Corpus-Based Instruction and Crow

Corpus-based methods for teaching (writing) provide instructors with opportunities to heighten students’ awareness of language patterns through exposure to authentic texts (Charles, 2011; Shin et al., 2018); such exposure allows students to make contextually and rhetorically appropriate choices when writing in a given register/genre. Exposing students to corpora can aid in building students’ lexical variety and their ability to communicate successfully in academic contexts by providing students with examples of phrases, frequency lists, and lexicogrammar commonly used in particular contexts (e.g., nursing, engineering, business), or within particular genres. For example, if students are learning how to incorporate sources, corpora can provide developing writers with authentic examples of phraseology used for rhetorical conventions such as reporting from sources (e.g., reporting verb forms and functions).

While there are quite a few corpora available for educators and researchers to utilize (e.g., COCA, BAWE, MICUSP), there are limited learner corpora available focused on Foundations Writing (composition and writing classes specific to the early undergraduate experience). In this article, we showcase Crow (the Corpus and Repository of Writing), which is a unique learner corpus compiled of texts written by first-year students from the University of Arizona, Northern Arizona University, and Purdue University. A learner corpus is one comprised of learners of any given language, whether they are L1 or L2 students, as opposed to a corpus comprised of published journal articles, newspapers, or published stories by expert users.

Like any corpus, a learner corpus aims for balance and representativeness of the domains it is trying to mirror. Learner corpora include texts that are representative of the genres/registers that students are asked to engage with; in Crow there are 26 genres, representing the variety of assignments that students are asked to write at our research sites including argumentative papers, literature reviews, literacy narratives, and genre redesigns. Learner corpora have a variety of benefits for the SLW classroom including representation of topics that interest first-year writers, and accessibility to the writing level; notably, in both projects outlined below, the texts that were used for corpus-informed instruction were produced by and read by students of the same context.

From our corpus, we have been able to both conduct research and generate evidence-based pedagogical materials that target learners’ needs. Two such projects are detailed below.

The Reporting Verb Project

Based on results from an empirical study on L2 learners’ uses of reporting verbs in the Crow corpus (Kwon et al., 2018), we developed pedagogical materials using literature reviews from Crow. The goal was to observe whether or not corpus-informed instruction would increase students’ variety of reporting verbs and improve their functional understanding and use of reporting verbs. Using these corpus-informed materials, we conducted a 45-minute workshops in three international sections of Foundations Writing courses.

The workshops took place after students wrote their first draft, but before they began revising. We measured the effectiveness of our workshop by examining changes between drafts, and compared changes in the workshop group to drafts from three randomly selected sections of the same course from the same semester that did not receive corpus-informed instruction. We coded for lexical variety using four main categories: Argue, Show, Find, and Think (access Charles, 2011), and we coded functional uses under three main categories: reporting, self-reference, and uncited generalizations (access Shin et al., 2018).

After coding the data, we observed differences in students’ lexical variety and functional uses between the first and final drafts. Our study provided evidence that corpus-informed instruction leads to students using more variety of reporting verbs; however, to see a more meaningful difference in students’ understanding of the functional uses of reporting verbs, students would need to be engaged in ongoing instruction.

The Reporting Verb Pedagogical Materials

The materials created for our workshop serve as a model for corpus-informed instruction that can be easily adapted and used within most L2 writing classes. For this article, we wanted to provide a walkthrough of how to create these types of materials and implement them into classroom instruction.

We chose to use literature reviews since we expected that students would be citing sources in these texts. We selected excerpts of literature reviews from academic articles as well as from learners in our corpus. For each sample, we bolded the reporting verbs. Figure 1 shows an example handout created for this activity. Before handing out this activity, we provided a general introduction to reporting verbs including their uses, purposes, definition(s), and example sentences from the Crow corpus. Next, students were asked to identify the reporting verbs in the excerpts we provided. For each of the excerpts, students identified and observed the placement of reporting verbs, then noted and unpacked differences between the academic articles and student writing. It should be noted that we included a positive example from the learner corpus, so that students could see how writers at their level successfully implemented reporting verbs. Finally, students applied this knowledge to their own writing by highlighting the reporting verbs in their drafts, identifying which ones they used more frequently, and choosing new ones from the list (Figure 2) we provided focusing on revising for appropriateness, variety, and accuracy.

Figure 1. Published and student excerpts of literature reviews with reporting verbs bolded.

Figure 2. Reporting verb table with definitions and examples.

The CUES Project

The aim of the second project, sponsored by the Center for University Educational Scholarship, is to harness corpus tools to advance the teaching, learning, and assessment of writing in Foundations Writing courses at the university. In year one of the multiyear project, we chose to develop, implement, and assess materials for English 107, a genre-based writing course developed for first-year international students.

The project advanced in two phases. First, we held focus groups with instructors teaching English 107 to conduct a needs analysis. We then used their feedback to develop corpus-based activities to accompany units focusing on two different genres that are typically taught in English 107 and for which we had a sufficient number of texts in the corpus: literacy narratives and genre analyses. We explored the two genres using multiple types of corpus analysis, including word frequencies, keywords, phrases, and concordance lines. Based on the patterns of language use that we identified in these genres and the instructors’ suggestions and feedback, we then designed activities to accompany each project.

In Phase 2, the instructors implemented the corpus-based activities in their classes, and we conducted classroom observations of the activities in use, surveyed instructors and students about their perceptions of the materials’ effectiveness, and collected examples of students’ writing to assess the efficacy of the materials. Students’ texts were added to Crow, where they are available to include in future corpus-based analysis and materials development.

CUES Pedagogical Materials

One affordance of corpus-based materials is that they can be used to raise students’ awareness of choices for language use without being prescriptive or error focused, which was a priority for us and our participating instructors. For example, instructors told us that their students often used sentence-initial transition words repetitively in their genre analysis papers; though they did not want to treat repetitiveness as an error, they requested materials that could help expand students’ repertoire of linguistic choices. We then explored Crow for different kinds of transitions used in genre analysis texts and developed a sequence of three activities focused on using transitions for exemplification, using “this” plus a summary word as a transition, and varying transitions throughout a text. Modeled on activities in Swales and Feak (2012) and Grammar and Beyond (Bunting et al., 2013) but modified for the Foundations Writing context, these activities made use of multiple types of excerpts, including a chart quantifying instances of different transition words, one-paragraph excerpts, concordance lines (Figure 3), and what we call “Crowcordance lines” (Figure 4). Crowcordance lines are two-sentence excerpts illustrating the use of a keyword in its surrounding context. We developed them in response to participating instructors’ concerns that traditional concordance lines do not provide enough context for exploring word placement at the sentence level.

Figure 3. Example of concordance lines.

Figure 4. Example of “Crowcordance” lines.

Students and instructors responded positively to the materials; most students reported that the activities helped them with their writing projects. From classroom observations, we confirmed that corpus-based activities work best when integrated with other types of activities. Students also expressed that they appreciated multiple chances to work with corpus-based activities, since it took some time to adjust to them, especially activities based on quantitative corpus data. We are creating videos to provide a better introduction to the materials and the corpus so students understand how the quantitative data were generated and why they are meaningful to analyze.

Conclusion

Our projects demonstrate the value and importance of evidence-based teaching while also providing opportunities for students to engage with texts that are both accessible and representative of the types of writing they encounter during their undergraduate studies, particularly in a first-year writing context. We invite readers to learn about Crow here https://writecrow.org/, and to consider using our web-based tool for their teaching purposes https://crow.corporaproject.org/. Access is freely available to instructors and researchers after filling out a short survey.

References

Bunting, J. D., Diniz, L., & Reppen, R. (2014). Grammar and beyond Level 4. Cambridge University.

Charles, M. (2011). Adverbials of result: Phraseology and functions in the Problem–Solution pattern. Journal of English for Academic Purposes, 10(1), 47-60.

Kwon, M. H., Staples, S., & Partridge, R. S. (2018). Source work in the first-year L2 writing classroom: Undergraduate L2 writers' use of reporting verbs. Journal of English for Academic Purposes, 34, 86-96.

Shin, J., Velázquez, A., Swatek, A., Staples, S., & Partridge, R. (2018). Examining the effectiveness of corpus-informed instruction of reporting verbs in L2 first-year college writing. L2 Journal, 10(3), 31-46.

Swales, J. M., & Feak, C. B. (2012). Academic writing for graduate students (3rd ed.). University of Michigan.

Nina Conrad is a PhD student in second language acquisition and teaching at the University of Arizona.

Dr. Shelley Staples is associate professor of English applied linguistics and director of second language writing at the University of Arizona.

Kevin Sanchez is an undergraduate student at the University of Arizona, studying English and creative writing.