This chapter gives an account of our data and method, specifically outlining how we collected and prepared the texts containing extremist language that are the subject of this book, along with the different tools and techniques that were used for analysis. The chapter then carries out preliminary analyses of the data, using Biber’s multidimensional approach before moving on to describe a methodology called Corpus Assisted Discourse Studies (CADS) which involves a collection of approaches that are united by their use of software to identify linguistic patterns in large, electronically-encoded sets of data. We also describe how we obtained and classified keywords across the three sub-corpora which were used as the basis of focussing our analysis on a manageable set of lexical items.