Building on Hancock, intersectionality is a social science paradigm: it is both a way to approach information and a way to explain data patterns. I argue that for intersectionality to be a successful paradigm employed by the majority of social scientists, we need to better identify and contend with intersectionality’s biggest methodological challenges. Considering survey data, three of the major, interrelated challenges are: (a) matching the best methodological solutions to each of intersectionality’s different theoretical strands; (b) appropriate measurement of intersectional groups; and (c) the small n problem. To contend with these methodological challenges intersectionality is in need of methodological innovations.
In this paper, I weave together three such innovations in survey design and analysis: mixed methods, survey data harmonization, and big data. Mixed methods allows researchers to generalize across populations while using qualitative approaches to delve deep into how they think and behave. Survey data harmonization (ex ante or ex post), at a large enough scale, turns into big data with a sufficient numbers of cases to construct and analyze nuanced intersectional groups. Qualitative textual data collected in a mixed method design or from other sources has potential to be harmonized with quantitative survey data and, at a large enough scale, can be used to analyze the thoughts and behaviors of intersectional groups. Advances in mixed methods, harmonization and big data analytics should lead to advances in surveys specifically designed to account for intersectionality. While each innovation introduces its own challenges, taken together, they can be creatively used to contend with intersectionality's longstanding methodological challenges for data design and analysis.
