Computer aided thematic analysis: introducing descending hierarchical classification

Thematic analysis and a number of related techniques focus attention on identifying "themes" or categories of information. Typically, early stages of the process involve extensive working over the source materials to identify and validate the identification of these themes before further analysis can proceed. Computer aided thematic analysis techniques can facilitate these processes by semi-automatically categorizing the source materials thus providing a replicable basis for producing materials for subsequent analysis. A number of such techniques are now available including topic modelling, propositional analysis of text, network analysis of text, and descending hierarchical classification (DHC).

Descending hierarchical classification was developed in France in the 1980s, and has been implemented in open source software (IRaMuTeQ) which I will describe in this webinar. The DHC process utilises document punctuation to identify segments of contiguous text likely to convey meaning. These segments are then classified using a clustering algorithm which groups similar text segments together thus approximating themes of the speakers or authors. Researchers can then interrogate these outputs assigning meaning to the clusters. The technique has been used for analysing interview and focus group transcripts, responses to open questions on surveys, and for literature reviews, and in combination with grounded theory. It has been used in a variety of fields including political science, health care, and analysis of social representations.

In this webinar I will introduce the technique with a focus on understanding how the text is processed, how we can interpret the output, and considering the implications for thematic analysis.

Speaker

Stephen Gourlay, Kingston Business School, Kingston University.

I have been teaching research methods to Masters’ and Doctoral level students in the Business School, principally on human resource management and occupational psychology programmes, for over 20 years. My focus has largely been on text analysis techniques and particularly on mixed methods approaches. I have been involved in a number of research projects on employee engagement, healthcare managers’ information use, and nursing work.