From Qualitative to Quantitative Analysis in a Large-Scale Migration Project

In recent years mixed methods research has gained traction in political science. It has become customary to begin with a large-n quantitative analysis and then proceed to the qualitative stage (quantitative → qualitative). Influential works by Lieberman (2005), Dunning (2007), Seawright and Gerring (2008), Rohlfing (2008), Brady and Collier (2010), among others, have emphasized how large-n analysis can be supplemented by the in-depth study of typical, outlier and other cases, identified by their relation to a regression line. Such research designs are appropriate for relatively well established research programs, but not for new research programs, where potential independent and dependent variables are not yet well defined, and where the relationships between them are not well tested. A rigorously designed qualitative analysis is thus necessary before any quantitative research design can be drafted (qualitative → quantitative). This paper presents a sequence of analyses conducted within the large-scale European Research Council Project “Diasporas and Contested Sovereignty.” The project seeks to understand different factors driving transnational diaspora mobilizations for homeland-oriented goals and what caused these mobilisations to be either sustained or episodic in nature. The research is designed in order to prepare for a large-scale survey to be conducted in five countries in Europe in 2016. First, qualitative comparative analysis has been conducted by four researchers in six cases studied in five European countries. Samples of interviews from these six cases were coded using an inductive method, which has been followed by the identification of most recurrent explanatory variables and their conceptual and property spaces. Deductive coding on a larger sample of interviews was then introduced using several external coders, inter-coder tests conducted, and an inter-coder reliability test implemented. A preliminary analysis of 40 sampled interviews was then subjected to correspondence analysis seeking to identify the most likely causes for sustained vs episodic diaspora mobilization. This paper will provide a template and practical advice for qualitative → quantitative research sequences, with emphasis on the importance of proper coding and inter-coder reliability testing.