Text Mining for Qualitative Data Analysis in the Social Sciences

Abstract

Two developments in computational text analysis widen opportunities for qualitative data analysis: amounts of digital text worth invest- igating are growing rapidly, and progress in algorithmic detection of semantic structures allows for further bridging the gap between qualitative and quantitative approaches. The key factor here is the in- clusion of context into computational linguistic models which extends simple word counts towards the extraction of meaning. But, to benefit from the heterogeneous set of text mining applications in the light of social science requirements, there is a demand for a) conceptual in- tegration of consciously selected methods, b) systematic optimization of algorithms and workflows, and c) methodological reflections with respect to conventional empirical research. This book introduces an integrated workflow of text mining applications to support qualitative data analysis of large scale document collections. Therewith, it strives to contribute to the steadily growing fields of digital humanities and computational social sciences which, after an adventurous and creative coming of age, meanwhile face the challenge to consolidate their methods. I am convinced that the key to success of digitalization in the humanities and social sciences not only lies in innovativeness and advancement of analysis technologies, but also in the ability of their protagonists to catch up with meth- odological standards of conventional approaches. Unequivocally, this ambitious endeavor requires an interdisciplinary treatment. As a political scientist who also studied computer science with specialization in natural language processing, I hope to contribute to the exciting debate on text mining in empirical research by giving guidance for interested social scientists and computational scientists alike.