Comments
	The purpose of qualitative research is to understand and describe participants' experiences, to allow them to "tell their story." One means of doing this is to identify the common themes that emerge when participants describe their experiences in their own words. The goal is to produce an account of their experience that is faithful to what they have reported, that extends to other related contexts, and that can be audited in terms of the researcher's decisions. We want a re-telling of their stories that they will look at and say, "Yes, that says what it's like better than I could." Quantitative findings can lead to questions that need to be explored qualitatively just as interviews can lead to questions that need to be tested quantitatively. Thus, we take qualitative research to be complementary to quantitative research, providing a different view of a phenomenon. Researchers are involved with participants. Personal rapport and trust are crucial. The clinical intake interview is a good model for the interviews. The interviewer has a general direction for the interview to go ("What brings you here; who are you?" or "What is your experience of X?"). While keeping to that direction, the interviewer lets the participants speak for themselves and welcomes surprise turns in the interview. The interview is allowed to range over any topics that the researcher or participant feels are relevant. The researcher's personal feelings about the research and participants are important, although not to the participants. Guidelines about self-disclosure from clinical psychology are important, i.e., self-disclose only if it is in the interest of the participant. "Bracketing" is a useful strategy for recognizing and setting aside researchers' own reactions. It is OK, even necessary, for the interviewer to listen to and follow her/his intuition. Keep track of it, however, by leaving a "decision trail." Identify your choices and why you made them. This allows "auditing" of your findings. Below is one example of a qualitative strategy based on in-depth interviews. There is often overlap between these steps. Parts of the process are circular. For instance, the literature search can suggest reformulations of the research question and early interviews can suggest questions for later interviews.
Preliminary Steps
	1. Formulate the research question in general terms: What are participants' experiences of the phenomenon of interest? 2. Literature search. May include insights from literature, poetry, art as well as traditional scientific accounts. (Note: some argue that the literature search should be done after data collection and analysis so as not to bias the data collection. There seem to be strong arguments both ways.) 3. Define relevant characteristics of participants. Select participant pool. 4. Identify relevant areas to explore. Draft several open-ended questions. These questions will be used to start the interviews and other questions will be asked to keep the interview flowing. Record the process of deciding on questions for later auditing. 5. Pilot test questions and debrief with participants. 6. Finalize questions. Again, record the decision process. 7. Recruit participants. This step means deciding who will actually be interviewed. There are two approaches. One is to select participants, interview them, and analyze the themes until "saturation" is reached, until no new themes emerge. (In one study of elderly men, this meant 10 interviews per group.) Sometimes participants are selected for special characteristics: most and least successful, for example. Generally, the number of interviews is not very high.
Data Gathering
	8. Conduct interviews. Start with standard, open-ended questions and add new questions as they arise. The aim is to allow the participant to describe his/her own experience. Tape record and take notes to help guide the interview. Notes do not have to be thorough, the recording will do that. Rather, they can refer to non-verbal communication and other aspects of the interview the tape recorder does not get. Some researchers take pictures to help recreate the context of the interview. Of course, informed consent will always be obtained and other ethical principles followed. 9. Soon after the interview, record your own feelings and reactions. Record the decision trail that occurred within the interview. "Why did you ask that particular question?" 10. A shorter second interview may be conducted. See if any new information occurred to participants and clear up confusions from first interviews. NOTE: Steps 8-13 can go on simultaneously across informants. That is, early interviews may be transcribed while later ones are being conducted. Early interviews can raise issues that you feel should be asked in subsequent interviews.
Data Analysis
	11. Transcribe the interviews. This does not have to be done by the interviewer. 12. The interviewer reads the transcription while listening to the tape. Note feelings, nonverbal language, etc. 13. List topics (or "meaning units") with minimal editing or analysis. Examples of topics include behaviors, meanings, practices, episodes, feelings, roles, and relationships. There is computer software (such as Agenda, the Ethnograph, and Hypercard) to aid this process. The old standby-by is 3x5 cards organized into piles on the kitchen table. 14. Organize topics into themes and categories. There may be several levels of categories. Categories can be empirically-driven, letting the data speak for themselves, and they can be "theory"-driven, using categories that have already been established. And there should always be an "Other" category. Also, there may be a lot of trial and error in this stage, trying out themes and categories. 15. Identify categories: name, define, discuss, illustrate with examples from transcripts. Construct accounts of participants experiences: what is important. Importance may come from a high frequency of reports or from having a particularly meaningful impact. 16. Present descriptive statistics such as frequencies of the most common themes. 17. Search out negative instances. Welcome surprise. Discuss findings that stand out or don't fit. 18. Check categories examples, and accounts with others: staff, outside professionals, specific participants, "naive" commentators. Ask questions such as "Does this make sense, is it true to your experience, do you recognize your experiences here, what other questions does it raise for you?"