK, 1986; BERGER & HRADIL, 1990). Forms of political mobilisation which are only very distantly related to class conflict (EDER, 1993) may supersede the reproduction of class. As a result social group formation may cease to mirror class structure. The traditional focus of qualitative research methods is diverging from other indicators of class, such as the unequal distribution of resources. Therefore an analysis of social group formation must be combined with an analysis of social structure in order to offer a comprehensive analysis of class. 
Pierre BOURDIEU offers this. His empirical research is focused on the social and cultural aspects of class reproduction. Nevertheless, his theory has stressed the importance of resource distribution. BOURDIEU proposes that sociology should apply the concept of class to groups of people who share a similar capital portfolio. In his understanding of class, class is a theoretical concept, not an empirically observed social group.
"So the main error, the theoreticist error that you find in Marx, seems to consist in treating classes on paper as real classes, in concluding from the objective homogeneity of conditions, of conditionings, and thus of dispositions, which all come from the identity of position in the social space, that the people involved exist as a unified group, as a class" (BOURDIEU, 1990, p.129). 
According to BOURDIEU, sharing a similar position in social space does not lead to the formation of a social class. It will, however, result in "classes on paper" developing a common habitus, i.e. an internalised practical and mental perspective on the world, which responds to the options which people with that particular capital portfolio have in life. 
Empirically BOURDIEU uses a variety of methods in order to analyse class. In "Distinction" (BOURDIEU, 1984) he combines theoretical considerations and ethnographic observations with quantitative studies about taste. His later work "The Weight of the World" uses qualitative interviews (BOURDIEU, 1999), but distinguishes the plurality of individual perspectives from the generalising viewpoint which the sociology of class must take (BARLÖSIUS, 1999). 
Ralf BOHNSACK has proposed the "documentary method", by which similarities in habitus can be reconstructed in a systematic manner. He proposes that group interaction of persons with a similar position in social space can be seen as the actualisation of a similar life experience and orientation structure (BOHNSACK, 2004, p.216). An important feature shared by the documentary method and grounded theory approaches are the constant comparisons (NOHL, 2005; BOHNSACK & NOHL, 2003). The documentary method argues that different class positions should result in recognisable differences between the orientation structures of groups, which can be reconstructed from constant comparison. At the same time the limitations and chances common to one position in social space should also result in a shared habitus, i.e. in "habitual convergence". 
This article takes BOURDIEU's concept of class and its application in the documentary method1) as a point of departure for further consideration. Despite the many merits of BOURDIEU's approach to class, it still has one particular shortcoming. BOURDIEU is a methodological nationalist, i.e. he limits the perspective of sociology to an analysis of national societies (WIMMER & SCHILLER, 2002; BECK 2002). This article argues that concepts of class must be changed so as to include processes of transnationalisation (Part 2). In particular, the class position of migrants who live, work, earn and spend in more than one nation-state can only be understood by transcending methodological nationalism. 
This is an area of class research to which qualitative methods can significantly contribute. While quantitative research methods remain important for analysing the distribution of resources across national populations, they are deficient, in that they always need to determine first, what "representative" research is supposed to represent. More often than not, the presupposed frame is the nation-state. Qualitative research methods, on the other hand, have been very helpful for theory development and for an empirically based reflection of prevailing assumptions (GLASER & STRAUSS, 1967). The main parts of the article shows how a qualitative study about the class position of highly skilled migrants can further the development of a "grounded theory" of transnational class formation. The selection of the sample can be seen as a "qualitative experiment" (Part 3). By including diverse kinds of data into the analysis a comprehensive approach to class can be attempted (Part 4). This allows for the development of theoretical insights around transnational class relations, insights that are advanced in the final parts of the paper. 
2. Concepts of Class Transnationalised
When the concept of class was developed by MARX, it signified antagonistic positions in the system of production. As the economic system was international even then, the concept of class was too. To some extent the theoretical deduction of class from the economic system was contradicted by MARX's desire to see classes develop into political agents. MARX hoped that it would be possible to unite the "working class" so that workers would become a social group and political actor. In a political system of nation-states, this entailed classes becoming national. 
The ambiguous spatial relations of class have been expressed most clearly by WALLERSTEIN:
"Within a world-economy, the state structures function as ways for particular groups to affect and distort the functioning of the market. (…) Furthermore, this explains the ambiguity of class as a concept, since class refers to the economy which is worldwide, but class consciousness is a political, hence primarily national, phenomenon" (WALLERSTEIN, 1979, p.61). 
In the meantime not only has the economy become transnational but to some extent politics has as well (KEOHANE & NYE, 1973). A transnationalisation of the class structure has become more likely as a result of globalisation processes (HELD, GREW, GOLDBLATT & PERRATON, 1999). 
It is, however, undecided as to how a transnationalisation of class structures should be attempted. Applying a BOURDIEUIAN concept of class, Michael HARTMANN has focused his empirical research on the social aspects of class formation. With respect to a possible transnationalisation of class structure, he has looked at the national origin, international experience and educational history of top managers in the 100 largest British, French, German and U.S. corporations—including transnational corporations (HARTMANN, 2000, 2002). His research shows that the top echelons of corporations have shown little tendency to move across national borders or to adopt transnational lifestyles.2) 
Following a Marxist concept of class, Leslie SKLAIR is less interested in the transnationalisation of the life course. He emphasises a shared (economic) interest in the promotion of globalisation, which is supported by a common and transnational "culture-ideology of consumerism" (SKLAIR, 2002, p.6). With that definition of transnational class formation, SKLAIR can show that the elite of corporate executives, whom he interviewed in more than 80 major transnational corporations, can be seen as an emergent transnational capitalist class. 
The discussion about transnational class formation in this paper aims to build upon the above and engage with the undoubted ambiguities that now surround the concept of class.3) HARTMANN is looking for the formation of a transnational milieu sharing a common life-world. SKLAIR, on the other hand, is content with a shared economic position (supported by a common ideology). For the definition of transnational class formation we suggest an approach similar to BOURDIEU: de-emphasising the social aspects of class. It is of course relevant and interesting to look at group formation and to ask who is pursuing transnational lifestyles and careers. 
The core of any emergent transnational classes will probably be constituted by migrants, who live and work, earn and spend in more than one nation-state (WEISS, 2005). Most likely it will be easier to find the kind of transnational class, which HARTMANN is expecting, in the upper-middle classes and not the top corporate elites (XIANG, 2002). However, social group formation and physical mobility should not be treated as the only issues at stake. It is possible that a person remains sedentary, but is embedded in globalise labour markets (IREDALE, 2001) and epistemic communities (COE & BUNNELL, 2003). In this case he/she becomes structurally transnationalised and it is likely that his/her class position can be better understood in a transnational context than in the framework of the nation-state. 
Apart from considering social and structural transnationalisation it is also important to distinguish transnationalisation from "non-national" phenomena. Transnational class formation must not mean that the nation-state is becoming empirically irrelevant for class formation. As the value of resources depends on the geographical, social, and political context in which they are grounded (THERBORN, 2001), the relation of people (or their resources) to the nation-state system will continue to be an important dimension structuring class position besides the distribution of capital. This is especially true for migrants, but also for sedentary persons who are structurally transnationalised. It is exactly for this reason that a sociology which treats the state as an epistemologically "neutral" frame of research fails to empirically grasp the state as a structuring force, which must be included in the perspective of class analysis. 
Instead of forcing migrants and structurally transnationalised persons into the frame of the nation-state,4) we must determine where their resources (e.g. qualifications) are acknowledged. If their resources are transnationally valid and if they can gain physical, social and political access to most parts of the world, their transnational versatility will be an important structural aspect of their class position (WEISS, 2005). If their resources are devalued by a change of nation-state, or if they fail to reach a location of their liking, the nation-state system becomes an important aspect of their class position. It is necessary to transcend the nation-state frame epistemologically, so that it becomes possible to determine empirically whether specific persons can be seen as a transnational class. 
3. Highly Skilled Migrants as a Qualitative Experiment about the Hypothesis of Transnational Class Formation
The external validity of qualitative research is not enhanced by the number of persons involved but by the variety of the sample. For developing a "grounded theory" (GLASER & STRAUSS, 1967; STRÜBING, 2002) from empirical research, "theoretical sampling" is a necessary first step (see Table 1). Within this framework, highly skilled migrants should be seen as a prototype social group who are socially and structurally transnational, having overcome personal ties to a specific nation-state and political barriers to migration. They form a particularly educated tranche of the middle classes, many possessing transnationally valid forms of cultural capital, such as IT-experience, medical expertise or esteemed MBAs. Such qualifications and experience are effectively upper-middle class "passports" within global labour markets (IREDALE, 2001). Highly skilled migrants of this type change countries according to both external demand and internal labour market forces, usually within particular transnational corporations. Nation-states have tended to reduce barriers to free movement for these sought-after migrants and, in contrast to other educated migrants,5) they experience little depreciation of their cultural capital when they cross national borders. Living as migrants and being highly skilled should therefore maximise social and structural transnationalisation. 
By choosing highly skilled migrants6) to research the transnationalisation of class, we have purposefully focused on an "extreme" group, which is more likely than others to experience social and structural transnationalisation. At the same time, the sampling has tried to maximise the influence of the nation-state system by selecting only highly skilled migrants who passed a threshold of inequality (VOBRUBA, 1999; BOMMES, 2000) in the course of their migration. More specifically the project compares 11 IT-specialists who have migrated individually from developing and transformation societies to Germany with seven expatriates who have been sent to developing and transformation societies by German employers.7) These groups are rarely considered together, because they are separated not only by the direction of their migration, but by other differences as well. Experimental research of the kind proposed here may not be able to grasp every aspect of class formation/ position. Maximising contrasts between the groups should, however, highlight what they have in common despite of their differences. The social and cultural background of the sample is relatively diverse: the IT-specialists came from India (3), Algeria, Bulgaria, Brazil, China, Congo, Czech Republic, Tunisia, and Ukraine. The expatriates also lived in many countries in the course of their career.
Table 1: Maximising contrasts in the sample 
Qualitative research has an advantage over quantitative enquiry in that it is more amenable to cases that contradict prior theory building. The qualitative experiment (KLEINING, 1991) is different from the laboratory experiment in that it does not seek to change one independent variable with the goal of finding "probable" results from this change. Instead, cases are chosen so that they ensure a maximum of variation with respect to many "variables". Following the logic of "qualitative experimenting", the sample has been selected to maximise the chance of falsifying the hypothesis of transnational class formation by increasing the likelihood of "black swans". Looking for contradictions to a hypothesis the logical deductions from qualitative experiments can reach a level of generalisation which exceeds the results which probabilistic reasoning can gain from similarly small samples. In this research project, the likelihood that the structuring force of the state will appear as a prominent influence has been maximised by the fact that migrants: had crossed a threshold of inequality; had come from diverse nation-states and cultures; and, were migrating in both directions across this national gradient. Should the class position of the highly skilled migrants prove to be similar, or put more succinctly, should the research be able to explain "black swans" by something other than influence of the nation-state, we would have a very strong argument for an emergent transnational class(es). 
4. Strategies for Data Analysis
In our research on highly skilled migrants different types of data and analytical arguments were used to ensure a comprehensive class analysis. Our analysis combined the social reproduction of class (i.e. habitus, distinctions, etc.) with an examination of resource distribution. Even so, the front stage of social group formation cannot always be connected convincingly to the theoretical hypotheses surrounding the back stage structures (EDER, 2001). The general difficulty of grasping the structuring forces of class positions is further complicated by a transnational research perspective. This means that the frame of the empirical research is not predetermined but must be explored and developed during the course of the research. Ultimately, methods must be found for a class analysis that can do without the nation-state frame but nevertheless grasp the structures behind resource distribution and the social reproduction of class. 
The core of the empirical data consists of 18 interviews, between one to four hours long, with the kind of highly skilled migrants described above. The interviews were semi-structured but did encourage the interviewees to develop their own narratives and arguments (HOPF, 1978). All interviews covered the following topics: qualification and position in the labour market; educational and migration history; social networks; lifestyle preferences; legal and financial situation; experiences with differential treatment; and, hopes for the future. The interviews contain explicit and comparable information documenting interviewees' social position, as well as individual narrations from which habitus and practices of distinction can be reconstructed (BOURDIEU, 1984, 1999). The interviews were supplemented by a short questionnaire containing questions around socio-economic status and other factual information. 
It is neither necessary nor possible here to expand upon the contents and results of the research. Instead we will describe the different types of data and analytical arguments on a general level, before selecting examples in order to discuss migrants' economic capital and the habitus that surrounds spending money.8) 
Looking at the general level, the grounded theory paradigm does not limit the researcher to one set of data (STRÜBING, 2004, p.18). In fact all of the above mentioned interview and survey topics are relevant for a comprehensive analysis of class. The following quotation is particularly useful in this respect as it contains the major types of information which can be used to reconstruct and analyse class position:
"Thus we are living in a more luxurious manner than ever before. And for a kind of money that would not allow us to do so here. Ok, that is about 800 Dollars (...) per month. (...) for that we would not get a house here. For rent. We are very much aware of that <C. laughs shortly, interviewer does, too> neither would we get a tropical garden in addition" (translation by AW).9) 
These are the words of Christine Berg-Grande10) an expatriate working in development for a large governmental organisation. In the quotation she provides factual information on the rent she is paying. A self-assessment comparing her economic situation before and after the migration shows that she has moved to a more luxurious position in social space. By stressing that she is very much aware of her privileged position, she is taking a moral stance habitually distinguishing herself and her development work from the more economically oriented who take uninhibited pride in the spoils of expatriation. 
Factual information, self-assessment and habitus cannot be treated as separate entities. Self-assessments in particular combine the other two types of data. It does make sense, however, to analytically distinguish between factual information about resource distribution on the one hand, and habitus on the other because these types of data relate to different methodological arguments. 
For example, most of the IT-professionals felt that they earned about as much money as native Germans would in a similar position.11) They are quite certain about this self-assessment and it can be supported by a comparison of their income with the income of IT-professionals in the European Social Survey (N=487). Both groups vary but with a similar range around the same median. At the same time, the IT-professionals find it very hard to state conclusively whether or not they earn more in their country of origin or more in Germany.12) In absolute numbers they make more in Germany, but the cost of living also is higher. They are more likely to be able to afford real-estate in their country of origin, but then the long-term value of their investment is less certain. And so the complexity goes on. Thus, for an analysis of factual information about resource distribution it is important to keep in mind that the value of the reported resources depends upon the context in which they are used.13) If the framework in which the value of resources is assessed is a predetermined one, comparison is easy. If, as in transnational social space, the same amount of currency can purchase different amounts of goods depending on where it is spent, the comparison is much more complex.14) For migrants who live, work, earn and spend in different countries, several possible frames of analysis may apply. 
An analysis of habitus, on the other hand, does not depend on standardised measurements and a prior frame of reference. Nevertheless, a reconstruction of habitus may prove to be difficult in an intercultural context. The documentary method contends that empirical cases should not be interpreted from the perspective of the researcher, but only by comparison with each other (BOHNSACK, 2004). Ideally only one relevant characteristic should be varied at a time, so that a comparison can pinpoint the reason for observed difference (NOHL, 2001a; BOHNSACK, 2003). Consequently the intercultural applications of this method (NOHL, 2001; NOHL, 2001b; SCHITTENHELM, 2005) have focused on one age group and contrasted up to three cultural contexts which are represented by several cases each. 
For a research design which casts doubt on national categories, it is not possible to represent all possible lines of comparison in the sample. Instead a comparative analysis of habitus will focus on some predetermined lines of differentiation, and include some more which appear in the course of the research. Still, in a transnational context, we are likely to miss some determining factors for lack of understanding. Taking the habitual attitudes towards spending15) as an example, we can distinguish between three types. First, a large group of respondents emphasised that they were not "mean" people, which is to say that spending generously is part of their way of life ("investors"). For most of them, giving to charity became part of their identity and they actively looked for opportunities to "invest" in enterprises and people. A second group found financial security important, but suffered from the necessity to "invest". Another group lived more thriftily and did not mention charity at all ("savers"). 
The theoretical sampling of the study ensured that several factors thought to structure class position were kept constant, for example, high cultural capital, whilst some potential structuring forces were varied in order to ensure a strong contrast (e.g. the direction of migration across a threshold of inequality). The gender, family status and class background of migrants' families varied sufficiently in order to allow comparison between migrants of different background. None of these factors could completely explain the differences in habitus identified above. While some differences in habitus seem to be almost self-explanatory—those who had a hard time making ends meet were typically "savers"—"black swans" can always be found: some of the more wealthy people also lived thriftily. Likewise, coming from lower, middle or upper class families could not explain the habitual difference. 
In both the documentary method and grounded theory approaches results of this kind should prompt further field work. However, considering the wealth of possible structuring forces in a transnational social space it is realistic to assume that some variation in habitus must remain unexplained. Instead of explaining the structuring forces behind the types of habitus, we are content to note that the observed differences in habitus do not distinguish clearly between expatriates, who migrated from Germany to developing and transformation societies and IT-professionals who migrated from these societies to Germany. Members of both groups are represented in the three types delineated above. This makes it very unlikely that the difference between the types relates to national differences in wealth. Yet the—possibly transnational—factors which do, in fact, explain the observed differences in habitus, could only be found with further research. 
The second part of this article has argued that the search for a proper frame of reference is desirable from a theoretical point of view: a transnational epistemological perspective and the empirical analysis of transnational classes cannot presuppose a frame of reference. We have now seen that this theoretical consideration also translates into methodological concerns. In transnational social spaces neither factual information nor indicators of habitus can be placed in one single frame of reference. Nevertheless it is possible to reconstruct some structuring forces from the qualitative interviews. 
5. Assessing the Structuring Force of the Nation-State System
Lack of clarity with regard to the frame of reference for, and limited data on, transnational class formation potentially compromises qualitative research in this emergent field. Still, the openness of the grounded theory paradigm shows how conceptual and theory formation is possible even with limited existing knowledge. We will now show how the complex process of discovery (KLEINING & WITT, 2001) or abductive reasoning (REICHERTZ, 1999) operates to help us develop the link between the nation-state system and the economic position of highly skilled migrants. 
In a situation where it is difficult to decide who should be compared with whom, analysis can start with the comparative self-assessment of the interviewees themselves. As mentioned above, highly skilled migrants tend to compare their income to the average income in their country of origin and relate this to the cost of living there. For those living in Germany, it seems that they earn as much as native IT-professionals. German expatriates, in contrast, feel that they earn quite a lot more when they are abroad than when they stay in Germany. Both the IT-specialists and the expatriates argue that highly skilled persons who work for local employers or the state in developing and transformation societies earn a lot less than those who are employed by global players in Germany or abroad. 
When comparing interviewees' self assessments we must consider that as well as reporting knowledge about the labour market, they also try to justify their migration decision and distinguish themselves from others. Judgments about justice should not be confused with the structuring forces of inequality. It is, however, possible to reconstruct the reference points of their comparisons. It seems as if the economic situation of the highly skilled migrants is structured by nation-states (i.e. exchange rates, cost of living, inflation) as well as labour markets. For their labour market position qualification is important as is the (trans-) national status of the employing organisation. 
With regard to the influence of the nation-state, the different kinds of data point towards (apparently) contradictory results. Several of the migrants to Germany who felt that their standard of living was better in Germany than at home gave factual information about their economic situation which seemed to contradict their self assessment. For example, one of the poorer couples in the sample had a house and rode a motorcycle to work in India, whereas in Germany they shared a one-room-apartment and rode a bicycle. Still they felt that their overall standard of living was better in Germany than it was in India. 
The contradiction between self-assessment and factual information could be resolved by a theoretical consideration. Theories about public goods (KAUL, GRUNBERG & STERN, 1999) can show that life chances are not only determined by individual assets, but also by the infrastructure which an environment offers. Jessica Ramachandran of the above mentioned couple, for example, emphasised how she now lived in a beautiful and clean environment in the countryside. Her place of work could be seen from her house thereby avoiding long commutes, and she enjoys flexible working times. Taking public goods into account, the self assessment that she has bettered herself becomes convincing, despite the loss of some amenities which she enjoyed in India. This consideration points toward the important role of the state in providing public goods and thereby structuring transnational class positions. Living in a wealthy state, the IT-professionals gain access to public goods and exposure to these goods improves their situation, independently of their personal assets. 
The question now is whether this hypothesis can be generalised, or whether it is contradicted by "black swans". In contrast to the IT-professionals, the expatriates do feel that they have "bettered their lot" significantly through migration. Their income after migration is approximately 30% higher because of tax and social security exemption.16) However, despite this greater wealth, the argument about public goods still holds. Looking at the factual information expatriates provided about the standard of living, many reported problems with violent crime and related issues. Their spare time was limited to clubs and hotels, as they don't understand the language and found it difficult to gain access to local social networks. In their self-assessment expatriates felt they had moved to the upper class particularly because they sent their children to elite schools and because they could employ servants: indicators of an upper class position in their countries of origin. Looking at the expatriates' standard of living one could also argue that they needed elite schools, because the state schools do not work, that they need private security because the state cannot sufficiently control violence, and that even with private security their standard of living is still worse than that of a person who lives without private security in a safe state. 
In this case, inductive comparison of these two cases helps to point towards an important general finding. The IT-professionals and the expatriates both feel that they have bettered their lot. The IT-professionals earn similar amounts of income,17) irrespective of where they are located within a globalised labour market, yet their quality of life is improved by public goods and in particular a strong welfare state. The German expatriates gain personal income when they move to developing and transformation societies. Yet, to some extent this serves to compensate for a loss of amenities commonly shared in their country of origin. So their class positions are influenced by the public goods not offered by the state. 
The structuring force of the nation-state system influences self-assessment and factual information in an apparently contradictory manner. This prompted us to examine in further theoretical depth the reasons for the contradiction and the extent to which the findings could be generalised across different groups in the interview sample. We found that, in this example at least, it was possible to determine a background factor (the nation-state system) structuring the front stage (i.e. the habitus and resources) of the migrant groups. 
Starting from a BOURDIEUian concept of class the article has argued that a comprehensive analysis of class should take into account habitus as well as resource distribution. Both can be structured by the labour market, epistemic communities, the nation-state, and many other structuring forces. These forces cannot easily be determined from "front stage" information (EDER, 2001). 
Looking at the processes of transnationalisation in the political and economic sphere it is likely that class positions are increasingly structured by transnational rather than national economic and political forces. An analysis of transnational class formation is particularly relevant for migrants, who live in several nation-states and possibly inhabit a variety of transnational social spaces (PRIES, 1997, 1999). Transnational class positions can only be understood if we determine the extent to which specific states, the entire system of nation-states or transnational structures, influence class position. 
This means that transnational research cannot operate in a predetermined frame of reference. As "representative" research strategies have to determine first, what they want to represent, it would be quite difficult to use a quantitative paradigm of class analysis. Qualitative approaches to class, on the other hand, have stressed social group formation at the expense of the "back stage" structuring class positions. Comparative qualitative research has focused on specific cultures or milieus accepting only a limited degree of cultural variation. 
Considering the limits of both quantitative and qualitative approaches to class, the article has proposed early steps toward an analysis of class which does not presuppose the nation-state frame. We have suggest that grounded theory approaches work particularly well and have used the "qualitative experiment" to support the process of empirically-based theory building. This empirical-theoretical process is especially well suited for new fields of inquiry, such as this. 
Choosing a sample which maximises differences can be seen as a qualitative experiment which cannot prove, but could at least falsify the hypothesis of an emergent transnational middle class. Processes of data collection which include diverse types of data and analytical strategies increase the likelihood of "black swans", i.e. cases which (seemingly) contradict general findings. Open research strategies make paradoxical and contradictory observations possible and thereby promote the discovery of new findings. For example, the study about highly skilled migrants presented here shows that class position in transnational social spaces is structured by the public goods a nation-state provides. Nevertheless, it is also clear that whilst highly skilled migrants in global labour markets reach an economically similar position in social space, they do so along different paths. 
By choosing a sample from many different nations in order to highlight transnational similarities rather than intercultural differences, the study sacrificed the option of in-depth cultural analysis. This methodological strategy was critical, and it was a strategy that was intricately connected to the fact that the field of study is an emergent one, and that contemporary class analysis is a complex and multi-dimensional endeavour. In fact, the study's methodological approach diverges from typical qualitative approaches to migration; whilst it may overlook cultural specificities, it is able to engage in theory building and enhance our structural understanding of the social, economic and political dimensions underpinning highly skilled migration and transnational class formation. 
I would like to thank Karin SCHITTENHELM for important suggestions, Sam SCOTT for excellent editorial work and the German Research Association for funding this research.
1) The conceptual compatibility of the documentary method with BOURDIEU's theory has been discussed elsewhere (WEISS, 2001). It is important to add that the documentary method favours group discussions, but has also been applied to biographical interviews (BOHNSACK, 2003; NOHL, 2005). This article refers to semi-structured interviews which include extensive narrations, with the documentary method applied selectively to the narrative parts of the interviews. <back>
2) This may be attributed to top elites belonging to an older generation. Research in elite business schools has shown a tendency towards internationalisation (MARCEAU, 1989). <back>
3) In migration research the definition of transnationalism comprises diverse phenomena ranging from migration trajectories (GLICK SCHILLER, BASCH & BLANC-SZANTON, 1992) to networks, political actors (KEOHANE & NYE, 1973) and spaces (PRIES, 1999). <back>
4) Replacing the "national" frame with the term "world society" cannot avoid the problems of methodological nationalism (BRENNER, 2000; BECK, 2004). <back>
5) That definition is different from the common definition of the highly skilled which uses academic degree as the benchmark (see AURIOL & SEXTON 2002). <back>
6) The project entitled: "Highly skilled migrants: the transnationalization of social inequality" (2002 to 2005) was funded by the German Research Association and headed by the author. <back>
7) The project also included interviews with: contrast groups of migrants whose cultural capital depreciated in the course of their migration; sedentary professionals; and, African academics migrating to South Africa. <back>
8) For a focus on cultural capital see Anja WEISS (2005). <back>
9) "Insofern leben wir groesser luxurioeser als je zuvor. Und mh fuer ein Geld wofuer man hier das nich tun koennte. Also das sind irgendwie achthundert Dollar (...) im Monat. (...) dafuer wuerden wir hier kein Haus kriegen. Gemietet. Das is uns auch sehr bewusst. <C. lacht kurz, I auch> Und auch keinen tropischen Garten dazu" (3: 131). The numbers refer to the case number (3) and paragraph (131) of the transcript. <back>
10) Names and other details have been changed in order to ensure anonymity. <back>
11) Two very competitive young men stress that they make more than many German employees. Two women with families to support, who work as programmers outside of the IT-industry, feel they make less than many German employees. <back>
12) Exceptions are some migrants from very poor countries who find that their savings in Germany translate into a little fortune when transported to their home country. <back>
13) This argument is more obvious for educational titles as they must be acknowledged (see WEISS, 2005). <back>
14) Problems of this kind are not limited to self-assessment or qualitative research. Robert WADE (2001), for example, shows that an assessment of global inequality leads to different results depending on case construction and on whether one uses purchasing power parities or exchange rates. <back>
15) This interpretation uses findings by Claudia BAUER in her diploma thesis (2004). <back>
16) I have shown in another publication (WEISS, in print) that the IT-professionals pay twice (publicly and privately) for public goods whereas the expatriates tend to be relieved from taxes both at home and abroad. This explains much of the difference in their net income. <back>
17) Taking into account the cost of living. <back>
Auriol, Laudeline & Sexton, Jerry (2002). Human resources in science and technology. Measurement issues and international mobility. In Organisation for Economic Co-operation and Development Seminar on International Mobility of Highly Skilled Workers (Eds.), International mobility of the highly skilled [Collection of papers prepared for the Seminar on International Mobility of Highly Skilled Workers: From Statistical Analysis to the Formulation of Policies, Paris, June 11-12, 2001] (pp.13-38). Paris: OECD.
Barlösius, Eva (1999). "Das Elend der Welt". Bourdieus Modell für die "Pluralität der Perspektiven" und seine Gegenwartsdiagnose über die "neoliberale Invasion". BIOS, 12(1), 3-27.
Beck, Ulrich (1986). Risikogesellschaft. Auf dem Weg in eine andere Moderne. Frankfurt/M: Suhrkamp.
Beck, Ulrich (2002). Macht und Gegenmacht im globalen Zeitalter. Neue weltpolitische Ökonomie. Frankfurt/M: Suhrkamp.
Beck, Ulrich (2004). Der kosmopolitische Blick oder: Krieg ist Frieden. Frankfurt/M.: Suhrkamp.
Berger, Peter A. & Hradil, Stefan (Eds.) (1990). Lebenslagen, Lebensläufe, Lebensstile. Göttingen: Otto Schwartz & Co.
Bertaux, Daniel & Thompson, Paul (1997). Pathways to social class. A qualitative approach to social mobility. Oxford: Clarendon Press.
Bohnsack, Ralf (2003). Rekonstruktive Sozialforschung: Einführung in qualitative Methoden, Opladen: Leske + Budrich.
Bohnsack, Ralf (2004). Group discussion and focus groups. In Uwe Flick, Ernst. v. Kardorff & Ines Steinke (Eds.), A companion to qualitative research (pp.214-221). London: Sage.
Bohnsack, Ralf & Nohl, Arnd-Michael (2003). Youth culture as practical innovation. Turkish German youth, "time out" and the actionisms of breakdance. European Journal of Cultural Studies, 6(3), 367-386.
Bommes, Michael (2000). National welfare state, biography and migration. Labour migrants, ethnic Germans and the re-ascription of welfare state membership. In Michael Bommes & Andrew Geddes (Eds.), Immigration and welfare. Challenging the borders of the welfare state (pp.90-108). London: Routledge.
Bourdieu, Pierre (1984). Distinction. A social critique of the judgement of taste. Cambridge, MA: Harvard University Press.
Bourdieu, Pierre (1990). Social space and symbolic power. In Pierre Bourdieu, In other words (pp.123-139). Stanford: Stanford University Press.
Bourdieu, Pierre (Ed.) (1999). The weight of the world. Social suffering in contemporary society. Oxford UK: Polity Press.
Brenner, Neil (2000). The urban question as a scale question: Reflections on Henri Lefebvre, urban theory and the politics of scale. International Journal of Urban and Regional Research, 24(2), 361-378.
Coe, Neil M. & Bunnell, Timothy G. (2003). Spatializing knowledge communities. Towards a conceptualization of transnational innovation networks. Global Networks, 3(4), 437-456.
Eder, Klaus (1993). The new politics of class. Social movements and cultural dynamics in advanced societies. London: Sage.
Eder, Klaus (2001). Klasse, Macht und Kultur. Zum Theoriedefizit der Ungleichheitsforschung. In Anja Weiss, Cornelia Koppetsch, Albert Scharenberg & Oliver Schmidtke (Eds.), Klasse und Klassifikation. Die symbolische Dimension sozialer Ungleichheit (pp.27-60). Opladen: Westdeutscher Verlag.
Glaser, Barney G. & Strauss, Anselm L. (1967). The discovery of grounded theory: Strategies for qualitative research. Chicago: Aldine.
Glick Schiller, Nina; Basch, Linda, & Blanc-Szanton, Cristina (Eds.) (1992). Towards a transnational perspective on migration: Race, class, ethnicity and nationalism reconsidered. New York: New York Academy of Sciences.
Hartmann, Michael (2000). Class-specific habitus and the social reproduction of the business elite in Germany and France. The Sociological Review, 48(2), 241-261.
Hartmann, Michael (2002). Die Spitzenmanager der internationalen Grosskonzerne als Kern einer neuen "Weltklasse". In Rudi Schmidt, Hans-Joachim Gergs & Markus Pohlmann (Eds.), Managementsoziologie. Themen, Desiderate, Perspektiven (pp.184-208). München: Rainer Hampp.
Held, David; Grew, Anthony Mc; Goldblatt, David & Perraton, Jonathan (1999). Global transformations. Politics, economics and culture. Cambridge: Polity.
Hopf, Christel (1978). Die Pseudo-Exploration – Überlegungen zur Technik qualitativer Interviews in der Sozialforschung. Zeitschrift für Soziologie, 7(2), 97-115.
Iredale, Robyn (2001). The migration of professionals: Theories and typologies. International Migration, 39(5), 7-26.
Kaul, Inge; Grunberg, Isabelle & Stern, Marc A. (1999). Global public goods. International cooperation in the 21st century. New York, Oxford: Oxford University Press, UNDP.
Keohane, Robert O. & Nye, Joseph S. (Eds.) (1973). Transnational relations and world politics. Cambridge: Harvard University Press.
Kleining, Gerhard (1991). Das qualitative Experiment. In Uwe Flick, Ernst v. Kardorff, Heiner Keupp, Lutz v. Rosenstiel & Stephan Wolff (Eds.), Handbuch Qualitative Sozialforschung (pp.263-266). Weinheim: PVU.
Kleining, Gerhard & Witt, Harald (2001). Discovery as basic methodology of qualitative and quantitative research [81 paragraphs]. Forum Qualitative Sozialforschung / Forum: Qualitative Social Research [On-line Journal], 2(1), Art. 16. Available at: http://qualitative-research.net/fqs-texte/1-01/1-01kleiningwitt-e.htm [May 3, 2006].
Lamont, Michèle (1992). Money, morals & manners. The culture of the French and the American upper-middle class. Chicago: Unviersity of Chicago Press.
Lamont, Michelle (2000). The dignity of working men: Morality and the boundaries of race, class, and immigration. New York: Russell Sage Foundation.
Nohl, Arnd-Michael (2001). Migration und Differenzerfahrung. Junge Einheimische und Migranten im rekonstruktiven Milieuvergleich. Opladen: Leske + Budrich.
Nohl, Arnd-Michael (2005). Dokumentarische Interpretation narrativer Interviews. Available at: http://www.bildungsforschung.org/Archiv/2005-02/interview?set_language=de/ [Dec 23, 2005].
Pries, Ludger (Ed.) (1997). Transnational migration. Baden-Baden: Nomos.
Pries, Ludger (Ed.) (1999). Migration and transnational social spaces. Aldershot: Ashgate.
Reichertz, Jo (1999). Gültige Entdeckung des Neuen? Zur Bedeutung der Abduktion in der qualitativen Sozialforschung. Österreichische Zeitschrift für Soziologie, 24(4), 47-64.
Schittenhelm, Karin (2005). Soziale Lagen im Übergang. Junge Migrantinnen und Einheimische zwischen Schule und Berufsausbildung. Wiesbaden: VS Verlag.
Sklair, Leslie (2002). The transnational capitalist class and global politics. Deconstructing the corporate-state connection. International Political Science Review, 23(2), 159-174.
Strübing, Jörg (2002). Just do it? Zum Konzept der Herstellung und Sicherung von Qualität in grounded theory-basierten Forschungsarbeiten. Kölner Zeitschrift für Soziologie und Sozialpsychologie, 54(2), 318-342.
Strübing, Jörg (2004). Grounded Theory. Zur sozialtheoretischen und epistemologischen Fundierung des Verfahrens der empirisch begründeten Theoriebildung. Wiesbaden: VS Verlag.
Therborn, Göran (2001). Globalization and inequality. Issues of conceptualization and explanation. Soziale Welt, 52(4), 449-476.
Vobruba, Georg (1999). Die soziale Dynamik von Wohlstandsgefällen. Prolegomena zur Transnationalisierung der Soziologie. Soziale Welt, 46(2), 326-341.
Wade, Robert (2001). Global inequality. Winners and losers. The Economist, April 28th, 93-97.
Wallerstein, Immanuel (1979). The capitalist world-economy. Cambridge: Cambridge University Press.
Weiß, Anja (2001). Rassismus wider Willen. Ein anderer Blick auf eine Struktur sozialer Ungleichheit. Westdeutscher Verlag: Opladen.
Weiß, Anja (2005). The transnationalization of social inequality. Conceptualizing social positions on a world scale. Current Sociology, 53(4), 707-728.
Weiß, Anja (in print). Hochqualifizierte MigrantInnen. Der Kern einer transnationalen Mittelklasse? In Florian Kreutzer & Silke Roth (Eds.), Internationale Karrieren und transnationale Mobilität. Biografien, Lebensführung und Identitäten. Frankfurt/M: Campus Verlag.
Willis, Paul E. (1978). Learning to labour. How working class kids get working class jobs. Franborough: Saxon House.
Wimmer, Andreas & Glick Schiller, Nina (2002). Methodological nationalism, the social sciences, and the study of migration. An essay in historical epistemology. International Journal of Migration, 43(2), 576-610.
Xiang, Biao (2002). Ethnic transnational middle classes in formation. A case study of Indian information technology professionals. Paper presented at the 52nd Annual Conference of Political Studies Association (UK) "Making Politics Count" University of Aberdeen, 5.-7. April 2002.
Dr. Anja WEISS is a researcher and lecturer at the Department of Sociology, Ludwig-Maximilians-University Munich, Germany. She is currently working on a transnational theory of social inequality which builds on results from the project "Highly Skilled Migrants" funded by the German Research Association. With Arnd-Michael NOHL, Karin SCHITTENHELM and Oliver SCHMIDTKE she is heading the international study group "Cultural Capital During Migration" funded by the VW Foundation. Her general research interests are: social inequality, racism and interethnic conflict and migration.
Dr. Anja Weiss
Institut of Sociology
Weiss, Anja (2006). Comparative Research on Highly Skilled Migrants. Can Qualitative Interviews Be Used in Order to Reconstruct a Class Position? [46 paragraphs]. Forum Qualitative Sozialforschung / Forum: Qualitative Social Research, 7(3), Art. 2, http://nbn-resolving.de/urn:nbn:de:0114-fqs060326.