Abstract
Simply put, social network analysis is an umbrella term for the study of social reality in which actors and relations are treated as nodes and edges in a graph. Thanks to its abstract core, social network analysis can attract social scientists from across a highly diverse landscape of empirical data and methodologies. The history of social network analysis is often and correctly portrayed as a fruitful marriage between ethnography and mathematics. Indeed, one of the joys of attending the yearly meeting of social network analysts, International Network for Social Network Analysis (INSNA), is the unconditional exchange of analytical ideas and empirical results in which the tiresome distinction between ‘quality’ and ‘quantity’ is largely absent. However, it cannot be denied that progress and methodological evolution in social network analysis has been and is much focused on developing more sophisticated computational, mathematical, and statistical tools. While this has certainly benefited a closer integration between the social and other sciences, it has also biased the field of social network analysis. Against this background, Domínguez and Hollstein’s volume on mixed methods is an interesting contribution to the social networks library.
Growing out of sessions on qualitative and mixed methods organized over the years at the INSNA meetings mentioned above, this volume provides an overview of the use of mixed methods research in the study of social networks. The editors have quite successfully managed to put together interesting contributions from 21 international experts into an accessible and appetizing volume. The book, which is a little short of 400 pages, consists of 12 chapters divided into three parts: Part I contains the introductory chapter and some general background chapters on social network analysis; Part II has four chapters that illustrate social network research that mixes ethnographic and different quantitative data; and Part III has another four chapters on what is termed ‘new methodological approaches’.
Hollstein’s introduction (Chapter 1) gives an overview of the whole project but more importantly disentangles and systematizes the idea of mixed methods in social network analysis, here defined as studies that: (a) are based on numerical network data and textual data; (b) use qualitative and quantitative interpretation; and (c) integrate analytical data analysis and/or interpretation. Needless to say, this definition covers a wide range of empirical studies. I am not fully convinced that the typology for classifying mixed methods research that is discussed in this chapter (distinguishing between sequential, parallel, fully integrated, embedded and conversion designs) has general analytical value. However, it does serve its purpose very well in organizing the different chapters into one coherent volume. Drawing on a wide-ranging literature and a genuine knowledge of social network analysis, this chapter itself is a very useful review and introduction to mixed methods social network research. While the remaining three chapters in Part I offer an introduction to social network research (Carrington), network data triangulation (Wald) and interpretation (Häussling), the core of this volume is Parts II and III.
The chapters on applications (Part II) all draw on significant empirical research projects to illustrate in different ways how mixed methods are being used in social network research. Bernardi and colleagues describe a data collection instrument applied in a comparative project on fertility decisions that combines semi-structured interviews with standardized network data collection tools. One of the benefits from such a parallel data collection design is the simultaneous collection of information on agency and structure. Maya-Jariego and Domínguez present a variation of such mixing of psychometrics and interviews, using psychometric network data and interviews to explore personal networks and the individual experience of acculturation. The iterative and sequential approach to data collection and analysis is extended further in Avenarius and Johnson’s study of network positions and beliefs about justice in rural China. They describe an ambitious longitudinal design that uses five instruments to collect data over a three-year period, constructing an increasingly deeper understanding of the role and meaning of social structure in everyday life. Concluding Part II, Gluesing and colleagues show how qualitative interview and observational data are used to validate and contextualize email communication data in a study of innovation networks, thus combining automated quantitative data-mining and ethnographic methods.
Where the chapters in Part II teach mixed methods research by example, the chapters in Part III discuss in more detail some selected methodological approaches and how they can be put to use in mixed methods network research. In turn, these four chapters discuss fuzzy-set qualitative comparative analysis (Hollstein and Wagemann), text analysis (Verd and Lozares), visualization (Molina et al.), and agent-based modelling (Rogers and Menjívar). Part III made less of an impression than Part II, perhaps because there are more comprehensive texts on the market about these topics, perhaps also because a section labelled ‘new approaches’ created higher expectations; that said, all contributions to Part III share valuable and insightful reflections on social network research design and analysis, drawing on rich research experiences. Moreover, the chapter by Molina and colleagues on the use of visualization for collecting and analysing personal network data provides an excellent methodological frame for some of the chapters in Part II.
Mixed Methods Social Networks Research details a sophisticated range of network studies and offers valuable lessons from network analysis in the making. I will remark that there is a bias towards small-n data and personal (ego-centric) networks. It would have been nice to also see research focusing on structures of large-n complete networks, moving beyond the (false?) idea that ethnography in some form may be a necessary component of mixed-methods research. However, this remark should not deflate the value of any of the chapters or of the volume as a whole. This volume gives a unique insight into the enormous potential of social network analysis and offers rich examples of how to tackle and solve challenges in research design, offering something for the novice and the experienced researcher alike.