The ever-increasing application of Digital Humanities techniques to social scientific research questions calls for continuous reflection on how they can contribute to scholarly research in combination with other more common text analysis methods. This article explores the various dimensions along which scholarly text analysis can be performed, using combinations of Natural Language Processing (NLP) and qualitative content analysis methods. The authors describe a practical methodological position based on a mixed-methods approach, arguing that a careful mix of automated language analysis using NLP and manual qualitative text analysis can help respond to the depth/breadth dilemma often encountered in case study research, whereby more detailed studies cover fewer cases, and vice versa (e.g. Gerring, 2007). The article demonstrates that using NLP and qualitative text analysis within a research workflow or cycle can allow researchers to move beyond the examination of the same data through the lenses of different methods for purposes such as corroboration, and towards a deeper exploration that reacts to the findings arising from different methods to hone and generate new avenues of research.