QUALITATIVE RESEARCH (Part 3)

CONDUCTING INTERVIEWS








Types of Interviews. Qualitative researchers are interested in using interviews to make sense of people's actions in naturalistic settings. The language used by respondents is a powerful means to accomplish this goal. Basically, the interview is a face-to-face or indirect (for example, telephone) interaction be¬tween the investigator and one or more participants in which the investigator asks the respondents questions in order to obtain information.
Interviews can be classified into three types: structured interviews, unstructured interviews, and group interviews (Fontana & Frey, 2000). In a structured interview, the questions and the order in which they are asked are determined a priori. Moreover, responses are classified into categories or are quantified according to some protocol that also is developed a priori. The interviewer develops rapport with the respondent but takes a neutral stance in that he or she does not show either approval or disapproval of responses and does not follow up on unusual, interesting, or uninformative responses. The advantage of structured interviews is that they are standardized across respondents and minimize variations. However, they have limited usefulness in qualitative research because they (1) shape data to conform to structures that emanated from the investigator's previously held beliefs about the phenomenon, (2) use a standard language across respondents (rather than questions customized to the language of particular respondents), and (3) restrict affective components of responses (Fontana & Frey, 1994). Although technically not an interview, the questions can be printed and administered to participants, which is similar to a questionnaire except that the responses are not constrained in any way.
On the other end of the continuum, unstructured interviews provide latitude to explore the responses of participants and to adapt questions for respondents. Qualitative researchers often use unstructured interviews to collect data, and the responses would not be quantified. The type of question asked in qualitative research is shaped by the type of research conducted; thus, ethnographers approach this endeavor differently from grounded theorists, for example. Because the questions are not determined a priori, the interviewer has great latitude to explore the phenomenon and ask probing questions to get a more complete description. The respondents are encouraged to use their own language to describe their experiences; however, the interviewers must take care not to shape the responses by covertly reinforcing certain types of responses (for example, following up only on responses that fit the researchers' assumptions). As Polkinghorne (2005) suggested, "although the produced account is affected by the researcher, it is important that the participant remain the author of the description" (p. 143). Also, in the case when more than one researcher is conducting interviews, interviewers need to discuss how each person's personal styles may influence the participants' responses and subsequent analytic results. Furthermore, it should be noted that that data from unstructured interviews take a lot more time to collect and organize. Researchers usually need to do follow-up interviews because new insights about the phenomenon (and thus, additional questions) may emerge as more participants are interviewed.
Most of the interviews used in counseling research span the structured/ unstructured continuum. Researchers often use semi-structured interviews to provide some consistency across interviews but also allow the respondents to have ample opportunity for offering richer responses. The challenge of utilizing the semi-structured format lies in the decision of how much structure should be imposed on the interview process. For an example of a relatively unstructured interview, consider Rennie's (1994b) study of clients' deference in psychotherapy, in which clients viewed a videotape of a recent counseling session (a variation of Interpersonal Process Recall; Kagan, 1975) and stopped the tape when they were experiencing something interesting or significant. The interviewer then "conducted a nondirective inquiry into the recalled experience" (Rennie, 1994b, p. 429), creating a description of the client's perceptions of the therapy in the client's own language without imposing predetermined categories for understanding these descriptions. As an example of a more structured approach, Frontman and Kunkel (1994), who also wanted to access clients' reactions to therapy, asked clients immediately after a session (rather than being stimulated by a videotape) to respond to the following probe: "As if you were making a personal journal entry, write what you felt was successful about the session" (p. 493). Because this probe was responded to in writing, no follow-up questions were used. Although the probe for this study was standardized, no preconceived categories were used to make sense of the responses, so that the conclusions were grounded in the data and not structured by the investigators' prior understanding of the phenomenon.
A third type of interview is the group interview, in which more than one..., person is interviewed simultaneously. Madriz (2000) describes a focus group as "a collective rather than an individualistic research method that focuses on the multivocality of participants' attitudes, experiences, and beliefs" (p. 836). Particularly exciting possibilities exist for focus groups, a modality that evolved from marketing research related to reactions to products, advertisements, and services. (For a brief history of the development of the focus group method, see Morgan, 1988, or Madriz, 2000.) In marketing research, the members of a focus group are strangers, but applications to qualitative research suggest possibilities for using intact groups, such as the staff of a mental health clinic. The goal of a focus group is to obtain the participants' opinions, not to reach a consensus or reconcile opposing views; the expression of different opinions is informative. Typically, the participants in a focus group are relatively homogeneous and are asked to reflect on a particular issue. The economy of using group interviews is self-evident, but a less obvious advantage of the group format is the interaction among the respondents, which can provide richer information as various members of the group provide more details, disagree on points, and reconcile differences of opinions. Having participants respond in a social context is thought to provide honest and responsible comments. However, the interviewer needs to ensure that minority opinions are allowed expression, given the natural tendency for such opinions to be suppressed (Madriz, 2000). A good illustration of this method has been provided by O'Neill, Small, and Strachan (1999) who used focus groups to explore the employment issues facing people with HIV/AIDS.
Methods of Interviewing. Counselors are well trained in the art of asking questions, and there are many similarities between the counseling interview and the qualitative research interview. However, note that "research interviewing-has different goals and requires different skills" (Polkinghorne, 2005, p. 143). Although the goals of qualitative interviewing will depend on the type of qualitative research conducted, a generic goal is to understand the experience of the participant; therapeutic actions should be avoided. Table 11.4, a typology of questions suitable for qualitative research, provides some examples of questions aimed at gaining understanding at various levels.

We now briefly examine the various steps in conducting qualitative interviews, adapted from Fontana and Frey (2000). The first step is to gain entree to the setting in a way similar to gaining entrée for observations. Wampold et al. (1995) were able to gain access to two chemistry groups by first obtaining the support of the laboratory leader. Although all the members subsequently agreed to be observed, several declined to be interviewed because of the time it took away from their work.
The second step involves preparing to understand the language and the culture of the interviewees. A qualitative investigator strives to have participants express their experiences in their own language. To achieve that goal, the questions should be understandable and nonoffensive to participants. Also, the interviewer must instantly understand what the interviewee means by various idioms and expressions so that appropriate follow-up questions can be asked. This knowledge is acquired from previous experience and preparation and is accumulated over the course of the interviews. Moreover, knowledge of language and culture is needed to make sense of the interviews in the data analysis phase, which is described in the next section.
The third step is to make a decision about self-presentation. Whereas quantitative researchers present themselves as objective scientists, qualitative researchers can choose among various self-presentations. For example, a researcher might present herself to abused women as a feminist as well as a researcher. Basically, the issue here is the degree to which the researchers should share of themselves and how it would affect the participants' willingness to share their stories or personal accounts.
The fourth step is to identify the interviewees. The central concept in this step is key informants, people whose perceptions are particularly important for understanding the context being studied. Of course, the key informants will likely be unknown to the investigator initially, but as knowledge is gained their identity will emerge. Key informants are often those who have a different experience than the norm, who are willing to divulge sensitive information, who are formal or natural leaders, and so forth. Moreover, key informants may not be directly involved in the phenomenon being studied. For example, the support staff in a mental health clinic could provide important information relative to counseling at the agency, even though they neither deliver nor receive treatment. In quantitative research the emphasis is on representative samples, but in qualitative research the goal is to obtain a depth of understanding, and thus the choice of interviewees is a crucial process.
The fifth step is to establish rapport with each interviewee, a process that is beneficial for two reasons. First, rapport leads to trust, which in turn leads to honest and descriptive responses. Second, an empathic stance enables the interviewer to better understand the interviewee's responses. Interviewers must take care, however, that the natural affinity that goes hand in hand with empathy does not cloud their assessment of the situation.
The sixth and final step is to decide how the interview data will be collected. If done unobtrusively, interviews should be recorded and subsequently transcribed for analysis. In any case, field notes should be taken. (Note that the comments regarding field notes for observations apply here as well.) Important information is contained in the nonverbal responses of the interviewees, and in the setting and its surroundings. Therefore, certain critical data may be missing when interviews are conducted over the phone or via email.
It should be noted that interviewing is the predominant mode of obtaining data in qualitative research in counseling psychology, and it is deemed one of the most difficult and advanced techniques to master (Fassinger, 2005; Hill et al., 2005; Polkinghorne, 2005). Therefore, neophyte qualitative researchers are encouraged to hone their skills in this area. Kvale (1996) and Douglas (1985) discuss qualitative interviewing in depth. In particular, the procedures for conducting focus group interviews can be found in Greenbaum (1998) and Vaughn, Schumm, and Sinagub (1996).

USING EXISTING MATERIALS
Existing materials are written text and artifacts, which can often inform qualitative research in ways that observations and interviews cannot; such materials are essential to any historical study for which observations are impossible. As always, the goal is to use this material to provide a richer understanding of the phenomena being examined.
Written documents are of two types, official records and personal documents (Lincoln & Guba, 1985). Official documents include government reports, licenses, contracts, diplomas, and so forth. Personal documents include diaries, letters, email, literature, field notes, and so forth. Artifacts include material and electronic traces, such as buildings, art, posters, nontextual computer files—essentially any disturbance of the natural environment created by people.
One type of existing material has assumed primacy in qualitative research in counseling: the counseling interview (see, for example, Elliott et al., 1994; Thompson & Jenal, 1994), either represented as a verbal record (that is, a tape recording) or as text (that is, a transcript). In such studies, interpretations are made of the records of counseling. For example, Friedlander, Heatherington, Johnson, and Showron (1994) examined family therapy sessions to under¬stand how sustaining engagement was important to change.
Of the throe sources of qualitative data discussed in this chapter (observations, interviews, and existing materials), existing materials present the most issues, even though such materials can be particularly valuable. The difficulty with existing materials involves the interpretation of the text or artifacts, particularly when the analyses of existing materials and use of other methods (e.g., observations or interviews) yield different or even opposing results. Hodder (1994) gives several examples of discrepancies between material traces and people's reports of their activities (e.g., report of alcohol use versus number of beer bottles found in garbage).

ANALYZING AND PRESENTING DATA
As is the case for the collection of data, the analysis and presentation of qualitative data depend on the particular qualitative strategy of inquiry utilized. Here we discuss three generic means to analyze data and present results: thick description, themes and relationships, and interpretation. For more detail, see Wolcott (1994), who presented a thorough discussion of these methods, including lengthy examples.

THICK DESCRIPTION
Thick description is the most basic means of presenting qualitative data. Essentially, a thick description is an unadulterated and thorough presentation of the data. The researcher may write some introductory and transitory material, but principally the presentation consists of lengthy excerpts from interviews, field notes, and existing materials (text and/or descriptions of artifacts).
The consumers of thick descriptions have the raw data, which for the most part have not been altered by the, investigator. In this way, "the data speak for themselves" and provide a rich account of what happened. Of course, the process of observing, interviewing, collecting materials, and deciding what to describe filters what is available to consumers. Still, even though the understanding achieved by readers of a thick description is affected by the research process, thick descriptions are closer to the phenomenon being studied than are any other means, qualitative or quantitative.
Despite the primary advantage that the data are minimally filtered, there are numerous disadvantages to using thick descriptions. True descriptions are too lengthy for journal articles, and many important thick descriptions languish as lengthy dissertations or unpublished reports. Another disadvantage is that the thick descriptions may be unfocused and uninteresting: "Readers are likely to get the idea that the researcher has been unable to sort out (or unwilling to throw away) data and has simply passed the task along. . . . [Djata that do not 'speak' to the person who gathered and reported them are not likely to strike up a conversation with subsequent readers either" (Wolcott, 1994, pp. 13-14). It is hoped that the story, as told by the participants or as observed by the investigator, will be fascinating enough to hold the interest of readers.
Rarely are qualitative investigators able to present all data verbatim in a thick description, and thus they must use various strategies to present a condensed but still relatively complete description. As discussed earlier in this chapter, the methods of condensing and presenting the data vary in accordance with the strategies of inquiry adopted. Regardless, the investigator has the critical task of deciding what must go, and there are few guidelines. Crucial decisions about data should be made through a process that acknowledges the investigators' suppositions. Moreover, the process involved in examining these suppositions and making sense of the data in the context of these suppositions should be a prominent part of the presentation so that readers can understand how the descriptions were distilled to a reasonable length.

THEMES AND RELATIONSHIPS
Themes are recurrent patterns in data that represent a concept, whereas relationships are the interconnections among the themes. Through development of themes and relationships, the essence of a phenomenon is revealed. Extracting themes and establishing relationships are the essence of the three strategies of inquiry discussed earlier (i.e., grounded theory, phenomenology, and CQR). Essentially, the relationships of themes constitute the theory or the structure of the phenomenon, which is grounded in (is inductively developed from) the data.
The first step in this process is to generate codes for various themes that are intrinsic to the phenomenon being studied. Often tentative codes are first generated from initial observations or field notes on a word-by-word, sentence by sentence, or paragraph-by-paragraph reading of the text, and are then written directly on the transcripts. These codes name various patterns. For example, in a grounded theory study of psychotherapy, Watson and Rennie (1994) developed the code "reflective self-examination" to describe clients' observations of themselves that were "spurred by the questions they posed about their behavior, their feelings, and their interactions with the world" (p. 503).
As the investigator progresses through the data, the explanatory power of the various codes becomes apparent. The codes are revised and refined until a clear theme is developed. When the theme is relatively robust, the investigator reexamines the data to understand how the theme functions with the phenomenon (e.g., what is the context needed to give rise to the phenomenon?). In the Watson and Rennie (1994) study, the authors answered these questions by presenting the complex relationship among themes:
The analysis of clients' subjective experiences during the exploration of problematic reactions provided access to the internal operations that clients engaged in to effect changes in their behavior. The model of the clients' experiences during the event highlights two important foci of clients' attention and activity during the session: client operations and session momentum. When engaged in the former, clients alternated between two primary activities: symbolic representation of their experience and reflective self-examination.... These two processes, together with therapists' operations, are related to clients' making new realizations and engaging in activities to alter their behavior. (p. 506)

The model of themes and their relationships provides an emerging grounded theory. As the model is developed, the investigator returns to existing data or collects additional data to test crucial aspects of the model. In their study of chemistry groups, Wampold et al. (1995) had a tentative model that described the chemists' avoidance of social interaction involving encoding and decoding emotion. Given that strong affect is intrinsic to conflictual situations, they returned to observe the chemists during conflict in order to under¬stand more fully the interplay between the chemists' expression and understanding of emotion and their social interactions. The point is that the themes and relationships model should always be considered emergent and should guide further collection of data.

INTERPRETATION
Interpretation, as described by Wolcott (1994), is aimed at extracting meaning and identifying context. Rather than providing a theory of a specific phenomenon by relating themes, interpretation addresses more global issues, such as "What is the role of race and ethnicity in American society?" and "How do cultural factors affect conceptions of mental health and treatment?" Wolcott noted that "at the interpretive extreme, a researcher-as-writer may seem merely to swoop down into the field for a descriptive morsel or two and then retreat once again to the lofty heights of theory or speculation" (p. 11) and that interpretation "is well suited to mark a threshold in thinking and writing at which the researcher transcends tactual data and cautious analyses and begins to probe into what is to be made of them" (p. 36). Clearly, novice researchers need to acquire more training and experiences in order to master these types of higher level skills.


TRUSTWORTHINESS OF QUALITATIVE RESEARCH
Finally, we will briefly discuss the criteria used by qualitative researchers to evaluate the rigor or trustworthiness of a qualitative research design. Some researchers have argued for alternative ways to evaluate the rigor of data collection and analytic procedure in qualitative research (Pidgeon, 1996). For example, compared to the criteria used by quantitative researchers for evaluating a quantitative study (e.g., internal validity, external validity/generalizability, reliability, and objectivity), a different set of criteria (i.e., credibility, transferability, dependability, and confirmability) were proposed by Lincoln and Guba (1985) to evaluate the scientific worth of qualitative research. Many qualitative researchers have since followed Lincoln and Guba's recommendations. Several guidelines also have been published by other qualitative researchers (see Cobb & Hagemaster, 1987; Elder & Miller, 1995; Elliott, Fischer, & Rennie, 1999; Morrow, 2005). Interested readers should refer to these publications for more details.


SUMMARY AND CONCLUSIONS
In this chapter we discussed the definition and understanding analysis of qualitative research. The key myths and facts regarding qualitative methodology were also explored. Occasionally, we have encountered students who have undertaken qualitative research under the misguided notion that qualitative research is easier or less technical than quantitative research. Rigorous qualitative research certainly is different from quantitative research, but selecting this approach will lead to disappointment if the primary motivation is the avoidance of work or numbers. Qualitative research is time consuming and thought provoking. Be prepared to stay up late thinking about the meaning of field notes or transcripts, attempting to make sense of intrinsically ambiguous data, and writing, writing, writing.
This chapter has only scratched the surface of qualitative research. Volumes have been written about this subject, and one could devote all of one's energy to learning about only one particular strategy of qualitative inquiry. We have sought to provide the reader an appreciation of qualitative research, an overview of the various strategies of inquiry, general methods, guidelines for evaluating the rigor of qualitative research design, and references for those who want to pursue the topic in greater depth (also see Ponterotto, 2005a, for books and journals with a qualitative research focus). In addition, helpful examples of theses and dissertations that utilized various strategies of inquiry can be found in Wang, Heppner, and Heppner (2004a, 20046).
Finally, we recommend that students seek out ongoing qualitative research teams and volunteer to participate. As suggested by McLeod (2001), we believe that the personal qualities of qualitative researchers (e.g., integrity, perseverance, and willingness and ability to struggle with ambiguity), rather than the methodology per se, have a greater impact on the potential contribution of the studies. The experiences encountered when conducting a qualitative research apprenticeship with veteran qualitative researchers not only can hone a person's methodological skills, but also can help to facilitate the development of these notable personal qualities.