The social space of relations: Local and nonlocal networks of migrants in Istanbul

Residential segregation in metropolitan cities is a long-standing phenomenon, and the research on residential segregation includes an extensive amount of works from different fields. It is still a popular subject because residential segregation is an aspect of the structural changes that the cities are facing under shifting global and local conditions. Residential segregation patterns are reproduced by both intra-urban mobility of residents and location choices of new migrants (both internal and international), which can be seen as responses to the changing conditions. Therefore, migrants' residential segregation in metropolitan cities is among the important subjects to be explored. Quite a number of studies on residential segregation of migrants do not consider the relationships of migrants at all. However, residential segregation of migrants could not be explained by only the intrinsic characteristics of the migrants. Social relations within and between groups play an important role on the formation and transformation of segregation patterns. The reverse effect is equally true; physical spaces operate as settings in which relations emerge (or do not emerge), strengthen or disappear. Thinking of the interplay between social networks and residential segregation, it is obvious that each influences the other to create a spiral structure. It is not a matter of which is the first to have an influence on the other, and which is the one that is influenced by the other. The important matter about the interplay between residential segregation and social networks of residents is to reveal people's networks by referring to their spatial and social aspects in order to explain how local and nonlocal ties operate for people with different characteristics in different contexts (i.e. local settings). Therefore, this thesis explains residential segregation of migrants by focusing on the relationships. This thesis aims to investigate the structure and composition of the social networks of migrants who are residentially segregated within the city of Istanbul, and to reveal the forms of local and nonlocal ties of the migrants. In order to do this, the thesis is built around the concept of social space. Social space from a network perspective is the space of social relations which enables us to take relations as the unit of analysis without making a priori definitions and categories of actors. Structure and composition of networks and patterns of relationships correspond to social space. Further, social space is a generated space which is defined during the process of analysis. The aim of this thesis is not to find causal relationships between variables, but to capture different contexts in the generated space of social relations. The attributes and behaviours of individuals will then be examined within different contexts. At the same time, the geographical units in which these relationships occur are evaluated. This kind of approach is not used in other studies which try to find linear associations between social networks and migrants' attributes. This thesis will introduce such an approach, which derives its starting point from the concept of social space, and explores the social space of relationships of the migrants in Istanbul. According to the results of the Correspondence Analysis (see Chapter 4) that reveal the districts in which internal and international migrant groups concentrate, by reason of the high over-representation of international migrants who came from Macedonia, Albania and old Yugoslavian countries in the district of Bayrampasa; high over-representation of internal migrants who came from big metropolitan cities of Turkey, and slight over-representation of international migrants who came from Macedonia, Albania and Sandzak region in the district of Besiktas, the study areas are designated as Bayrampasa and Besiktas districts. In this thesis, the focus is on the international migrants who came from Macedonia and Sandzak region and who live in Bayrampasa and Besiktas, and internal migrants who came from the second biggest metropolitan city Ankara and its border city of Eskisehir and who live in Besiktas. These two districts are different from each other in terms of their geographical locations in the city, land use and social structures. The empirical analysis of this thesis consists of the combined use of Social Network Analysis, Multiple Correspondence Analysis and Hierarchical Cluster Analysis of a dataset which is obtained by 126 face-to-face personal network surveys including information about the migrants' attributes, attributes of the persons (24 alters for each migrant) in their personal networks, and the relationships between migrants and the persons in their networks (3024 ties in total). The relationship between residential segregation and local and nonlocal composition and quality of the networks of migrants is explored by using the information on the locations of the alters in migrants' networks. The community-saved and community-liberated arguments are both in effect for different migrants groups. Further, the networks with a high share of local (i.e. neighbourhood) ties are relatively resource-poor, while the networks with a low share of local ties are relatively resource-rich. The homophily principle is in effect for almost all migrants in that the high-educated migrants' network members are also high educated, which renders their networks resource-rich, while the reverse is equally true. Although Bayrampasa is a district in which the migrants from Macedonia, Albania and Sandzak region highly concentrate, it is also demonstrated that the migrants from Ankara and Eskisehir (and to a lesser extent the migrants from Macedonia, Albania and Sandzak) are concentrated in Besiktas. However, the structure and composition of the networks of the migrants in these two districts are quite different from each other. Therefore, it is not a foregone conclusion to say that concentration of migrants in a neighbourhood or district hinders forming or maintaining ties with other groups, since there are many other factors in the complex world of relationships. Moreover, for all migrant groups, the metropolitan area bounds the field of interaction more than the neighbourhood does, in that almost all migrants have connections within the metropolitan city no matter whether their neighbourhood ties take more or less space within their networks. The combined use of Social Network Analysis, Multiple Correspondence Analysis and Hierarchical Cluster Analysis helped in recognizing the patterns hidden in the data and assigned positions to the networks and ties so that I did not lose any case which does not have similar characteristics with the general profile. Neither the relations nor the individuals were partitioned, and their entirety was not lost during the analyses. I had the chance to find the relations and individuals who are connected to them after I generated the social space of relations. Therefore, further qualitative research can be made by interviewing the migrants (or even alters in their networks) who conform with the general tendencies, and those who have unique features. The conceptualization of social space as the space of relations allowed me to bring physical space into the analysis by considering the residential locations of the migrants and alters in their networks. I was able to typify the networks and ties with a direct reference to place by examining how residentially segregated migrants relate to the outside world (their neighbourhoods, districts, metropolitan city, place of origin, or elsewhere in the world) over their personal networks which lead me to demonstrate both the community-saved and liberated arguments are to be in effect in the city.