Applied Linguistics 2011: 32/1: 77–94 ß Oxford University Press 2010

doi:10.1093/applin/amq036 Advance Access published on 28 September 2010

Interview ‘Problems’ as Topics for Analysisy

KATHRYN ROULSTON
Department of Lifelong Education, Administration, and Policy, The University of

Downloaded from http://applij.oxfordjournals.org/ at University Library, University of Illinois at Chicago on June 24, 2014
Georgia
E-mail: roulston@uga.edu

In this article the author argues that interactional difficulties and questioning
practices identified in the methodological literature on qualitative interviewing
as ‘problems’ provide topics of analysis. Methodological examinations of
interview data drawing on conversation analysis also explicate how interview
‘problems’ may be conceptualized in relation to various theorizations of inter-
views. Two common conceptualizations of qualitative interviews include
a neo-positivist approach in which a neutral interviewer elicits participants’
descriptions that reflect ‘true’ states of affairs that are employed in realist reports,
and a romantic approach in which a reflexive interviewer facilitates genuine
rapport with participants to generate confessional data. The author uses a
constructionist conceptualization of interviewing to examine several issues
identified in methodological literature as ‘problems’, including the use of the
interview guide as a spoken survey, the use of closed questions, providing
possible responses in questions and asking questions that include assumptions
about participants’ life-worlds. These analytic demonstrations are used to argue
that researchers in the field of applied linguistics might use this approach to
examine interview interaction in order to develop insights into moderating
their interview practice as well as considering questions pertaining to research
design.

INTRODUCTION
The research interview has long been an important method of data generation
used by researchers to elicit descriptions from participants concerning their
experiences, perspectives, beliefs, and opinions. Widely used by qualitative
researchers across disciplines, the research interview is used by researchers
in the field of applied linguistics in conjunction with participant observation,
audio- and video-recording of naturally occurring interaction, and analysis of
documentary and textual data. As evidenced by numerous research reports
published in applied linguistics journals, researchers use research interviews as
a way to access participants’ descriptions of their life worlds.
In this article I propose an approach that may be used by researchers
in applied linguistics to examine interview interactions identified as
y
Part of 2010 Special Issue: Qualitative interviews in applied linguistics: Discursive
perspectives
78 INTERVIEW ‘PROBLEMS’ AS TOPICS FOR ANALYSIS

‘problematic’. Through such examination, researchers can account for the re-
searcher’s part in the co-construction of research data, show how participants
of research interviews artfully formulate their accounts in response to inter-
viewers, and demonstrate the ways in which problems might be viewed from
different theoretical perspectives. Such examinations show how ‘interview
problems’ become visible only in relation to particular theorizations of

Downloaded from http://applij.oxfordjournals.org/ at University Library, University of Illinois at Chicago on June 24, 2014
interviewing.
For example, in methodological literature on interviewing, common prob-
lems include interviewers posing questions with possible responses included
or ‘leading questions’ (e.g. Arksey and Knight 1999), or expressing personal
opinions and experiences (e.g. Weiss 1994). From a neo-positivist perspective
of interviewing, characterized by Alvesson (2003) as one in which the inter-
viewer focuses on generating ‘context-free truth’ or ‘facts’ through taking an
‘objective’ and ‘neutral’ stance, these actions are indeed problematic if the
interviewer is aiming to generate ‘valid’ data (Foddy 1993). For researchers
aiming to promote critical dialog with participants, the practice of asking lead-
ing questions might be viewed as a useful resource, rather than a problem
(Brinkmann 2007; Tanggaard 2007, 2008). For interviewers taking a ‘roman-
tic’ approach to interviewing, characterized by Alvesson (2003) as a perspec-
tive in which the interviewer strives to develop rapport with interviewees in
an effort to generate authentic, in-depth dialog that focuses on participants’
meanings, sharing one’s own experiences may be deemed as ‘good’ practice, as
in feminist interviewing (Oakley 1981).
Recent methodological work that draws on ‘constructionist’ approaches to
interviewing—that is, conceptualizations of the interview as a socially-situated
encounter in which both interviewer and interviewee play active roles
(Holstein and Gubrium 1995) takes the co-construction of interview data as
a topic of examination, rather than as a transparent resource. I argue that this
approach to interviewing provides a theoretically-grounded approach to ana-
lysis of interview data that researchers in applied linguistics might use to exam-
ine interview practices and inform interpretations of data. This type of
methodological analysis may be used by researchers to develop insights into
moderating future practice and to consider how studies might be designed.

CONCEPTUALIZING QUALITATIVE INTERVIEWS

In my reading of the methodological literature on qualitative interviews,
I have discerned at least six conceptualizations of research interviews.
Although multiple labels have been used to describe these approaches, I use
the terms ‘neo-positivist’, ‘romantic’, ‘constructionist’, ‘postmodern’, ‘trans-
formative’, and ‘decolonizing’ for heuristic purposes to consider different
ways of conceptualizing the human subject and data (Roulston 2010). Each
of these conceptualizations reflects different assumptions about the generation
and use of interview data for social research. For the purpose of this article,
I review the assumptions underlying the first three of these approaches to
 K. ROULSTON 79

interviewing, since I have found these to be the most common approaches to

interviewing in qualitative research. The other emerging conceptualizations
are discussed elsewhere (Roulston, 2010).
In a neo-positivist conception of interviewing (Alvesson 2003), the skillful
interviewer tends to mimic prescriptions provided for standardized survey
interviewers (e.g. Foddy 1993; Weiss 1994; Seidman 2006) by asking neutral

Downloaded from http://applij.oxfordjournals.org/ at University Library, University of Illinois at Chicago on June 24, 2014
questions—often in a particular sequence, and working to minimize researcher
bias through taking an objective role as interviewer. Providing affirmative re-
sponses such as ‘uh huh,’ ‘yes,’ or ‘OK’ to interviewees’ answers, for example,
is thought to bias responses (Seidman 2006: 89). The purpose of this form of
qualitative interview is to generate credible knowledge concerning the beliefs,
perceptions, experiences and opinions of the authentic self of the interviewee.
Seidman (2006: 9) describes the purpose of in-depth interviewing as ‘under-
standing the experience of other people and the meaning they make of that
experience.’ Gaining such understanding, Seidman argues, ‘provides access to
understanding their action’ (10). Holstein and Gubrium (2004: 144) describe
this approach as the ‘vessel-of-answers’ approach to interviewing, in which the
interviewer’s work consists of asking questions in a way that avoids biasing
‘what lies within the subject.’ Thus, the interviewer will usually refrain from
providing personal opinions and perspectives during the interview, and
self-disclosure is discouraged (Weiss 1994: 79). Open, rather than closed ques-
tions are used, and interviews are likely to be semi-structured or structured,
rather than unstructured or conversational.
Although neo-positivist assumptions about interviews are evident in much
research, there have been a number of important critiques of research that
relies on interview data—although these critiques do not delineate the theor-
etical conceptions of interviewing applied in this article. These are that partici-
pants do not necessarily tell the truth, or do what they say they do, the
researcher’s subjectivities and beliefs may ‘bias’ the data through expressions
of personal opinions in interview interaction, and representations of interview
data do not account for the researcher’s part in the co-construction of data
(Hammersley and Gomm 2008). The most common responses to these cri-
tiques address the first two issues. Multiple methods of data collection are
used for the purposes of ‘triangulation’ (e.g. observations provide opportu-
nities to check the accuracy of what participants have said); sustained field-
work over time allows researchers to check the stability of participants’ reports
and validity of interpretations; and researchers frequently seek participants’
responses and feedback concerning findings (‘member checking’).
The issue of researcher bias is addressed by a second conception of inter-
viewing, that of the romantic (Alvesson 2003: 16), or ‘emotionalist’ interview
(Silverman 2001). In contrast to the neo-positivist conception of the interview,
the interviewer–interviewee relationship here is one in which the interviewer
strives to develop genuine rapport and trust with participants in order to gen-
erate intimate and self-revealing conversations (e.g. Oakley 1981; Douglas
1985; Reinharz 1992; Rubin and Rubin 2005). To accomplish this, interviewers
80 INTERVIEW ‘PROBLEMS’ AS TOPICS FOR ANALYSIS

provide personal accounts in order to generate self revelatory and confessional

data which are then used to produce in-depth and intimate portraits of par-
ticipants’ life-worlds, and the beliefs, perceptions, experiences and opinions of
the authentic selves of interview subjects. For example, Douglas (1985: 25)
writes that ‘creative interviewing’ involves the use of strategies based on ‘an
understanding of friendly feelings and intimacy, to optimize cooperative, mutual

Downloaded from http://applij.oxfordjournals.org/ at University Library, University of Illinois at Chicago on June 24, 2014
disclosure and a creative search for mutual understanding’ (italics in original).
Some forms of feminist interviewing take a romantic approach to the gen-
eration of research data. One of the first to articulate what a feminist approach
to interviewing might look like, Oakley (1981: 41) asserted that
in most cases, the goal of finding out about people through inter-
viewing is best achieved when the relationship of interviewer and
interviewee is non-hierarchical and when the interviewer is pre-
pared to invest his or her own personal identity in the relationship.
Many feminist researchers have applied this approach, and Reinharz (1992:
32–33) provides examples of interviewers who have used self-disclosure
in order to initiate ‘ ‘‘true dialog’’ by allowing participants to become ‘‘co-
researchers’’ ’. From this perspective, data are seen to be co-constructed by
the interviewer and interviewee, and since it is likely that the researcher
will reveal his or her subjectivities relative to the participant and research
topic, researchers make few or no claims that they have generated ‘objective’
data. Thus, research reports may include the interviewer’s utterances in rela-
tion to those of the interviewee, in addition to subjectivity statements con-
cerning the researcher’s interests and relationships to the topic of study and
participants. Although this approach to the interview takes seriously argu-
ments that it is neither possible, nor necessary to eliminate ‘researcher bias’
in open-ended qualitative interviews, the notion that researchers are able to
access the ‘authentic self’ of the interview subject via interview talk still under-
lies this approach. This view has been seriously questioned by some (e.g. Potter
and Hepburn 2005), who argue that interview data cannot accurately reflect
the interior states of minds of participants.
A third theoretical conception of qualitative interviews is that of the con-
structionist or ‘localist’ (Alvesson 2003) approach, which questions whether
any access to ‘authentic selves’ is possible via interviews. I use the label ‘con-
structionist’ to refer to the central notion that interviews are co-constructed by
both interviewer and interviewee, and that the data generated provide ‘access
to a cultural universe and its content of moral assumptions’ (Silverman 1993:
108). From this perspective, interview data generated within the social setting
of the research interview exemplifies but one cultural event within the life
world of the participant. Rather than viewing interview data as reports that are
directly reflective of ‘interior’ states of mind or ‘exterior’ states in the world;
instead, they are viewed as ‘accounts’—or practical displays of the local organ-
ization of social order by speakers (Baker 2002, 2004). Thus, the interviewer’s
contributions are subject to the same kind of analytic focus as that of the
 K. ROULSTON 81

interviewee, and both structural and topical features of talk are examined. This
approach to interviews has been critiqued as having an overly narrow focus on
form (how interview interaction occurs), and the aim of examining both ‘how’
data are constructed and ‘what’ the topic of talk has been questioned (see
Silverman 2001: 97–98).
This brief overview of three theoretical stances toward qualitative interview-

Downloaded from http://applij.oxfordjournals.org/ at University Library, University of Illinois at Chicago on June 24, 2014
ing reveals varied assumptions about knowledge production and the human
subject. Given this variety in the way in which qualitative interviews are
conceptualized by researchers—what might be made of the methodological
advice given to novice researchers concerning the conduct of quality inter-
views? There are numerous guides within the field of qualitative inquiry to
what researchers might do in interviews in order to generate quality data
(e.g. Kvale 1996; Patton 2002; deMarrais 2004; Rubin and Rubin 2005;
Seidman 2006), in addition to specific guides for researchers in particular
fields [e.g. Richards (2003) addresses qualitative interviewing for researchers
in TESOL]. Novice interviewers are commonly advised to avoid certain kinds
of interactions when asking questions of others for research purposes. Yet what
is revealed when interactions defined as ‘problematic’ are closely examined?
In the next section, I provide an outline of the data sets from which I have
drawn sequences of talk to illustrate various kinds of interview problems.

ANALYTIC METHODS AND DATA SOURCES

The studies from which interview excerpts are drawn include an evaluation
project of a federally-funded training grant in Mind Body Medicine (MBM) at
a family practice residency program (Study 1); a research project examining
clients’ perspectives of services offered by a conference center conducted as an
authentic project in a qualitative interview class (Roulston et al. 2008) (Study
2); and a study examining novice researchers’ conceptualizations of research
problems (Study 3). All projects were conducted in the USA. For the purpose
of illustration, I have selected excerpts that demonstrate interactions on the
part of the interviewer that are recognized as ‘problems’ in methodological
literature on qualitative interviewing. These include:
Using the interview guide as a spoken survey;
Asking closed questions;
Providing possible responses in questions; and
Asking questions that include assumptions about participants’
life-worlds.
Conversation analysis (CA) as a method of investigating talk-in-interaction
was developed by Harvey Sacks and his colleagues (Sacks 1992; Silverman
1998; Schegloff 2007). CA focuses on examining the conversational resources
used by members in everyday interaction (Psathas 1995; Pomerantz and Fehr
1997; Hutchby and Wooffitt 2008; Have 2007), and was originally used to
analyze mundane talk rather than research interviews. More recently, scholars
82 INTERVIEW ‘PROBLEMS’ AS TOPICS FOR ANALYSIS

have used methods drawn from CA, ethnomethodology (EM) (Garfinkel 1967,
2002; Have 2004) and membership categorization analysis (MCA) (Baker
2002, 2004) to analyze interview data (e.g. Mazeland and Have 1998;
Rapley 2001; Roulston 2006; Bartesaghi and Bowen 2009; Schubert et al.
2009).
Since questions and answer (Q–A) sequences are integral to research inter-

Downloaded from http://applij.oxfordjournals.org/ at University Library, University of Illinois at Chicago on June 24, 2014
views, my analysis begins by examining how Q–A sequences which can be
construed as problematic unfolded on a turn-by-turn basis. As Sacks (1992)
has shown, Q–A sequences are one of a class of utterances known as adjacency
pairs that prefer a specific kind of utterance in the next turn. In an adjacency
pair, once the question or ‘first-pair part’ is uttered; a slot becomes available for
the answer, or ‘second-pair part’, although this may not occur. In the analyses
presented below, I ask:
How are questions formulated by the interviewer?
How do interviewees orient to the interviewer’s questions?
What actions are accomplished in the talk?

USING THE INTERVIEW GUIDE AS A ‘SPOKEN’ SURVEY

Beginning interviewers frequently use the interview guide as a script rather
than a flexible resource for asking questions. Novice interviewers are
frequently advised to use the interview guide as just that—a guide. If inter-
viewees have already responded to questions listed on the interview guide,
then it is thought to be good practice to refrain from asking questions again
(e.g. deMarrais 2004: 67). This interview ‘problem’ then, reflects assumptions
about the interviewer–interviewee relationship that align with a romantic con-
ception of interviewing. That is, if interviewers are truly listening to partici-
pants, then it is seen to be poor practice and damaging to rapport-building to
ask a question that has already been answered by a participant. Surveys are
seen to be a more effective means of generating data if it is important as part of
the research design for interviewers to ask the same questions of multiple
participants. Consider the following sequence in which an interviewee is
asked to respond to a specific question that ostensibly had been addressed in
prior talk. From a romantic conception of interviewing, this interaction
would be correctly read as an ‘interview problem,’ and indeed, the breach of
interactional norms that occurs is made accountable by speakers in the
interaction.

Excerpt 1, Study 1 (October 2007, 26:08–28:12)

IR: Interviewer
IE: Interviewee

1. IR! u::::m (.) I think you probably have already answered this question
2. in various ways but I’m just going to state the question as I’ve got it
 K. ROULSTON 83

3. written here (.) so how- what do you see as the sig"nificance for a a
4. general practitioner um having knowledge and skills with mind-body-
5. spirit interventions (.) so what are your perceptions of the significance
6. of that to the work of being a (.) a family medicine practitioner
7. (1.0)
8. IE! u:::m well I don’t know how to rephrase what I al- said already=

Downloaded from http://applij.oxfordjournals.org/ at University Library, University of Illinois at Chicago on June 24, 2014
9. ((shared laughter))
10. IR! YEAH that’s why I said I think you’ve already answered this=
11. IE =yeah so=
12. IR =yeah=
13. IE =going back to that if I want to say one two three four I would say well
14. u::::h (2.0) as a general practitioner you see an entire spectrum of
15. patients=
16. IR =uh huh=
17. IE =irrespective of whether they’re ill or not they might be coming to us
18. fo::r moral support sometimes (.) they don’t get to a specialist for that
19.
20. [lines omitted]
21.
22. IE so that way a general practitioner could make a lot of difference=
23. IR =hm mm=
24. IE =um (.) I think that’s the way I would put it=
25. IR =yes yeah=
26. IE =so=
27. IR =yeah thank you=
28. IE =it’s going to be more useful in a general practice=
29. IR =yes=
30. IE =set up and I think all general practitioners should be aware of it
31. whether they are going to incorporate it or not ( )=
32. IR =yes=
33. IE =is dependent on [their way of looking at it
34. IR [yes yeah

In this sequence occurring mid-way through the interview, a topic change

from talk immediately prior is initiated by the interviewer via a question
(lines 1–6). The question is prefaced with an acknowledgement by the inter-
viewer that the interviewee had apparently already answered the question.
The potential delicacy of asking a question that has already been answered is
addressed in the interviewer’s comment: ‘I think you probably have already
answered this question in various ways’. The institutional obligations of her
work as an external evaluator are referred to by the interviewer: ‘I’m just going
to state the question as I’ve got it written here’ (lines 2–3). In responding, the
interviewee acknowledges that he has already answered the question, ‘u:::m
84 INTERVIEW ‘PROBLEMS’ AS TOPICS FOR ANALYSIS

well I don’t know how to rephrase what I al- said already’ (line 8), and both
speakers respond with shared laughter. The packaging of utterances by the
interviewer at lines 1–3, 10, and the interviewee’s initial response at line 8
demonstrate that both speakers are aware of the breach in interactional order
posed by asking a question that has already been answered, and that this is
a sensitive matter which is accountable. Given that both speakers begin

Downloaded from http://applij.oxfordjournals.org/ at University Library, University of Illinois at Chicago on June 24, 2014
by acknowledging in this sequence that this question had already been an-
swered—was there any value in the restatement of this question?
At line 13, the interviewee begins his answer to the question by foreshadow-
ing a list of items (‘one two three four’), and then re-stating a view that he has
proposed earlier (that general practitioners see patients who exhibit a range of
problems). What is new in the sequence that follows is that the interviewee
emphasizes the distinction between the work of ‘specialists’ and the work of
‘general practitioners (GPs)’ in a way that had not been highlighted prior to
this sequence of talk. The interviewee sums up his response to the question
posed at lines 1–6 from lines 22–33. Here he repeats an idea stated earlier in
the interview (that GPs should be open to learning about new treatments); and
adds a new idea that whether GPs will actually incorporate this knowledge
depends on their perspective.
In this sequence we see the interview guide being used in a standardized
way in that a question on the interview guide that had already been answered
by the interviewee is asked. From a neo-positivist conception of interviewing
this is acceptable practice, especially if there are multiple research participants
from whom researchers would like to elicit responses about the same topics.
From a romantic conception of interviewing, this is likely to be read as poor
interview practice in that it overlooks the importance of recognizing the con-
tent of prior talk. A constructionist perspective of interviewing shows in detail
how the data were co-constructed by participants in this sequence. Here, in
response to the interviewer’s question—which is formulated in a way that
orients to the delicacy of restating a question that has already been an-
swered—the interviewee used the opportunity to restate his views by sum-
marizing ideas that he had presented at different points throughout the
interview (that there would likely be more opportunity to use MBM in general
practice than in specialist work; and that GPs should be open to learning about
new treatments); but also adding to his earlier accounts by emphasizing the
differences between the work of specialists and GPs, and acknowledging that
GPs’ perspectives and applications of MBM vary.
While this interactional move on the interviewer’s part might be seen as
problematic from a romantic perspective of interviewing, given that the
research purpose of this project was to elicit participants’ perspectives and
evaluations of a particular program, the information generated here was
useful, in that the participant summarized and highlighted his perspectives
relevant to this question. Thus, while not re-stating questions that have
been answered before they have been asked is a useful rule-of-thumb, deviat-
ing from this advice may not always be fruitless. What close examination of
 K. ROULSTON 85

this sequence reveals, however, is that the interviewer does extensive work to
pose the question in a way that shows that she has listened to and heard prior
talk and that she is aware of the potential breach of interactional norms in
re-stating a question that the interviewee has already answered.

USING CLOSED QUESTIONS

Downloaded from http://applij.oxfordjournals.org/ at University Library, University of Illinois at Chicago on June 24, 2014
Closed questions are usually conceptualized as a ‘problem’ for qualitative inter-
viewers—given that open-ended questions are deemed to elicit the kind of de-
scriptive data that qualitative researchers seek to generate, and closed questions
encourage yes/no responses or short answers (Patton 2002; deMarrais 2004;
Brenner 2006; Seidman 2006). Yet, follow-up questions posed in a closed
format may be used by interviewers to clarify their understanding or request
specific information relevant to prior talk. Take, for example, two closed ques-
tions shown in Excerpt 2 that are asked of a participant in Study 3. The purpose
of the study was to elicit doctoral students’ accounts of how they went about
selecting research topics and formulating research questions. In Excerpt 2, at
lines 1 and 5, the interviewer asks what Stewart et al. (2006) label ‘implicit
closed questions,’ in which the response type has been implied but not provided.
These authors distinguish implicit closed questions from those that are ‘explicit,’
or provide possible responses in the question (e.g. on a scale of 1–5 with 1 being
excellent, and 5 being very poor, how do you rate x?).
The first of the closed questions posed in Excerpt 2 seeks further information
concerning a prior topic of talk—whether the interviewee has presented
or published findings from his master’s thesis. This question implies a yes/no
response in relation to two options—presentation and publication. The se-
cond closed question requests further information concerning the name of
the conference at which he gave a presentation.
Excerpt 2, Study 3 (27 May 2009, 15:48–16:15)
1. IR! yeah did you present (.) or publish that (.) piece of work?=
2. IE =um (.) I think I’ve presented it at a conference=
3. IR =uh huh=
4. IE =but I’ve never tried to have it published no no=
5. IR! =yeah yeah and what conference was that?=
6. IE =actually the::: (1.0) 8is that the only conference?8 the National Black
7. Speech As[sociation Conference=
8. IR [uh huh
9. IR =uh hm uh hm=
10. IE =I did it prior right before I defended it I presented it there
11. IR =uh huh=
12. IE =at the National Black Association Speech conference=
13. IR =uh huh=
14. IE =so
86 INTERVIEW ‘PROBLEMS’ AS TOPICS FOR ANALYSIS

This questioning sequence elicits specific details from the interviewee: he

affirms that (i) he did present work from his master’s thesis and that (ii) he has
never sought to publish from it. The interviewee provides the name of the
conference at which he presented, and adds detail concerning when his con-
ference presentation occurred.
This use of closed questions elicited specific information concerning the par-

Downloaded from http://applij.oxfordjournals.org/ at University Library, University of Illinois at Chicago on June 24, 2014
ticipant’s prior research experiences. It is not possible to judge whether the
participant would have provided this level of detail had an open question been
asked after he had talked about his thesis (e.g. ‘Tell me more about that.’).
Here, closed questions generate factual information concerning this partici-
pant’s prior experiences in research, which was one of the interview questions
that had been posed earlier. If researchers are seeking specific kinds of infor-
mation from participants, then closed questions can be used as a resource, and
may be particularly useful as follow-up questions. For researchers seeking to
elicit narratives from participants, however, closed questions of this sort are
likely to interrupt the story-flow and could possibly inhibit rapport building
between speakers (for examples of the kinds of responses generated from
closed questions, see Patton 2002: 356).
In Excerpt 3 drawn from Study 1, I provide a second example of closed
questions used as a resource to assess the accuracy of my understandings of
an interviewee’s prior talk. These are found in lines 8 and 17, and relate to
formulations (Heritage and Watson 1979) of the participant’s prior utterances.
This physician had been asked to indicate her position in relation to my inter-
pretation of the range of perspectives of MBM expressed by physicians in an
earlier round of interviews. These were presented on a printed diagram, and
related to the usefulness of learning about MBM:
Strongly Critical Skeptical Advocacy
Prior to Excerpt 3 below, this interviewee had indicated that she was both
‘strongly critical’ and an ‘advocate’ for MBM.

Excerpt 3, Study 1 (3 June 2008:58–4:41)

1. IR strongly critical so:: let me get this straight=
2. IE =heh heh=
3. IR =so:: (.) um you would "advocate for the importance of this i- for u::m
4. someone in a residency program
5. IE yeah <for anyone>=
6. IR =OK=
7. IE =for "anyone "period
8. IR! in family care?=
9. IE ="absolutely=
10. IR =OK=
11. IE =yeah
12. IR so (.) and then but at the same time because of your "prior experience
 K. ROULSTON 87

13. (.) you: were: strongly critical about its useful[ness in the
14. IE [yes
15. IR program here
16. IE yeah=
17. IR! =would that be correct?=
18. IE =yeah I think so

Downloaded from http://applij.oxfordjournals.org/ at University Library, University of Illinois at Chicago on June 24, 2014
19. IR uh huh can can [y-
20. IE [<which has to do with the implementation> not the:
21. IE [core::
22. IR [uh huh
23. IE (.) you know not the core of <what we’re talking about>
24. [that has to do with the delivery and how it was gonna affect you
25. IR [uh huh
26. IE know my time=
27. IR =uh huh=
28. IE =and all of that=
29. IR =uh huh

Unlike Excerpt 2, in which closed questions had been used to elicit specific
information concerning prior talk, in Excerpt 3, closed questions are specific-
ally marked as checking the interviewer’s understandings of prior talk, and to
check an issue about which she is unclear. This is indicated in line 1: ‘so:: let
me get this straight’. In both cases the closed questions prefer yes/no responses,
with confirmations delivered by the participant in lines 9 and 18. The first
closed question, ‘in family care?’ follows the formulation, ‘so:: (.) um you
would "advocate for the importance of this i- for u:m someone in a residency
program’ (lines 3–4), and seeks clarification of the participant’s response to the
formulation. Here, the interviewer sought to check whether the participant is
referring to any physician, or family care physicians as served by this residency.
The second formulation followed by a closed question, posed as ‘at the same
time because of your "prior experience (.) you: were: strongly critical about its
usefulness in the program here would that be correct?’ (lines 12–13) seeks to
elicit a confirmation from the interviewee concerning my understandings of
her utterances: while this physician has self-identified as an ‘advocate’ of
MBM, she is also ‘strongly critical’ of the training program in MBM. At line
19, the delivery of an additional follow up question is overlapped by the par-
ticipant—who preempts the question by providing an account that attends
to the apparent contradiction between the positions she has claimed (lines
20–24). In Excerpt 3, closed questions are used in combination with formula-
tions to gain further understanding of this participant’s reasoning for positions
which the interviewer has oriented to as problematic at line 1 (‘let me get this
straight’).
Although the use of closed questions in open-ended interviews may not
generate the detailed stories sought by many qualitative researchers, these
88 INTERVIEW ‘PROBLEMS’ AS TOPICS FOR ANALYSIS

excerpts show that they may be valuable resources in instances in which re-
searchers want to learn about specific details of a participant’s experience
(Excerpt 2), or to accomplish on-site member checking of their understanding
of topics discussed (Excerpt 3). Kvale (1996: 145) suggests that interviewers
verify their understandings of talk throughout the interview interaction. In
Excerpt 3, the interviewer poses formulations of prior talk in conjunction with

Downloaded from http://applij.oxfordjournals.org/ at University Library, University of Illinois at Chicago on June 24, 2014
closed questions in order to check her understanding of what the participant
had said. Rather than suggest that questions in which interviewers challenge
participants to account for their prior talk be used in any kind of qualitative
interview, researchers might evaluate their use of such a resource in order to
use it in a context-sensitive way that will align with their theoretical assump-
tions about interviewing, and the purpose of the research study.

POSING QUESTIONS WITH POSSIBLE RESPONSES

Methodological literature concerning qualitative interviews suggests that
interviewers refrain from providing possible responses in the questions, and
ask one question at a time (e.g. Arksey and Knight 1999; Patton 2002;
Seidman 2006). Excerpt 4 below shows an example in which a novice inter-
viewer conducting an interview with a client of a conference center as part of
a class project asked multiple questions (Study 2). The interviewer provides
possible responses for the interviewee—in this case, seeking a comparison of
the room in which they were sitting to the conference rooms usually used.

Excerpt 4, Study 2 (May 2004)

1. IR what about the setup of the of the rooms what type of a setup do do you
2. generally use or you do you have different types of formats that they run
3. say like a conference style we’re sitting at in this room or is it classroom
4. seats or=
5. IE =no actually kind of like theater-style=
6. IR =OK=
7. IE =that’s the way that [the Conference] Hall is set up permanently=
8. IR =OK=
9. IE =is theatre style and then um there is a stage at at the front and (.) um our
10. symposia is really composed "mostly of roundtable discussions
11. and so there is a group of people seated on stage (.) and a moderator
12. behind a podium there’s usually I think six (.) five or six people seated on
13. stage and the [Conference] Center setup for that which is the one
14. we’ve always used is very very elegant the chairs are not like this
15. and they’re not straight chairs they’re they’re very nicely appointed
16. furniture=
17. IR =mm hm=
18. IE =and there are tables in between the chairs with flowers um (.) uh
 K. ROULSTON 89

19. water pitchers (.) and the um (.) the participants in the roundtable are
20. always miked (.) and the audiovisual people at the [conference] center
21. are "excellent when it comes to that um (.) I don’t be- we hardly ever
22. have any communications type electronic type glitches when it comes to
23. the sound’s always really good and they make people very comfortable
24. when they have to be clipping on a little microphone (.) so that always

Downloaded from http://applij.oxfordjournals.org/ at University Library, University of Illinois at Chicago on June 24, 2014
25. goes off without a hitch

In lines 1–4, the IR poses multiple questions concerning the ‘type of setup’
usually used, and includes two candidate responses concerning the ‘different
types of formats’ employed at the conference center. In response, the inter-
viewee selects an alternative to those offered which she describes in detail,
including what ‘theater style’ entails; and the Conference Center staff’s
multi-faceted work in arranging that (lines 9–25). In formulating her response,
the participant provides a strong endorsement for the quality of the services
and facilities offered by the Conference Center even though the interviewer
had not yet asked her to evaluate the services. Yet, in answering the research-
er’s question, the interviewee addressed the larger topic of the interview—that
of eliciting opinions and evaluations of services rendered.
In Excerpt 4, the interviewer’s questions that included possible responses
generated a detailed and spontaneous description from the interviewee that
responded to the question concerning room arrangement, as well as the inter-
viewee’s ‘usual’ experience in conference planning. While multiple questions
and questions configured with possible responses in them have been viewed as
‘problematic,’ an investigation of Excerpt 4 shows otherwise. Puchta and
Potter (1999, 2004) have shown that focus group moderators use ‘elaborate
questions’ as a resource to secure participation of group members as new topics
are introduced in focus groups. These kinds of questions include multiple
components and candidate responses which provide guidance to research par-
ticipants in how to respond. In Excerpt 4, the interviewer’s initial question
includes features of an elaborate question, and is taken up in a way that
generated detailed descriptions pertinent to the study in question.
From a neo-positivist conception of interviewing, questions that include
possible responses are thought to ‘lead’ the interviewee towards possible
responses, thus ‘biasing’ the data. From a romantic conception of interviewing,
these kinds of questions provide possible wordings, and are seen to prevent
participants from formulating narrative responses in their own terms. In
Excerpt 4, a close examination of the participant’s response shows that she
selected the terms in which to respond to these questions, and rejected the
interviewer’s offerings. Again, what occurs in this excerpt cannot be extrapo-
lated to make claims about the routine ways in which qualitative interviews
are accomplished. Yet this instance shows a possible way in which interview
talk might unfold in a way that does not align with prescriptions offered by
methodological literature.
90 INTERVIEW ‘PROBLEMS’ AS TOPICS FOR ANALYSIS

QUESTIONS THAT INCLUDE ASSUMPTIONS ABOUT

PARTICIPANTS’ LIFE-WORLDS
Excerpt 5 from Study 1 shows an example of how interviewees can take up
interviewers’ invitations for particular kinds of responses in a way that dis-
agrees with the premise of the question. In essence, focus group participants in

Downloaded from http://applij.oxfordjournals.org/ at University Library, University of Illinois at Chicago on June 24, 2014
Excerpt 5 do not orient to the assumption embedded in the question posed. In
this focus group, the moderator (M) formulated prior talk to transition to a
question on the focus group guide—that of a discussion of ‘drawbacks’ of the
training program in which participants were involved. Given that this was an
evaluation project in which participants of the study were invited to discuss
both positive and negative features of an innovation (‘benefits’ and ‘draw-
backs’), this kind of question assumes that participants are able to provide
descriptions and opinions about both ‘positive’ and ‘negative’ features of a
particular experience. Excerpt 5 begins with a summary of what had been
said by one of the focus group participants in a way that linked his comment
to the a priori framework of the interview protocol. What is interesting in
lines 1–3 is that the moderator selectively focused on one aspect of interviewee
8’s response (not included here). In the turn prior to this sequence, interview-
ee 8 mentioned two ‘disadvantages’ of the use of MBM in the work of GPs, but
his response also made mention of four advantages. In the question posed at
lines 1–3, the moderator drew on a single topic of prior talk (here formulated
as ‘one of the drawbacks’) to move to the next question on the interview
guide.

Excerpt 5, Study 2 (Focus group, October 2007, 31:45–32:42)

M: Moderator
IE: Interviewee 8
IE?: Unidentified interviewee/s
1. M IE 8’s mentioned one of the drawbacks- well actually this is a drawback
2. of maybe "implementing it is that of clear guidelines that there’s a lack
3. of clear guidelines for practice? is that accurate?=
4. IE8 =mm hm=
5. M! =u:m do you see any drawbacks for u:m this kind of intervention program
6. being in a residential um residence family residence program (.) training
7. program?
8. IE8 well I think if we don’t try that in a residency then who else is going to
9. try=
10. M =uh huh uh huh=
11. IE8 =so ( ) people are still open for learning=
12. M =uh huh=
13. IE8 =and the patient maybe (helped) come to the expectancy of trying
14. something new so I think both the investigator and the patient would
 K. ROULSTON 91

15. be more open to give it a try=

16. M =uh huh (.) so any other perspectives on that one? so drawbacks for
17. actually learning about it?
18. (2.0)
19. IE? mm hm I think it’s excellent=
20. IE? =yeah I think it’s a great opportunity=

Downloaded from http://applij.oxfordjournals.org/ at University Library, University of Illinois at Chicago on June 24, 2014
21. M =uh huh

At line 4, interviewee 8 agrees with the interviewer’s restatement of his prior

talk as ‘lack of clear guidelines.’ However three group members’ responses
(lines 8–15, 19–20) to the question posed at lines 5–7 (‘u:m do you see any
drawbacks for u:m this kind of intervention program being in a residential um
residence family residence program (.) training program? ’) indicate that the
basic premise of the question—that they might have something to say about
‘drawbacks’—has not been taken up. In lines 19–20, two group members pro-
vide strong endorsement for the view expressed in lines 8–15 by interviewee 8
that the residency is an appropriate place to include such training. Interviewee
8 initiates the turn with ‘well,’ which indicates a forthcoming disagreement:
‘well I think if we don’t try that in a residency then who else is going to try.’
At lines 11 and 13–15, interviewee 8 expands on this statement by providing
the reasoning for the claim that the training program is an appropriate inclu-
sion in a family care residency because both physicians and patients may be
open to learning about new treatment options.
This raises the question of whether interviewers should ask questions with
prior assumptions embedded in them. Certainly, if research topics rely on
participants having certain experiences that they are able to describe, then it
is important for researchers to recruit people who can talk about those. In the
example provided in Excerpt 5, in which three participants of this group pro-
vided positive evaluations in response to a question seeking any negative views
(or ‘drawbacks’), the data generated is valuable in that it indicates that by
disagreeing with the very premise of the question, their positive evaluations
may be seen to be even more strongly articulated then if they had simply
responded to the question concerning ‘benefits’ of the program. Rather than
discard such questions from interview protocols altogether, when participants
do not orient to the assumptions embedded in questions as illustrated in
Excerpt 5, it is useful for researchers to consider what the question was to
which they responded. In this case, participants responded to a different ques-
tion than had been asked. The question they appear to have responded to is
‘what is your opinion of the MBM training program at the family residency?’
Responses such as these, then, might be used to inform the construction of
questions for future interviews. That is, for this particular study, in succeeding
rounds of data generation, opportunities for participants were purposefully
provided to describe the positive or useful outcomes of the program that are
implied in this sequence with this group of participants. Interviewee 8 was also
92 INTERVIEW ‘PROBLEMS’ AS TOPICS FOR ANALYSIS

interviewed individually in order to elicit further information concerning his

viewpoints.
From a neo-positivist perspective of interviewing, asking questions that in-
clude assumptions about participants’ life-worlds is problematic, because it fails
to situate the interviewer as neutral and objective. From a romantic perspec-
tive of interviewing, asking these kinds of questions might also be seen as

Downloaded from http://applij.oxfordjournals.org/ at University Library, University of Illinois at Chicago on June 24, 2014
problematic, albeit for different reasons. In providing a priori assumptions to
which participants respond, rather than providing interviewees opportunities
to set the agenda for talk within the interview, interviewers fail to build rap-
port with participants through sharing their own experience, and may subse-
quently instigate disagreement with participants (e.g.Weiss 1994: 107–110;
Patton 2002: 357; deMarrais 2004: 67). If the interview is seen from a con-
structionist perspective, participants’ rejections of interviewers’ questions are
also data to be analyzed. The usefulness of data generated by these kinds of
questions, however, is heavily dependent on the researcher’s recognition of
the ways in which participants disagree with researchers’ assumptions in both
subtle and overt ways.

CONCLUSION
In this article I have demonstrated a method by which researchers in the field
of applied linguistics can ask questions of problematic interactions that have
taken place in research interviews. I selected the ‘problems’ based on my ex-
periences as an interviewer, teacher of qualitative interviewing and reader of
methodological literature. I have mostly selected to use talk from my own
interviews, rather than those of my students, since I take the position that
studying one’s own interview talk can lead to a mindful consideration of one’s
role in the generation of data for research purposes. While the advice literature
in qualitative interviewing advocates certain kinds of ‘best’ practices that inter-
viewers should follow—there is no way of knowing what interviewees
will do in response to the interviewer’s actions, nor what interviewers will
do next. In the examples I have shown in this article, I have shown through
line-by-line analysis that interactions defined as ‘problems’ in methodological
literature on interviewing, when seen in context, may only be read as prob-
lematic in relation to one’s theoretical conception of interviews and research
purposes.
Clearly, methodologists offer advice for the purpose of providing rules-of-
thumb that need to be applied in nuanced ways. When acquiring new skills,
however, novices, tend to apply rules in ‘context-free’ ways, as Dreyfus and
Dreyfus (1986) have shown. As an example, after conducting a research inter-
view for the first time, one of my students reflected that he had interrupted his
interviewee several times, and was working to amend this practice. In listening
to the interview, however, these ‘interruptions’ could also be construed as
well-placed probes for specific detail from an artful story-teller who skillfully
held the floor and steered the conversation. While interrupting participants by
 K. ROULSTON 93

probing for further detail is not a practice that is usually recommended, in this
case it may have been warranted. By examining interview problems and puz-
zling interactions in detail, interviewers and researchers may examine the
variety of actions that take place in interview contexts and what the outcomes
are for the data generated. In doing this kind of work, then, my aim is to
encourage researchers to look closely at their own interview interaction to

Downloaded from http://applij.oxfordjournals.org/ at University Library, University of Illinois at Chicago on June 24, 2014
examine how data are generated by speakers. The purpose of analyzing tran-
scripts as I have demonstrated here is not to illuminate a set of prescriptive
rules for others to follow, but rather for each one of us to mindfully consider
our actions and interactions with others.

SUPPLEMENTARY DATA
Supplementary material is available at Applied Linguistics online.

ACKNOWLEDGEMENTS
Thanks to Steven Landry, Marty Thomas and Seon Joo Kim for their assistance in preparation of
this manuscript.

REFERENCES
Alvesson, M. 2003. ‘Beyond neopositivists, Brinkmann, S. 2007. ‘Could interviews be epi-
romantics, and localists: A reflexive approach stemic? An alternative to qualitative opinion
to interviews in organizational research,’ polling,’ Qualitative Inquiry 13/8: 1116–38.
Academy of Management Review 28/1: 13–33. deMarrais, K. 2004. ‘Qualitative interview
Arksey, H. and P. Knight. 1999. Interviewing for studies: Learning through experience’
Social Scientists: An Introductory Resource with in K. deMarrais and S. D. Lapan (eds):
Examples. Sage. Foundations for Research: Methods of Inquiry in
Baker, C. D. 2002. ‘Ethnomethodological Education and the Social Sciences. Lawrence
analyses of interviews’ in J. Gubrium and Erlbaum Associates, pp. 51–68.
Holstein (eds): Handbook of Interviewing: Douglas, J. D. 1985. Creative Interviewing. Sage.
Context and Method. Sage, pp. 777–95. Dreyfus, H. and S. Dreyfus. 1986. Mind
Baker, C. D. 2004. ‘Membership categorization over Machine: The Power of Human Intuition
and interview accounts’ in D. Silverman (ed.): and Expertise in the Era of the Computer. Free
Qualitative Research: Theory, Method and Practice Press.
2nd edn. Sage, pp. 162–76. Foddy, W. 1993. Constructing Questions for Inter-
Bartesaghi, M. and S. P. Bowen. 2009. views and Questionnaires: Theory and Practice in
‘The acquisition of memory by interview Social Research. Cambridge University Press.
questioning: Holocaust re-membering as Garfinkel, H. 1967. Studies in Ethnomethodology.
category-bound activity,’ Discourse Studies Prentice-Hall.
11/2: 223–43. Garfinkel, H. 2002. Ethnomethodology’s Program:
Brenner, M. E. 2006. ‘Interviewing in educa- Working out Durkheim’s Aphorism. Rowman and
tional research’ in J. L. Green, G. Camilli, Littlefield.
P. B. Elmore, and A. Skukauskaite_ (eds): Hammersley, M. and R. Gomm. 2008.
Handbook of Complementary Methods in Education ‘Assessing the radical critique of interviews’
Research. Lawrence Erlbaum Associates, in M. Hammersley (ed.): Questioning Qualitative
pp. 357–70. Inquiry: Critical Essays. Sage, pp. 89–100.
94 INTERVIEW ‘PROBLEMS’ AS TOPICS FOR ANALYSIS

Have, P. T. 2004. Understanding Qualitative Richards, K. 2003. Qualitative Inquiry in TESOL.

Research and Ethnomethodology. London, Sage. Palgrave Macmillan.
Have, P. T. 2007. Doing Conversation Analysis: Roulston, K. 2006. ‘Close encounters of the
A Practical Guide, 2nd edn. Sage. ‘CA’ kind: A review of literature analyzing
Heritage, J. C. and D. R. Watson. 1979. talk in research interviews,’ Qualitative
‘Formulations as conversational objects’ Research 6/4: 515–34.
in G. Psathas (ed.): Everyday Language: Studies Roulston, K. 2010. Reflective Interviewing: A Guide

Downloaded from http://applij.oxfordjournals.org/ at University Library, University of Illinois at Chicago on June 24, 2014
in Ethnomethodology. Irvington, pp. 123–62. to Theory and Practice. Sage.
Holstein, J. A. and J. F. Gubrium. 1995. The Roulston, K., V. J. McClendon, A. Thomas,
Active Interview. vol. 37. Sage. R. Tuff, G. Williams, and M. Healy. 2008.
Holstein, J. A. and J. F. Gubriuim. 2004. ‘The ‘Developing reflective interviewers and
active interview’ in D. Silverman (ed.): reflexive researchers,’ Reflective Practice 9/3:
Qualitative Research: Theory, Method and Practice. 231–43.
Sage, pp. 140–61. Rubin, H. J. and I. S. Rubin. 2005. Qualitative
Hutchby, I. and R. Wooffitt. 2008. Conversation Interviewing: The Art of Hearing Data, 2nd edn.
Analysis, 2nd edn. Polity Press. Sage.
Kvale, S. 1996. InterViews: An Introduction to Sacks, H. 1992. Lectures on Conversation.
Qualitative Research Interviewing. Sage. Blackwell.
Mazeland, H. and P. T. Have. 1998. ‘Essential Schegloff, E. A. 2007. Sequence Organization in
tensions in (semi-)open research interviews’ Interaction: A Primer in Conversation Analysis I.
in I. Maso and F. Wester (eds): The Deliberate Cambridge University Press.
Dialogue: Qualitative Perspectives on the Interview. Schubert, S. J., S. Hansen, K. R. Dyer, and
VUB University Press, pp. 87–113. M. Rapley. 2009. ‘ ‘‘ADHD patient’’ or ‘‘illicit
Oakley, A. 1981. ‘Interviewing women: A drug user’’ ? Managing medico-moral mem-
contradiction in terms’ in H. Roberts (ed.): bership categories in drug dependence ser-
Doing feminist research. Routledge, pp. 30–61. vices,’ Discourse and Society 20/4: 499–516.
Patton, M. Q. 2002. Qualitative Research and Seidman, I. 2006. Interviewing as Qualitative
Evaluation Methods, 3rd edn. Sage. Research: A Guide for Researchers in Education and
Pomerantz, A. and B. J. Fehr. 1997. the Social Sciences, 3rd edn. Teachers College.
‘Conversation analysis: An approach to the Silverman, D. 1993. Interpreting Qualitative Data:
study of social action as sense making prac- Methods for Analysing Talk, Text, and Interaction.
tices’ in T. A. van Dijk (ed.): Discourse as Sage.
Social Interaction. Sage, pp. 64–91. Silverman, D. 1998. Harvey Sacks: Social Science
Potter, J. and A. Hepburn. 2005. ‘Qualitative and Conversation Analysis. Polity Press.
interviews in psychology: Problems and possi- Silverman, D. 2001. Interpreting Qualitative
bilities,’ Qualitative Research in Psychology 2: Data: Methods for Analysing Talk, Text, and
281–307. Interaction, 2nd edn. Sage.
Psathas, G. 1995. Conversation Analysis: The Study Stewart, D. W., P. N. Shamdasani, and
of Talk-in-interaction. Sage. D. W. Rook. 2006. Focus Groups: Theory and
Puchta, C. and J. Potter. 1999. ‘Asking Practice, 2nd edn. Sage.
elaborate questions: Focus groups and the Tanggaard, L. 2007. ‘The research interview as
management of spontaneity,’ Journal of discourses crossing swords,’ Qualitative Inquiry
Sociolinguistics 3: 314–35. 13/1: 160–76.
Puchta, C. and J. Potter. 2004. Focus Group Tanggaard, L. 2008. Objections in research
Practice. Sage. interviewing International Journal of Qualitative
Rapley, T. J. 2001. ‘The art(fulness) of open- Methods, 7/3: at http://ejournals.library
ended interviewing: Some considerations on .ualberta.ca/index.php/IJQM/article/view/
analysing interviews,’ Qualitative Research 1/3: 1827/3449. Accessed 22 May 2009.
303–23. Weiss, R. S. 1994. Learning from Strangers: The
Reinharz, S. 1992. Feminist Methods in Social Art and Method of Qualitative Interview Studies.
Research. Oxford University Press. The Free Press.