Yes, but if you use NViVO or ATLAS ti...you can analysis your data without coding first

Coding texts allows is crucial to identify common themes and linguistic devices used, particularly when assessing the evolution of a discourse across time. The need for software assisted coding depends in large part on the size of your corpus. You may not need to do so for a small number of texts, but it is a great way to manage, organize and categorize analysis, and most packages have excellent functionality for re-arranging and representing relations that you find.

Critical discourse analysis (CDA) is primarily concerned with the effects of power on the production and dissemination of texts, as well as their ability to be authoritative. Therefore, an explicit theory of power should also inform you approach to coding. I have engaged CDA methodologically in a number of studies which may be of interest to you. See:

McDonagh, N. (2016). How to Frame a Bank Bailout: Lessons from Ireland during the Global Financial Crisis. Przestrzeń Społeczna/SocialSpace, 12(2), 84-112.

McDonagh, N. (2019). One-dimensional times: a dialectical response to the Irish Times’ coverage of the global financial crisis. Culture and Organization, 25(1), 32-51.

I have found Nvivo is a great package for qualitative coding, but have not yet used CAQDAS.

Good luck!

I think you need to step back from the question of coding and using software to instead start by examining some examples of CDA that you think were particularly well done. For those studies, you should examine how much data was actually presented, how that data was selected, and how it was analyzed. Then, for each of those questions, you need to consider the extent that the author(s) used coding and software.

Naoise McDonagh Thank you for your kind and thorough explanation. I have read the 2nd paper and I must admit it is eye-opening. One question persists though, do the themes arise out of the usual coding-to-themes (thematic analysis) route OR as a result of directly examining the discourse in each paper/doc? I also saw this paper from Mullet (2018) [Article A General Critical Discourse Analysis Framework for Educatio...

David L Morgan Thank you for the response. I think I have a fairly huge number of docs from diverse sources. So most probably I might be needing the software.

Second, many of the studies about/using CDA seems to conveniently skip the stepwise strategies to CDA methods and most don't mention if they used software or the pen/paper method.

Adi Fahrudin Thank you for your suggestion. Can you be a little more elaborate please?

Oluwole Badmus I would agree that an iterative approach is helpful. Each document in a corpus should be initially read on its own terms for the discourses within. But it is also the case that no text is an island unto itself, but is dialectically related to its broader discursive environment (see Norman Fairclough's work on this relationship) and indeed to ongoing empirical events, given that the times in which we read a text impacts and transforms our interpretation.

You will also have a wider selected corpus within which a single text is hypothesized as having some relationship. That relationship will only emerge iteratively over time, as you read more texts within the corpus, note emergent thematics, and then begin the process of coding. It really is a fascinating and rewarding process, but takes a lot of reading, thematizing, re-reading, coding in a stepwise process to get there.

Hope that helps.

Naoise.