• No results found

You mentioned the term metadiscourse What does that mean to you?

Part III: Composition and FYC

Chapter 4: Methodology

Z: You mentioned the term metadiscourse What does that mean to you?

Honus: The points where the author is just talking about what he’s going to be doing. So, for example, a lot of times—for instance, I'm reading Dewey right now. Dewey— really consistently, every single chapter, he writes—the last couple of pages are a summary where he says this is what this chapter was about. So I make a note of that so if I want to come back to that chapter later, I can look there and remind myself that this was what the chapter was actually about. He tends to do the same at the beginning of the chapters. This is the problem we will be addressing. So I want to mark that point where he's not just engaging with the problem, but telling me the problem that he's engaging with— a higher level discourse.

Some transcription scholar-practitioners (Charmaz, 2002; Kvale & Brinkmann, 2009) argue that transcription requires a meticulous set of procedures that a researcher must disclose in considerable detail. While that may be true for some interviews—perhaps interviews where many informants are speaking at the same time, such as focus groups, or interviews conducted in multiple languages—that did not turn out to be the case for this study. Though arduous, my transcription process for the eleven interviews that I conducted was relatively straightforward. Over the course of three rounds I shaped, refined, then edited each transcription. During each round of transcription, I rewound the audio when necessary to ensure that I had met my goals for each iterative phase.

For the first round, I listened to each recording at approximately half-speed on

headphones, and using the “Speech Recognition” voice-to-text application on Google Docs, I narrated the interview out loud. During the second round, I listened to the interview at

approximately three-quarter speed and focused on including any words or phrases that weren’t captured during the first round. Here, I corrected any misspelled words. Finally, during the last round, I attended to rhetorical punctuation to capture the interview as a piece of authentic

communicative discourse.

In accordance with ethical, credible, and reflexive ethnographically-informed practice, I provided each interviewee with a transcription to ensure that member-checking standards were adhered to. Member-checking offers participants the chance to verify the accuracy of their transcribed responses, clarify certain statements if they feel compelled to do so, and on a holistic level, determine whether their comments seem to have been captured faithfully. None of the TAs decided to make any comments about or changes to the transcribed interviewees. Phases of Data Analysis

My analysis of the data that I collected in this mixed-methods study moved through five distinct phases. Each round of data analysis was driven by specific goals that would help me systematically explore my research questions in the most productive way possible. Those goals, in order, were to (1) determine which TAs to interview, (2) generate supplementary interview questions, (3) develop codes from the open-ended survey response and interviews, (4) analyze TAs’ three writing assignments as genre and develop codes from their typified features, and (5) compare codes across data points.

First, I began by exclusively focusing on the survey data to search for broad themes across TAs’ responses, along with the more individualized nuances that differentiated TAs’ perceptions—a distinction that helped me determine which TAs would likely offer the most productive follow-up interviews. Secondly, once I had pinpointed the specific TAs that I wished to interview, I examined their classroom resources—syllabi, assignments, and occasionally, handouts, and other in-class resources—to familiarize myself with each TA’s curriculum and search for any curiousities that might enhance our interviews. Then, thirdly, I returned to the survey data and examined it in conjunction with the interview data that I had just collected. I inspected these two data sources, together, in iterative stages. This third phase the most

formalized set of coding procedures. At this point, my goal was to develop a robust set of salient codes and exhaustive categories that would facilitate my exploration of the research questions grounding this project. In the fourth phase, I analyzed TAs’ assignments as typified genres with distinct textual features, and I created codes for those features. Finally, I used the codes

generated from phases three and four to triangulate the data—that is and paint a portrait of the ways in which this particular group of TAs attempt to guide students’ reading in their FYC courses.

Approach to Data Analysis

I took a grounded theory approach to analyzing the data across phases one through four. In the fifth phase—what could be considered my triangulation phase—I took a more deductive approach to determine whether the codes that I had generated via grounded theory in phases three and four were present in TAs’ assignments.

Although I had populated the survey questionnaire with fourteen reading behaviors— half of which surfaced through TAs’ own responses in pilot studies—I tried to let the data “speak” for itself in this study. That is, I attempted to avoid imposing pre-set categories onto the qualitative interview data. As Saldaña cautions, “Researchers should exercise caution with

Provisional Codes. A classic fieldwork saying goes, ‘Be careful: If you go looking for something, you'll find it,’ meaning that your preconceptions of what to expect in the field may distort your objective and even interpretive observations of what is ‘really’ happening” (p. 122). I felt that establishing the set of fourteen reading behaviors within the survey was an appropriate first step of this project, but I was not committed to finalizing each behavior as its own formalized category unless the data prompted such a decision. I realized that it was possible—perhaps even likely—that the qualitative data (the open-ended survey data and interview data) would

even if that meant expanding, collapsing, consolidating, or at the very least, qualifying any of the fourteen reading behaviors based on TAs’ responses. To enact this methodological goal, I adhered to grounded theory methodology and only coded the concepts that were present in TAs’ qualitative responses.