Chapter Ten Interview techniques and the analysis of interview data
An interview can be thought of as a dialogue or conversation between interviewers and research participants with the purpose of eliciting information from the participants. Interviews are a key tool for the clinician and the health researcher as a means of collecting information. Interviews may vary substantially in their structure, content and the way in which they are conducted. This chapter is concerned with interviews and the analysis of interview data.
Many researchers distinguish between structured and unstructured interviews. Sometimes the terms ‘formal’ and ‘informal’ or ‘guided’ and ‘open-ended’ are also used (see, for example, Morse & Field 2003).
Denzin & Lincoln (1994) distinguished between three forms of the interview: the schedule standardized interview in which the wording and order of all questions are exactly the same for every respondent; the non-schedule standardized interview where certain types of information are desired from all respondents but the particular phrasing of questions and their order are redefined; and the non-standardized interview in which no prescribed set of questions is employed.
If one defines an interview as a conversation, then a schedule standardized interview is a very rigid form of conversation, almost like a play with a fixed script. In its most structured form, a structured interview may involve the reading of a prepared questionnaire to respondents and then filling in an answer form or response sheet for them on the basis of their answers. The questions are provided in a systematic order, with minimal or no deviation from the prepared script. In a structured interview, the role of the interviewer is to ask the questions and the role of the respondent is to provide the answers with minimal extraneous information. Conversely, an unstructured interview may involve the interviewer in asking no direct questions, but simply prompting respondents to reflect on their current interests and concerns. Clearly, between these extremes lies a variety of different types of interview strategies and degrees of structure. The extent of ‘structure’ or ‘formality’ is determined by a number of factors, including the following:
It is useful to consider some of the advantages and disadvantages of the different types of interview approaches. These are summarized in Table 10.1.
Table 10.1 Advantages and disadvantages of structured and unstructured interviews
| Advantages | Disadvantages | |
|---|---|---|
| Structured interviews | Responses may not be recorded in the respondents’ own words | |
| Unstructured interviews |
The appropriateness of the different interview approaches is determined by the objectives of the researcher. If the researcher simply wished to collect some basic symptom data, an unstructured interview would be inefficient. However, if the researcher wished to study people’s conceptualization and interpretation of their illness, an unstructured interview may be quite suitable.
Some clinical interviews, such as history taking, are highly structured whereas other clinical interviews, such as those involving management of a long-term problem, may be less so.
Since an interview is a conversation, there exist several possible methods for conducting it. The interview may be conducted in person (‘face to face’) or by remote means such as by telephone. There are a number of advantages associated with face-to-face interviews. (These are also discussed in Ch. 9.) Face-to-face interviews permit the nonverbal reactions of the respondent to be observed and perhaps the development of a closer rapport arising from the more ‘natural’ setting. Interviewers may use their observations of non-verbal cues to supplement the verbal information being provided and use their own non-verbal cues in a similar fashion. However, the face-to-face interview may require a substantial amount of participant travel time and hence higher costs than for a telephone interview. With certain interview objectives, however, telephone interviews may not be suitable. If the interviewer and his or her credentials are not well known to the interviewee, it is unlikely that participants in a telephone interview would provide valid and reliable personal information about personal topics. Some people find disclosure of sensitive information to be easier by telephone or by the use of anonymous self-completion questionnaires. The face-to-face interview may be too confrontational or embarrassing for them.
Interviewers may use a number of different means of recording interview information, ranging from written summary notes of the interview to an actual video- or audiotaping of the live interview. These recording methods have a number of advantages and disadvantages, as discussed below.
The advantages and disadvantages of the various recording methods are summarized in Table 10.2. Thus, video recording is intrusive, requires substantial post-interview analysis and may result in less disclosure, yet provides very rich information that can be independently analysed. The use of a recording method such as a response sheet requires great trust in the judgment and recording abilities of the interviewer. There is a potential for bias arising from the interviewer ‘adjusting’ the information provided by the interviewee to fit the recording method and/or expectations of the interviewer. Further, there is no opportunity for re-analysis.
Table 10.2 Advantages and disadvantages of different ways of recording interview information
| Advantages | Disadvantages | |
|---|---|---|
| Video recording | ||
| Audio recording | ||
| Response sheets | ||
| Unstructured notes | Cheap and simple |
So which information-recording method for interviews is the most appropriate? The appropriateness of the method of recording interview information is determined by the needs of the person using the information. For example, if the information user simply wants some basic data such as the age, sex and symptom profile of a patient, it would be absurd to use video recording. This would be very time-consuming. Each tape would have to be made, then viewed again for analysis. In this instance, a simple response sheet or checklist would suffice. However, if the interviewer was interested in exploring reactions of interviewees to the death of a close relative, perhaps the use of audio or video recording would provide a richness of data suitable for that interest.
Although most interviews are conducted with one interviewer and one interviewee present, sometimes group interviews with many participants are conducted. The focus group, which is a form of group interview, involves a discussion among a small group of people, including a moderator or facilitator (Thomas et al 1992). The facilitator’s role is to introduce the topics or questions for discussion and to facilitate the contributions of the group participants. As a matter of historical interest, Merton (1946) originally proposed the focused interview, which was the forerunner for the focus group.
Focus groups differ fundamentally from the individual interview in that the researcher is outnumbered and the participants may interact with each other, modify each other’s responses and ask questions of each other. The researcher is less in control of the process than in other types of interview approaches. Focus groups often bring to the surface matters that would not have been raised in conventional interviews. Focus groups are now widely used in health research because they provide rich sources of insights and interpretations from the participants.
The manner in which interview data may be analysed is determined in part by how the data have been recorded and, in part, by the theoretical orientation of the researcher. In the previous discussion, we have seen that these formats include videotapes, audiotapes, completed response sheets and free-form summary interviewer notes.
The basis for many analyses of interview data is the interview transcript. The transcript of an interview is a verbatim written version of the conversation that took place between the participants. To provide an example, an excerpt from an interview transcript produced by Janet Doyle at La Trobe University follows (Doyle & Thomas 2000). The transcript is of an interview between a clinician and a client, concerning the client’s hearing loss.
And so the interview continued.
Let us now consider some of the analysis options under the quantitative and qualitative headings.
A number of quantitative analyses possibilities are presented with interview transcripts. For example, the researcher might count the number of words spoken by each participant to obtain a quantitative measure of their relative contributions to the conversational process. Another possibility would be to count the number of questions asked by the clinician. These quantitative measures could then be used to test various hypotheses. Analyses of interview transcripts similar to the example shown above, in the study from which it was taken, have demonstrated substantial sex differences between the number of questions asked by male and female clients. It seems that the male clients asked many more questions than the females. Thus, the quantitative researcher might use interview transcripts to count and analyse certain features of the transcripts.
Under the qualitative heading there is a broad variety of approaches to the collection and analy-sis of interview data. Such approaches may, however, be broadly categorized as descriptive or theoretical.
A descriptive qualitative study is often termed an ethnography. They are often written from the perspective of the participant(s) in the first person. The purpose of the ethnography is to provide a detailed description of a particular set of circumstances and to encourage readers to make their own interpretations. A celebrated example of such an ethnography is found in Bogdan & Taylor’s (1976) description of Ed Murphy’s life. Ed Murphy was a former resident of a home for the intellectually disabled in the USA.
Many qualitative studies, however, are theoretical in nature. That is, they attempt to develop theories and concepts and, often, to verify these concepts and theories.
A key approach to theoretical qualitative research is provided by Glaser & Strauss’ (1967) grounded theory (see also Strauss 1987, Strauss & Corbin 1998). Glaser & Strauss advocate two methods for the development of grounded theory: the constant comparative method in which the researcher codes and analyses data to develop concepts, and the theoretical sampling method in which cases are selected purposively to refine the ‘theory’ previously developed. Glaser & Strauss provide highly detailed examples of their analytic methods in the above references and the interested reader is referred to them for further detail. Glaser & Strauss’ approach is not universally accepted in that some qualitative researchers argue for the necessity both to develop and verify their theories. Notwithstanding these theoretical differences, many qualitative researchers share common analysis tools such as coding and thematic analysis.
Coding is used to organize data collected in an interview and, for that matter, in other types of documents such as field notes. Different qualitative researchers advocate different approaches to coding but it typically involves the following steps. The researchers closely study their materials, in this case the interview transcripts, and develop a close familiarity with the material. During this process, all the concepts, themes and ideas are noted to form major categories. For example, in interviews of nursing home residents, some themes that arise might include personal safety, autonomy and decision making, personal hygiene and so on. Often, the researcher will then attach a number or label to each category and record their positions in the transcript. Coding is an iterative process, with the researcher coding and recoding, as the scheme develops. Some computer programs are now available to assist with the coding analysis of machine-readable transcripts and these ease some of the clerical burden, although most qualitative researchers still employ manual coding methods. The researchers, having developed the codes and coded the transcripts, then attempt to interpret their meanings in the context in which they appeared. The reporting of this process typically involves ‘thick’ or detailed description of the categories and their context, with liberal use of examples from the original transcripts. The process of analysis of qualitative research materials is examined in more detail later in this book.
Interviews may be defined as a conversation between interviewers and interviewees with the purpose of eliciting certain information. Structured interviews generally involve a fixed set of questions or schedule, the use of preplanned response sheets, a greater proportion of closed-response questions and direction from the interviewer. Unstructured interviews tend not to have these attributes, with less structure and control. Interviews may be conducted face to face or by telephone and both these methods have certain advantages and disadvantages. The focus group is a valuable alternative to the individual interview. Interview data may be recorded in a number of different ways, but the transcript is often used. Transcripts may be analysed using quantitative or qualitative techniques.
Explain the meaning of the following terms: