Analyzing Interview Data: The Development and Evolution of a Coding System Cynthia Weston,1 Terry Gandell, Jacinthe Beauchamp, Lynn McAlpine, Carol Wiseman, and Cathy Beauchamp
This paper describes the process used by a research team to develop a coding system for analyzing data from interview transcripts and situates the process within approaches to qualitative analysis. Successive versions of the coding scheme illustrate its development over several years; the role of team members and veriﬁcation in this evolution are discussed. Several lessons emerge from our experience: a) coding is not what happens before analysis, but constitutes an important part of the analysis; b) a team builds codes and coding builds a team through the creation of a shared understanding of the phenomenon; and c) collaborative qualitative research requires a kind of rigor that an independent researcher might not be aware of or need. KEY WORDS: coding; interview data; qualitative research; research methodology; data analysis.
INTRODUCTION Much is written about methods for coding interview data: ways to think about it, how to do it, how to document it, and how to report it. We have coded mountains of verbal data in the form of interview transcripts in various contexts during the past decade, and have reported the results of our investigations (e.g., Weston et al. 1997; McAlpine et al. 1999a; McAlpine et al. 1999b). However, these reports are the tip of the iceberg. We have not written about what lies below the surface— the enormous task of developing coding systems and coding the transcripts that provide the very basis for reporting the results of our investigations. Our research team (the authors of this article) has worked together since 1994 to develop a 1 Correspondence
should be directed to Cynthia Weston, Department of Educational and Counselling Psychology and the Center for University Teaching and Learning, McGill University, Montreal, Quebec, Canada H3A 1Y2; e-mail: email@example.com. 381 C
2001 Human Sciences Press, Inc.
Weston, Gandell, Beauchamp, McAlpine, Wiseman, and Beauchamp
coding system to analyze and understand our current data. We have learned a great deal about coding during this time that may be of interest to other researchers who deal with similar data (interview transcripts), use similar methods (coding), and who work or imagine working collaboratively in qualitative research. Maxwell (1990) notes that there has been a great deal of abstract debate about different paradigms in education, but little analysis of actual examples of research that combine different methods to determine what works and why. In this article we provide an example of an approach to coding verbal data that we believe represents a combination of methods. For example, we used a priori theory to frame our questions, drive our interview protocols, and structure the initial levels of the coding scheme. We moved into a more grounded approach as we discovered codes working through the transcripts. Inﬂuenced by Berliner’s (1992) notion of telling stories, we share this story of how our coding system was developed and how it continues to evolve. We describe our context, the process we used to develop the coding system, how we went about verifying codes, and our codebook. We believe there are several lessons in this story which are woven through the account. First, coding is not what happens before analysis, but comes to constitute an important part of the analysis. Second, a research team builds codes and coding builds a team through the creation of shared interpretation and understanding of the phenomenon being studied. Third, collaboration in qualitative research requires a kind of rigor that a lone or independent researcher might not be aware of or need. In this account, we try to “keep a sense of voice present” (Eisner 1998, p. 2) so readers will know the people behind the...