As one can see, in-depth interviewing generates an enormous amount of text. The vast array of words, sentences, paragraphs, and pages have to be reduced to what is of most significance and interest (McCracken, 1988; Miles & Huberman, 1984; Wolcott, 1990). Most important is that reducing the data be done inductively rather than deductively. That is, the researcher cannot address the material with a set of hypotheses to test or with a theory developed in another context to which he or she wishes to match the data (Glaser & Strauss, 1967). The researcher must come to the transcripts with an open attitude, seeking what emerges as important and of interest from the text.
At the same time, no interviewer can enter into the study of an interview as a clean slate (Rowan, 1981). All responses to a text are interactions between the reader and the text (Fish, 1980; Rosenblatt, 1982). That is why it is important that the researcher identify his or her interest in the subject and examine it to make sure that the interest is not infused with anger, bias, or prejudice. The interviewer must come to the transcript prepared to let the interview breathe and speak for itself.
Marking What Is of Interest in the Text
The first step in reducing the text is to read it and mark with brackets the passages that are interesting. The best description I have read of this aspect of the winnowing process is Judi Marshall’s “Making Sense as a Personal Process” (1981). She acknowledges that what she can bring to the data is her sense of what is important as she reads the transcripts. She expresses confidence in being able to respond to meaningful “chunks” of transcript. She says that she recognizes them when she sees them and does not have to agonize over what level of semantic analysis she is doing. She affirms the role of her judgment in the process. In short, what is required in responding to interview text is no different from what is required in responding to other texts-a close reading plus judgment (Mostyn, 1985).
Marshall also talks about the dark side of this process: that time when, while working with interview data, you lose confidence in your ability to sort out what is important, you wonder if you are making it all up, and you feel considerable doubt about what you are doing. You become worried that you are falling into the trap of self-delusion, which Miles and Huberman (1984) caution is the bane of those who analyze qualitative data. Marshall (1981) calls it an anxiety that you learn to live with.
It is important that researchers acknowledge that in this stage of the process they are exercising judgment about what is significant in the transcript. In reducing the material interviewers have begun to analyze, interpret, and make meaning of it. The interviewer-researchers can later check with the participants to see if what they have marked as being of interest and import seems that way to the participants. Although member-checking can inform a researcher’s judgment, it cannot substitute for it (Lightfoot, 1983). That judgment depends on the researcher’s experience, both in the past in general and in working with and internalizing the interviewing material; it may be the most important ingredient the researcher brings to the study (Marshall, 1981).
Although I can suggest some of the characteristics that make interviewing texts meaningful to me, there is no model matrix of interesting categories that one can impose on all texts. What is of essential interest is embedded in each research topic and will arise from each transcript. Interviewers must affirm their own ability to recognize it.
There are certain aspects of individual experience and social structure to which I respond when they appear. I am alert to conflict, both between people and within a person. I respond to hopes expressed and whether they are fulfilled or not. I am alert to language that indicates beginnings, middles, and ends of processes. I am sensitive to frustrations and resolutions, to indications of isolation and the more rare expressions of collegial- ity and community. Given the world in which we live, I am sensitive to the way issues of class, ethnicity, and gender play out in individual lives, and the way hierarchy and power affect people (Kanter, 1977). I do not, however, come to a transcript looking for these. When they are there, these and other passages of interest speak to me, and I bracket them.
Even when working with a research team, I give little instruction about marking what is of interest in a transcript other than to say, “Mark what is of interest to you as you read. Do not ponder about the passage. If it catches your attention, mark it. Trust yourself as a reader. If you are going to err, err on the side of inclusion.” As you repeat the winnowing process, you can always exclude material; but materials once excluded from a text tend to become like unembodied thoughts that flee back to the stygian shadows of the computer file, and tend to remain there. (See Vygotsky, 1987, p. 210.) Despite my open instruction about marking transcripts, I have often found considerable overlap among my colleagues in what we have marked.
Source: Seidman Irving (2006), Interviewing As Qualitative Research: A Guide for Researchers in Education And the Social Sciences, Teachers College Press; 3rd edition.