Corpus-Driven Analysis in the Writing Center

Stinging Eyes, 'Perlin Sketch,' 2009.

Stinging Eyes, 'Perlin Sketch,' 2009.

In his May 19, 2015, blog entry, Thomas mentioned the empirical research that Isabelle Thompson (Emerita Professor, Auburn University) and I have been doing over the last seven or so years. Specifically, he mentioned our recent book Talk about Writing: The Tutoring Strategies of Experienced Writing Center Tutors (Routledge 2014). We have been pleased with the response we have received to the book, particularly the coding scheme for tutoring strategies that we developed and tested. It seems that writing center specialists are eager for more empirical research to guide our practices.

Isabelle and I have moved on to a new project, one that we think pushes beyond the call for more empirical research (see, for example, Pemberton, 2010) to a type of research (to our knowledge) unused in writing center scholarship: corpus-driven analysis.

A corpus can be any collection of naturally occurring language—talk or writing.[1] It could also be a collection of nonlinguistic communication, for example, gestures (see, for example, Caridakis et al. 2010). For our purposes, though, we use a narrower definition of a corpus: a sample of language from a particular discourse situation (Biber and Conrad 2009)—in our case, writing center conferences.

For our new book, The Aboutness of Writing Center Talk: A Corpus-Driven and Discourse Analysis, we report the results of a mixed-methods quantitative and qualitative analysis of more than 45 writing center conferences, explaining in detail the methodology as well as the analyses. .We have two main goals for the new book: (1) to show how quantitative corpus-driven analysis combined with qualitative discourse analysis robustly reveals the aboutness—the topics discussed and, as often as possible, how those topics are received by both tutors and student writers and (2) to inform tutor training by analyzing in detail at the micro- and macrolevels the specialized corpus of writing center conferences that we have gathered. To accomplish the first goal, we explain the theory and research that underlie and support our mixed-methods approach and examine studies of other instructional discourse contexts that have employed a combination of corpus and discourse analysis. To accomplish our second goal, we apply this mixed-method approach to our corpus. We are using corpus-driven analysis to identify and examine the basic measures of word counts, word frequencies, type/token ratios; key words; and word bundles (called n-grams) to understand what writing center discourse is about and, ultimately, to make research-driven recommendations for tutor training. We are supplementing our study with discourse analysis, a method that other writing center researchers have employed quite often (for example, Thonus 1999a, 1999b) and that we used in Talk about Writing and elsewhere as well (see Mackiewicz and Thompson 2013; Thompson and Mackiewicz 2014) so that we could analyze and reflect upon microlevel linguistic features identified through the corpus-driven analysis in their broader, macrolevel contexts.

We hope to publish our findings along the way as we work on Aboutness, and we hope to get your feedback on it as well.

[1] Naturally occurring language also includes sign languages, and corpora of sign languages have been developed. See, for example, the British Sign Language Corpus Project (2015) and the DGS Korpus, a corpus of German sign language (2015). 

Jo Mackiewicz is Associate Professor of Rhetoric and Professional Communication at Iowa State University.