1748390 – Wiley US ©

derator or interviewer of the group be familiar with group processes and with the range ofpossible roles as moderator (Barbour, 2008; Hennink, 2014; Krueger & Casey, 2015; Stewart& Shamdasani, 2015).

Finally, “focus groups work best for topics people could talk about to each other in theireveryday lives—but don't” (Macnaghten & Myers, 2004, p. 65). Obviously, a focus group isa poor choice for topics that are sensitive, highly personal, and culturally inappropriate totalk about in the presence of strangers. Of course, it's not always obvious ahead of timehow appropriate a topic might be. Crowe (2003) reports successful use of focus groups tocreate culturally appropriate HIV prevention material for the deaf community. Jowett andO'Toole (2006) report an interesting analysis of two focus groups—one of mature studentsand their attitude toward participation in higher education, and one of young women andtheir views of feminism. They found that the mature students' focus group was a failurebut the young women's group was a success. The authors had not anticipated “howingrained the sense of inadequacy is for some people who have felt excluded fromeducation” (p. 462), nor how the power imbalance among members of the maturestudents' group and between the researcher and the group inhibited participation. Finally,Stewart and Williams (2005) explore the practical and ethical issues of conductingsynchronous and asynchronous online focus groups.

Thus, as with any other data collection method, focus groups are appropriate to be usedwhen this is the best way to get the best data that addresses your research question. Andas with any other method, the advantages need to be weighed against the disadvantages;one also needs to develop the skills necessary for using this technique.

Online InterviewsThere is no question that the Internet has changed the world. It has also increased thepossibilities for the myriad ways that one can collect data through online venues inconducting qualitative research through various information communication technologies(ICTs) and computer mediated communication (CMC) tools (Salmons, 2015). Qualitativedata are collected from or through email, blogs, online discussion groups, Skype, tweets,texts, and various forms of social media. Here we discuss issues in conducting onlineinterviews.

One can conduct online interviews synchronously (in real time) through various CMC toolssuch as Skype or Adobe Connect. These are typically verbal interviews with a videocomponent that are more like face-to-face interviews; one can also conduct voice-to-voicereal-time interviews over the telephone. One can also conduct interviews asynchronously(where there is a lag time) over email or on an online discussion group; asynchronousinterviews tend to be text-based or written interviews. There are strengths and weaknessesto both synchronous and asynchronous venues. As will be discussed in more detail later,in general it is helpful to build rapport with participants when conducting qualitativeinterviews. Rapport building can be slightly more challenging in text-only asynchronousvenues (such as email), when visual cues are missing (James & Busher, 2012). Further,participants may not respond to email queries or not respond to certain questions over

1748390 – Wiley US ©

email that they would likely answer in synchronous video or voice-to-voice formats. On theother hand, text-based interviews over email provide the researcher with a ready-madetranscript, making it easy to document what was said, though the nonverbal cues andpauses in conversation are missing. Such an email “transcript” can save the researchertime and money in transcription costs.

Given the availability of various information and communications technology (ICT) toolsfor conducting online interviews in either synchronous or asynchronous formats, Salmons(2015), in her book on online interviews, presents a framework for considering what sherefers to as “e-interview research” (p. 4). She invites the researcher to explore key questionsin eight interrelated categories: (1) aligning the purpose of the research with the design;considering issues related to (2) choosing data collection methods and (3) one's positionas a researcher; determining (4) the e-interview style, (5) the type of ICT tools to use, (6)sampling issues, (7) ethical issues, and (8) actually collecting the data. While qualitativeresearchers always need to consider similar issues in all qualitative studies, Salmonsproposes questions and issues specifically related to the online environment.

There is a growing discussion of the availability of various ICT tools for conducting onlineinterviews, many of which are reviewed by Salmons (2015) and others, talking mainlyabout individual interviews. Tuttas (2015) focuses more specifically on lessons learnedfrom using real-time audiovisual web conferencing technology to carry out focus groupinterviews with nurses from geographically dispersed locations in the United States. Whileshe ultimately chose to use Adobe Connect, she considers the strengths and weaknessesof various web conferencing technology venues (Skype, ooVoo, GoToMeeting, and AdobeConnect) for her focus group interviews, which can provide some guidance for some ofthe available options.

Like any data collection method, conducting online interviews has its strengths andweaknesses. One of the obvious strengths is that the researcher is no longer constrainedby geography in considering participants. A researcher could interview participants acrossthe world, and could perhaps even conduct focus group interviews where all parties cansee each other. Another strength is that many CMC venues allow video recordings to bemade, which can be helpful if one wants to explore or review nonverbal cues later. Someobvious weaknesses are that not everyone has access to various CMC tools or theknowledge of how to use them. Further, technology is always subject to breakdowns.There can be problems with audio recording equipment, as voices sometimes break up oncell phones or over Skype or other computer mediated venues, which can cause frustrationfor both the interviewer and interviewee. Finally, there is always the chance ofconfidentiality being compromised when one uses CMC tools over the Internet. While thismay be unlikely, it is always a consideration for researchers and institutional reviewboards in dealing with ethical issues in doing research. In sum, all of the strengths andweaknesses of CMC tools in relationship to qualitative interviewing need to be consideredwhen undertaking a qualitative research study.

Asking Good QuestionsThe key to getting good data from interviewing is to ask good questions; asking goodquestions takes practice. Pilot interviews are crucial for trying out your questions. Not onlydo you get some practice in interviewing, but you also quickly learn which questions are

1748390 – Wiley US ©

confusing and need rewording, which questions yield useless data, and which questions,suggested by your respondents, you should have thought to include in the first place.

Different types of questions will yield different information. The questions you ask dependupon the focus of your study. Using the example of mentoring in the career developmentof master teachers, if you wanted to know the role mentoring played in careerdevelopment, you would ask questions about teachers' personal experience withmentoring and probably get a descriptive history. Follow-up questions about how they feltabout a certain mentoring experience would elicit information that is more affective innature. You might also want to know their opinion as to how much influence mentoringgenerally has in a teacher's career.

The way in which questions are worded is a crucial consideration in extracting the type ofinformation desired. An obvious place to begin is by making certain that what is beingasked is clear to the person being interviewed. Questions need to be couched in familiarlanguage. “Using words that make sense to the interviewee, words that reflect therespondent's world view, will improve the quality of data obtained during the interview.Without sensitivity to the impact of particular words on the person being interviewed, theanswer may make no sense at all—or there may be no answer” (Patton, 2015, p. 454).Avoiding technical jargon and terms and concepts from your particular disciplinaryorientation is a good place to begin. In a study of HIV-positive young adults, for example,participants were asked how they made sense of or came to terms with their diagnosis, nothow they constructed meaning in the process of perspective transformation (thetheoretical framework of the study) (Courtenay, Merriam, & Reeves, 1998).

Types of Questions, Good Questions, and Questions to AvoidAn interviewer can ask several types of questions to stimulate responses from aninterviewee. Patton (2015) suggests six types of questions:

1. Experience and behavior questions —This type of question gets at the things a persondoes or did, his or her behaviors, actions, and activities. For example, in a study ofleadership exhibited by administrators, one could ask, “Tell me about a typical day atwork; what are you likely to do first thing in the morning?”

2. Opinion and values questions —Here the researcher is interested in a person's beliefs oropinions, what he or she thinks about something. Following the preceding example ofa study of administrators and leadership, one could ask, “What is your opinion as towhether administrators should also be leaders?”

3. Feeling questions —These questions “tap the affective dimension of human life. Inasking feeling questions—‘how do you feel about that?’—the interviewer is looking foradjective responses: anxious, happy, afraid, intimidated, confident, and so on” (p. 444).

4. Knowledge questions —These questions elicit a participant's actual factual knowledgeabout a situation.

5. Sensory questions —These are similar to experience and behavior questions but try toelicit more specific data about what is or was seen, heard, touched, and so forth.

�. Background/demographic questions —All interviews contain questions that refer to theparticular demographics (age, income, education, number of years on the job, and so

1748390 – Wiley US ©

on) of the person being interviewed as relevant to the research study. For example, theage of the respondent may or may not be relevant.

Interestingly, Patton (2015) recommends against asking “why” questions because theytend to lead to speculation about causal relationships and they can lead to dead-endresponses. Patton recounts an amusing interview with a child in a study of openclassrooms. When a first grader responded that her “favorite time in school” was recess,Patton asked her why she liked recess. Her answer was because she could go outside andplay on the swings. When he asked, “why” she went outside, the child responded, “Becausethat's where the swings are!” (p. 456). Although “why” questions can put an end to a lineof questioning, it has been our experience that an occasional “why” question can uncoverinsights that might be speculative but that might also suggest a new line of questioning.

Another typology of different types of questions that we have found particularly useful ineliciting information, especially from reticent interviewees, is Strauss, Schatzman, Bucher,and Sabshin's (1981) four major categories of questions: hypothetical, devil's advocate,ideal position, and interpretive questions. Each is defined in Table 5.2 and illustrated withexamples from a case study of displaced workers participating in a job training andpartnership ( JTPA) program.

TABLE 5.2

Four Types of Questions with Examples from a JTPA Training Program Case Study.Type of Question Example1. Hypothetical questions—Ask what therespondent might do, or what it might be like in aparticular situation; these usually begin with“what if” or “suppose.”

Suppose it were my first day in thistraining program. What would it belike?

2. Devil's advocate questions—The respondent ischallenged to consider an opposing view orexplanation to a situation.

Some people would say thatemployees who lost their job didsomething to bring about being fired.What would you tell them?

3. Ideal position questions—Ask the respondentto describe an ideal situation.

Would you describe what you think theideal training program would be like?

4. Interpretive questions—The researcheradvances tentative explanations orinterpretations of what the respondent has beensaying and asks for a reaction.

Are you finding returning to school asan adult a different experience fromwhat you expected?

Hypothetical questions ask respondents to speculate as to what something might be likeor what someone might do in a particular situation. Hypothetical questions begin with“What if” or “Suppose.” Responses are usually descriptions of the person's actualexperience. In the JTPA study, for example, the hypothetical question, “Suppose it were myfirst day in this training program—what would it be like?” elicited descriptions of what itwas actually like for the participants.

1748390 – Wiley US ©

Devil's advocate questions are particularly good to use when the topic is controversial andyou want respondents' opinions and feelings. This type of question also avoidsembarrassing or antagonizing respondents if they happen to be sensitive about the issue.The wording begins, “Some people would say,” which in effect depersonalizes the issue.The response, however, is almost always the respondent's personal opinion or feelingabout the matter. In the JTPA example, the question, “Some people would say thatemployees who lost their job did something to bring it about. What would you say tothem?” usually revealed how the respondent came to be unemployed and thus involved inthe training program.

Ideal position questions elicit both information and opinion; these can be used withvirtually any phenomenon under study. They are good to use in evaluation studiesbecause they reveal both the positives and the negatives or shortcomings of a program.Asking what the ideal training program would be like in the JTPA example revealed thingsparticipants liked and would not want changed, as well as things that could have made ita better program.

Interpretive questions provide a check on what you think you are understanding, as well asoffer an opportunity for yet more information, opinions, and feelings to be revealed. In theJTPA example, the interpretive question, “Would you say that returning to school as anadult is different from what you expected?” allowed the investigator to confirm thetentative interpretation of what had been said in the interview.

Overall, good interview questions are those that are open-ended and yield descriptive data,even stories about the phenomenon. The more detailed and descriptive the data, the better.The following questions work well to yield this type of data:

Tell me about a time when…

Give me an example of…

Tell me more about that…

What was it like for you when…

Some types of questions should be avoided in an interview. Table 5.3 outlines three typesof questions to avoid and illustrates each from the JTPA study. First, avoid multiplequestions—either one question that is actually a multiple question or a series of singlequestions that does not allow the respondent to answer one by one. An example of amultiple question is, “How do you feel about the instructors, the assignments, and theschedule of classes in the JTPA training program?” A series of questions might be,“What's it like going back to school as an adult? How do instructors respond to you? Whatkind of assignments do you have?” In both cases the respondent is likely to ask you torepeat the question(s), ask for clarification, or give a response covering only one part ofthe question—and that response may be uninterpretable. If, for example, an intervieweeresponded to the question, “How do you feel about the instructors, the assignments, andthe schedule of classes?” with “They're OK—some I like, some I don't,” you would not knowwhether instructors or assignments or the schedule was being referred to.

1748390 – Wiley US ©

TABLE 5.3

Questions to Avoid.Type ofQuestion

Example

Multiplequestions

How do you feel about the instructors, the assignments, and theschedule of classes?

Leadingquestions

What emotional problems have you had since losing your job?

Yes-or-noquestions

Do you like the program? Has returning to school been difficult?

Leading questions should also be avoided. Leading questions reveal a bias or anassumption that the researcher is making, which may not be held by the participant. Theseset the respondent up to accept the researcher's point of view. The question, “Whatemotional problems have you had since losing your job?” reflects an assumption thatanyone losing a job will have emotional problems.

Finally, all researchers warn against asking yes-or-no questions. Any question that can beanswered with a simple yes or no may in fact be answered just that way. Yes-or-noresponses give you almost no information. For the reluctant, shy, or less verbalrespondent, they offer an easy way out; they can also shut down or at least slow the flowof information from the interviewee. In the JTPA example, questions phrased in a yes-or-no manner, although at their core they are seeking good information, can yield nothing.Thus asking, “Do you like the program?” may be answered yes or no; rephrasing it to,“What do you like about the program?” necessitates more of a response. The same is trueof the question “Has returning to school been difficult?” Asking, “How have you found theexperience of returning to school?” mandates a fuller response.

A ruthless review of your questions to weed out poor ones before you actually conduct aninterview is highly recommended. Ask the questions of yourself, challenging yourself toanswer as minimally as possible. Also note whether you would feel uncomfortablehonestly answering any of the questions. This review, followed by a pilot interview, will goa long way to ensure that you are asking good questions.

ProbesProbes are also questions or comments that follow up on something already asked. It isvirtually impossible to specify these ahead of time because they are dependent on howthe participant answers the lead question. This is where being the primary instrument ofdata collection has its advantages, especially if you are a highly sensitive instrument. Youmake adjustments in your interviewing as you go along. You sense that the respondent isonto something significant or that there is more to be learned. Probing can come in theform of asking for more details, for clarification, for examples. Glesne and Peshkin (1992)point out that “probes may take numerous forms; they range from silence, to sounds, to asingle word, to complete sentences” (p. 85). Silence, “used judiciously…is a useful andeasy probe—as is the bunched utterance, ‘uh huh, uh huh,’ sometimes combined with anodding head. ‘Yes, yes’ is a good alternative; variety is useful” (p. 86, emphasis in

1748390 – Wiley US ©

original). As with all questions, not just probes, the interviewer should avoid pressing toohard and too fast. After all, the participant is being interviewed, not interrogated.

Probes or follow-up questions—or as Seidman (2013) prefers to call it, “exploration”—canbe as simple as seeking more information or clarity about what the person has just said.These are typically who, what, when, and where questions, such as Who else was there?What did you do then? When did this happen? or Where were you when this happened?Other probes seek more details or elaboration, such as What do you mean? Tell me moreabout that. Give me an example of that. “Walk” me through the experience. Would youexplain that? and so on.

The following is a short excerpt (Weeks, n.d.) from an interview with a man in midlife whohad been retained (held back a grade) in grammar school. The investigator was interestedin how being retained had affected the person's life. Note the follow-up questions orprobes used to garner a better understanding of his initial reaction to being retained.

Interviewer: How did you feel about yourself the second time you were in first grade?Respondent: I really don't remember, but I think I didn't like it. It was probably

embarrassing to me. I think I may have even had a hard time explaining it tomy friends. I probably got teased. I was probably defensive about it. I mayeven have rebelled in some childlike way. I do know I got more aggressive atthis point in my life. But I don't know if being retained had anything to dowith it.

Interviewer: How did you feel about your new first grade teacher?Respondent: She was nice. I was very quiet for a while, until I got to know her.Interviewer: How did you feel about yourself during this second year?Respondent: I have to look at it as a follow-up to a period when I was not successful.

Strictly speaking, I was not very successful in the first grade—the first time.Interviewer: Your voice sometimes changes when you talk about that.Respondent: Well, I guess I'm still a little angry.Interviewer: Do you feel the retention was justified?Respondent: (long pause) I don't know how to answer that.Interviewer: Do you want to think about it for a while?Respondent: Well, I did not learn anything in the first grade the first time, but the lady was

nice. She was my Mom's best friend. So she didn't teach me anything, andshe made me repeat. I had to be retained, they said, because I did not learnthe material, but (shaking his finger), I could have. I could have learned itwell. I was smart.

The best way to increase your skill at probing is to practice. The more you interview,especially on the same topic, the more relaxed you become and the better you can pursuepotentially fruitful lines of inquiry. Another good strategy is to scrutinize a verbatimtranscript of one of your interviews. Look for places where you could have followed up butdid not, and compare them with places where you got a lot of good data. The differencewill most likely be from having maximized an opportunity to gain more informationthrough gentle probing.

1748390 – Wiley US ©

The Interview GuideThe interview guide, or schedule as it is sometimes called, is nothing more than a list ofquestions you intend to ask in an interview. Depending on how structured the interview willbe, the guide may contain dozens of very specific questions listed in a particular order(highly structured) or a few topical areas jotted down in no particular order (unstructured)or something in between. As we noted earlier, most interviews in qualitative research aresemistructured; thus the interview guide will probably contain several specific questionsthat you want to ask everyone, some more open-ended questions that could be followedup with probes, and perhaps a list of some areas, topics, and issues that you want to knowmore about but do not have enough information about at the outset of your study to formspecific questions.

An investigator new to collecting data through interviews will feel more confident with astructured interview format in which most, if not all, questions are written out ahead oftime in the interview guide. Working from an interview schedule allows the new researcherto gain the experience and confidence needed to conduct more open-ended questioning.Most researchers find that they are highly dependent upon the interview guide for the firstfew interviews but soon can unhook themselves from constant reference to the questionsand go with the natural flow of the interview. At that point, an occasional check to seewhether all areas or topics are being covered may be all that is needed.

New researchers are often concerned about the order of questions in an interview. No rulesdetermine what should go first and what should come later. Much depends upon thestudy's objectives, the time allotted for the interview, the person being interviewed, andhow sensitive some of the questions are. Factual, sociodemographic-type questions canbe asked to get the interview started, but if there are a lot of these, or if some of them aresensitive (for example, if they ask about income, age, or sexual orientation), it might bebetter to ask them at the end of the inter-view. By then the respondent has becomeinvested in the interview and is more likely to see it through by answering these questions.

Generally it is a good idea to ask for relatively neutral, descriptive information at thebeginning of an interview. Respondents can be asked to provide basic descriptiveinformation about the phenomenon of interest, be it a program, activity, or experience, or tochronicle their history with the phenomenon of interest. This information lays thefoundation for questions that access the interviewee's perceptions, opinions, values,emotions, and so on.

Of course, it is not always possible to separate factual information from more subjective,value-laden responses. And again, the best way to tell whether the order of your questionsworks or not is to try it out in a pilot interview.

In summary, then, questions are at the heart of interviewing, and to collect meaningfuldata a researcher must ask good questions. In our years of experience doing andsupervising qualitative research, the fewer, more open-ended your questions are, the better.Having fewer broader questions unhooks you from the interview guide and enables you toreally listen to what your participant has to share, which in turn enables you to betterfollow avenues of inquiry that will yield potentially rich contributions. Exhibit 5.1 is aninterview guide for a study of how older adults become self-directed in their health care(Valente, 2005). These open-ended questions, followed up by the skillful use of probes,yielded substantive information about the topic.

1748390 – Wiley US ©

Exhibit 5.1. Interview Guide.

1. I understand that you are concerned about your health. Tell me about your health.

2. What motivated you to learn about your health?

3. Tell me, in detail, about the kinds of things you have done to learn more aboutyour health. (What did you do first?)

4. Where do you find information about your health?

5. Tell me about a time when something you learned had a positive impact on yourhealth care.

�. What kinds of things have you changed in your life because of your learning?

7. Whom do you talk to about your health?

�. Tell me about your current interactions with your health care provider.

9. Tell me about what you do to keep track of your health.

10. What other things do you do to manage your health?

11. What kinds of challenges (barriers) do you experience when managing your healthcare?

12. What else would you like to share about your health-related learning?Source: Valente (2005). Reprinted with permission.

Beginning the InterviewCollecting data through interviews involves, first of all, determining whom to interview.That depends on what the investigator wants to know and from whose perspective theinformation is desired. Selecting respondents on the basis of what they can contribute tothe researcher's understanding of the phenomenon under study means engaging inpurposive or theoretical sampling (discussed in Chapter Four). In a qualitative case studyof a community school program, for example, a holistic picture of the program wouldinvolve the experiences and perceptions of people having different associations with theprogram—administrators, teachers, students, community residents. Unlike survey research,in which the number and representativeness of the sample are major considerations, inthis type of research the crucial factor is not the number of respondents but the potentialof each person to contribute to the development of insight and understanding of thephenomenon.

How can such people be identified? One way is through initial on-site observation of theprogram, activity, or phenomenon under study. On-site observations often involve informaldiscussions with participants to discover those who should be interviewed in depth. Asecond means of locating contacts is to begin with a key person who is consideredknowledgeable by others and then ask that person for referrals. Initial informants can befound through the investigator's own personal contacts, community and privateorganizations, advertisements on bulletin boards, or on the Internet. In some studies apreliminary interview is necessary to determine whether the person meets the criteria for

1748390 – Wiley US ©

participating in the study. For example, in Moon's (2011) study of the transformationalpotential of grieving in older adulthood, he first had to determine if prospectiveparticipants could identify some change in their sense of self or view of the world as aresult of grieving the loss of a loved one.

Taylor and Bogdan (1984) list five issues that should be addressed at the outset of everyinterview:

1. The investigator's motives and intentions and the inquiry's purpose

2. The protection of respondents through the use of pseudonyms

3. Deciding who has final say over the study's content

4. Payment (if any)

5. Logistics with regard to time, place, and number of interviews to be scheduled (pp.87–88)

Besides being careful to word questions in language clear to the respondent, theinterviewer must be aware of his or her stance toward the interviewee. Since therespondent has been selected by the investigator on purpose, it can be assumed that theparticipant has something to contribute, has had an experience worth talking about, andhas an opinion of interest to the researcher. This stance will go a long way in making therespondent comfortable and forthcoming with what he or she has to offer.

An interviewer should also assume neutrality with regard to the respondent's knowledge;that is, regardless of how antithetical to the interviewer's beliefs or values the respondent'sposition might be, it is crucial for the success of the interview to avoid arguing, debating,or otherwise letting personal views be known. Patton (2015) distinguishes betweenneutrality and rapport. “At the same time that I am neutral with regard to the content ofwhat is being said to me, I care very much that that person is willing to share with me whatthey are saying. Rapport is a stance vis-à-vis the person being interviewed. Neutrality is astance vis-à-vis the content of what that person says” (p. 457, emphasis in original).

There are several ways of maximizing the time spent getting an informant to shareinformation. A slow-starting interview, for example, can be moved along by askingrespondents for basic descriptive information about themselves, the event, or thephenomenon under study. Interviews aimed at constructing life histories can beaugmented by written narratives, personal documents, and daily activity logs thatinformants are asked to submit ahead of time. The value of an interview, of course,depends on the interviewer's knowing enough about the topic to ask meaningful questionsin language easily understood by the informant.

Interviewer and Respondent InteractionThe interaction between interviewer and respondent can be looked at from the perspectiveof either party or from the interaction itself. Skilled interviewers can do much to bringabout positive interaction. Being respectful, nonjudgmental, and nonthreatening is abeginning. Obviously, becoming skilled takes practice; practice combined with feedbackon performance is the best way to develop the needed skills. Role playing, peer critiquing,

1748390 – Wiley US ©

videotaping, and observing experienced interviewers at work are all ways noviceresearchers can improve their performance in this regard.

What makes a good respondent? Anthropologists and sociologists speak of a goodrespondent as an “informant”—one who understands the culture but is also able to reflecton it and articulate for the researcher what is going on. Key informants are able, to someextent, to adopt the stance of the investigator, thus becoming a valuable guide inunfamiliar territory. But not all good respondents can be considered key informants in thesense that anthropologists use the term. Good respondents are those who can expressthoughts, feelings, opinions—that is, offer a perspective—on the topic being studied.Participants usually enjoy sharing their expertise with an interested and sympatheticlistener. For some, it is also an opportunity to clarify their own thoughts and experiences.

Dexter (1970) says there are three variables in every interview situation that determine thenature of the interaction: “(1) the personality and skill of the interviewer, (2) the attitudesand orientation of the interviewee, and (3) the definition of both (and often by significantothers) of the situation” (p. 24). These factors also determine the type of informationobtained from an interview. Let us suppose, for example, that two researchers are studyingan innovative curriculum for first-year college students. One interviewer is predisposed toinnovative practices in general, while the other favors traditional educational practices.One student informant is assigned to the program, while another student requests thecurriculum and is eager to be interviewed. The particular combination of interviewer andstudent that evolves will determine, to some extent, the type of data obtained.

There has been much attention in recent literature to the subjectivity and complexityinherent in the interview encounter. Critical theory, feminist theory, critical race theory,queer theory, and postmodernism have been brought to bear on analyzing the intricaciesof the interview encounter. Although each of these perspectives challenges us to thinkabout what we are doing when interviewing, what they have in common is a concern forthe participants and their voices, the power dynamics inherent in the interview, theconstruction of the “story,” and forms of representation to other audiences.

Some of this discussion is framed in terms of insider-outsider status, especially withregard to visible social identities, most notably gender, race, age, and socioeconomicclass. Seidman (2013, p. 101) discusses how “our experience with issues of class, race,and gender…interact with the sense of power in our lives.” And, in turn, “the interviewingrelationship is fraught with issues of power—who controls the direction of the interview,who controls the results, who benefits.” Foster (1994), for example, explores theambiguities and complexities of the interviewer-respondent relationship in her study ofattitudes toward law and order among two generations. She analyzes her stance withregard to interactions with women versus men, the younger generation versus the older,middle class versus the working class.

Does a researcher need to be a member of the group being investigated to do a crediblestudy? Is it preferable for women to interview women or for Hispanics to interviewHispanics? What about the intersection of race, gender, and class? Are people more likelyto reveal information to insiders or outsiders? There are of course no single right answersto any of these questions, only the pluses and minuses involved in any combination ofinterviewer and respondent. Seidman (2013) suggests that along with being highlysensitive to these issues and taking them into account throughout the study, “interviewing

1748390 – Wiley US ©

requires interviewers to have enough distance to enable them to ask real questions and toexplore, not to share, assumptions” (p. 102).

Thus the interviewer-respondent interaction is a complex phenomenon. Both parties bringbiases, predispositions, attitudes, and physical characteristics that affect the interactionand the data elicited. A skilled interviewer accounts for these factors in order to evaluatethe data being obtained. Taking a stance that is nonjudgmental, sensitive, and respectfulof the respondent is but a beginning point in the process.

Recording and Transcribing Interview DataOf the three basic ways to record interview data, the most common by far is to audiorecord the interview. This practice ensures that everything said is preserved for analysis.The interviewer can also listen for ways to improve his or her questioning technique. Thepotential drawbacks are malfunctioning equipment and a respondent's uneasiness withbeing recorded. Most researchers find, however, that after some initial warinessrespondents tend to forget they are being recorded, especially if one uses an unobtrusivedigital recorder. Occasionally interviews are videotaped. This practice allows for recordingof nonverbal behavior, but it is also more cumbersome and intrusive than audio recordingthe interview.

A second way to record interview data is to take notes during the interview. Since noteverything said can be written down, and since at the outset of a study a researcher is notcertain what is important enough to write down, this method is recommended only whenmechanical recording is not feasible or when a participant does not want to be recorded.Some investigators like to take written notes in addition to recording the session. Theinterviewer may want to record his or her reactions to something the informant says, tosignal the informant of the importance of what is being said, or to pace the interview.

The third—and least desirable—way to record interview data is to write down as much ascan be remembered as soon after the interview as possible. The problems with thismethod are obvious, but at times, writing or recording during an interview might beintrusive (when interviewing terminally ill patients, for example). In any case, researchersmust write their reflections immediately following the interview. These reflections mightcontain insights suggested by the interview; descriptive notes on the behavior, verbal andnonverbal, of the informant; parenthetical thoughts of the researcher; and so on.Postinterview notes allow the investigator to monitor the process of data collection as wellas begin to analyze the information itself.

Ideally, verbatim transcription of recorded interviews provides the best database foranalysis. Be forewarned, however, that this is a time-consuming prospect; you can eithertranscribe the interview yourself or have someone do it for you. Hiring a transcriber can beexpensive, and there are trade-offs in doing so. You do not get the intimate familiarity withyour data that doing your own transcribing affords. Also, a transcriber is likely to beunfamiliar with terminology and, not having conducted the interview, will not be able to fillin places where the recording is of poor quality. If someone else has transcribed yourinterview, it is a good idea to read through the transcript while listening to it in order tocorrect errors and fill in blanks. However, hiring someone to transcribe allows you to spendtime analyzing your data instead of transcribing. We recommend that new and

1748390 – Wiley US ©

inexperienced researchers transcribe at least the first few interviews of any study, if at allpossible.

There are great benefits to transcribing the interview yourself, not the least of which isincreasing your familiarity with your data. If you do it yourself, you can write yourselfanalytic memos along the way. But even with good keyboarding skills, transcribinginterviews is a tedious process, though many of our students have found some voicerecognition software extremely helpful to this task, cutting down the time it takes. The onethat is mentioned most often and appears to be the most economical in terms of time andmoney is Dragon NaturallySpeaking. However, it generally recognizes only the voice of thetrained speaker. One of our students described her procedure for using it for transcriptionmuch the same way as is described on their website: using what they call the “parroting”function, which means to respeak the interview. She found this extremely helpful andrelatively quick; further, using it provided her the opportunity to become very familiar withher data in the process. The procedure she used is described on their website(www.nuance.com/dragon/transcription-solutions/index.htm):

Listen to the recording through the headphones of your Dragon headset and activateyour Dragon microphone and repeat the recorded text as you hear it.

Speaking the text aloud in your own voice enables Dragon to accu-rately transcribethe audio using the Dragon profile tuned to your voice.

Dragon turns your voice into text as quickly as you can speak the words—so there's noneed to constantly rewind the audio while you try to type out the corresponding text.

This of course is not the only transcription software available, but it is one that she foundvery helpful and strongly recommends to others. These technol