Computer-Aided Discourse Analysis: Approaching Discourse in Large Text Collections using Multiple Correspondence Analysis


Bio: Dr. Isobelle Clarke is a Senior Research Fellow at the Centre for Corpus Approaches to Social Sciences at Lancaster University. In May 2020 Isobelle was awarded the Leverhulme Trust's Early Career Researcher Fellowship to examine anti-science narratives from a corpus linguistic perspective, which she began in May 2021. Isobelle received her PhD from the University of Birmingham in 2020. Her PhD research introduced a new modified version of Multi-Dimensional Analysis (MDA) that works with short texts and she applied this technique to a corpus of general English tweets and trolling tweets. More recently, Isobelle has developed this short text MDA approach, which uses Multiple Correspondence Analysis on the presence/absence of grammatical features, to the analysis of keywords in an approach she calls 'Keywords MCA'. The approach successfully identifies patterns of keyword variation and these are interpreted for the discourse they signal. Isobelle, along with Distinguished Professor Tony McEnery and Dr. Gavin Brookes have applied keywords MCA to a corpus of UK national press articles mentioning Islam and Muslims.

Dr. Gavin Brookes, Centre for Corpus Approaches to Social Science; Lancaster University

Dist. Prof Tony McEnery, Centre for Corpus Approaches to Social Science; Lancaster University

This workshop will introduce both keywords and a new approach to their analysis. Keywords offer analytical signposts to discourses in large volumes of text. Yet their interpretation often requires analysis of their use within their wider textual settings. Clarke et al. (2021) introduce Multiple Correspondence Analysis (MCA) as a new approach to organizing keywords statistically based on their co-occurrence across the texts of the corpus. The approach overcomes many of the issues in traditional keyword analyses and has proven to be effective for providing a more nuanced account of keywords that is sensitive to the various senses and discourses that a single keyword can exhibit. This workshop will provide learners with: (1) the understanding behind the MCA approach to keywords; (2) the tools to create a data matrix of variables and individuals; (3) the ability to run MCA in R; and (4) the skills for interpreting the results.