DescriptionMuch has been written within the CADS literature that corpus linguistic tools offer the perfect combination between qualitative and quantitative analyses: “the numbers tell us where to look closer,” being a popular trope. At the same time, however, any findings gleaned through corpus linguistics can themselves be validated and enriched by other methods through triangulation. Denzin (1970) identifies 4 types of triangulation: investigator, data, theoretical, and methodological. Regarding the latter category, Marchi and Taylor (2009) make a distinction between between-method triangulation and mixed-methods: the former refers to scenarios where two methodologies are applied separately, before comparing findings; whereas the latter refers to two methods that are intertwined and interdependent on each other. CADS is, of course, a perfect example of this latter category. The corpus-assisted part twins with the discourse analytical part; it is, in a sense, already ‘triangulated’.
There have been two edited collections which see corpus linguistic techniques involved in methodological triangulation. The first is Baker and Egbert (2016) who gave the same dataset to 10 scholars and asked them to perform different corpus linguistic techniques on it. The second is Egbert and Baker (2020) who asked for datasets to be approached both from the CADS perspective, and also from the perspective of another method within linguistics. In both cases, the authors find that methodological triangulation enriches the analysis: “any single methodological approach is only giving us a fraction of the full picture” (Baker and Egbert, 2016: 205). At the same time, there is a growing body of evidence that CADS is now being used in disciplines other than linguistics (albeit often in collaboration with a linguist). This is welcome and speaks to the utility of CADS as a method. With this in mind, and building upon Baker and Egbert (2016) and Egbert and Baker (2020), we take the approach one step further: we test whether it is possible to combine CADS within another method from outside linguistics. As one of the most popular methods across the social sciences, an obvious place to start is content analysis.
Content analysis serves as a good middle ground between quantitative and qualitative endeavours: like CADS, it can do both. The benefit of content analysis is that, rather than starting with a lexical search, like in CADS, we can begin instead with a theme or phenomena. In this paper, we reflect on the utility of using these two methods in tandem, and present a worked example of this in action. This worked example is achieved by: 1) applying a content analysis codebook to a corpus; 2) extracting those statements identified as belonging to specific criteria/phenomena; 3) performing a CADS analysis on that extracted data, to see how those phenomena are linguistically constructed.
Manual annotation of corpus samples is akin to traditional content analysis. Through this case study, we wish to explore the potential usefulness of such a combination of methods, and discuss other ways in which the two could work together.
|Period||27 Aug 2022|
|Event title||6th Corpora and Discourse International Conference|
|Location||ItalyShow on map|
|Degree of Recognition||International|