PSYCHOTHERAPY RESEARCH METHODS
Qualitative research interviews
SARAH KNOX & ALAN W. BURKARD
Department of Counseling and Educational Psychology, Marquette University, Milwaukee, Wisconsin, USA
(Received 2 June 2008; revised 15 December 2008; accepted 17 December 2008)
Abstract
After presenting a brief overview of the complexity of the qualitative interviewing process used by psychotherapy
researchers, the authors discuss some of the major ideas that psychotherapy researchers using such interviews must consider
both before and during the interview process. They then offer thoughts regarding approaches to strengthen qualitative
interviews themselves.
Keywords: alliance; process research; psychotherapist training/supervision/development; qualitative research
methods
Much of qualitative psychotherapy research relies on
spoken interviews with participants to gather de-
tailed information regarding the phenomenon under
examination (Polkinghorne, 2005). In an activity
that calls on not only strong interviewing techniques
but also the very skills used when working with
clients, interviewers confront challenges inherent in
both domains: How do they conduct an incisive
interview that yields rich and meaningful data while
simultaneously helping participants feel safe enough
to explore in depth often difficult experiences with
a relative stranger? Perhaps complicating this pro-
cess, qualitative psychotherapy researchers also must
attend to the ethics of interviewing. (The ethics of
interviewing are beyond the scope of this article, but
interested readers are encouraged to see Haverkamp,
2005.) Such researchers, for instance, have often
been trained, and may even be credentialed, to
address others’ distress. When conducting research,
however, they tread a sometimes difficult line
between interviewer and therapist, an ethical chal-
lenge that other social science researchers may not
face (Haverkamp, 2005). In this article, we discuss
important considerations that psychotherapy re-
searchers must address, both before and during the
interview itself, as they engage in this approach to
data collection. We do so in the hope that our
discussion of these vital components of qualitative
interviewing will not only improve researchers’
execution of such interviews themselves but will
also strengthen qualitative research more broadly.
When possible, we integrate extant empirical evi-
dence and relevant theory and conclude by suggest-
ing fruitful research avenues for advancing our
understanding of the qualitative interview process.
We acknowledge, as well, that our focus is not
exhaustive: There are certainly additional topics
worthy of consideration, but we have included those
that have consistently been of most relevance in our
own research.
Considerations Before the Interview
Interview Protocol
Before any interview can occur, consideration must
be given to the very questions that will be asked,
because "at the root of . . . interviewing is an interest
in understanding the experience of other people and
the meaning they make of that experience" (Seidman,
1991, p. 3). The means to access those experiences
range widely, from open-ended, unstructured
Correspondence concerning this article should be addressed to Sarah Knox, Department of Counseling and Educational Psychology,
Marquette University, 114 Schroeder Health Complex, Milwaukee, WI 53201, USA. E-mail: sarah.knox@marquette.edu
Psychotherapy Research, July
September 2009; 19(4
5): 566
575
ISSN 1050-3307 print/ISSN 1468-4381 online # 2009 Society for Psychotherapy Research
DOI: 10.1080/10503300802702105

approaches that may seem more a friendly conversa-
tion than a data-gathering interview (Seidman, 1991)
to highly structured protocols with preset and stan-
dardized questions from which there is little variance.
On one end of this continuum, then, are relatively
unstructured approaches (e.g., ethnography,
grounded theory, phenomenology) that may use an
evolving set of questions, such that later participants
respond to queries quite different from those to
which earlier participants responded. As initial data
are gathered and analyzed, they lead to refinement of
the study’s central focus and thus to new questions
for participants (Glaser & Strauss, 1967). Such an
approach is in keeping with the sentiments of Kvale
(1996), who asserted that the design of qualitative
interview research is open ended in that it is more
concerned with being attuned to the participant than
with necessarily following the same path for all
respondents. In ethnography, for example, the inter-
view is more a ‘‘friendly conversation into which the
researcher slowly introduces new elements to assist
informants to respond’’ (Spradley, 1979, pp. 58-59)
and thus retains an open framework with little in the
way of preset queries. The basic themes or topic
areas of the investigation are likely determined ahead
of time, but not the sequence or the content of the
specific questions. As stated by Kvale (1996),
‘‘Sometimes only a first, topic-introducing question
is asked and the remainder of the interview proceeds
as a follow-up and expansion on the interviewee’s
answer to the first questions’’ ( p. 127). Unstruc-
tured interviews, although they may well yield
unexpected responses (Kvale, 1996), also make it
difficult to compare findings across cases if partici-
pants have not responded to the same questions.
Occupying the middle of the continuum are
semistructured interviews, in which a protocol using
open-ended questions based on the study’s central
focus is developed before data collection to obtain
specific information and enable comparison across
cases; interviewers nevertheless remain open and
flexible so that they may probe individual partici-
pants’ stories in more detail (DiCicco-Bloom &
Crabtree, 2006). The interviewer thus asks all
questions of each respondent but may pursue in
more depth particular areas that emerge for each
interviewee (Hill et al., 2005; Hill, Thompson, &
Williams, 1997) and may also vary the sequence in
which questions are asked. The protocol in such
semistructured interviews serves as a guide (Flick,
2002), a foundation on which the interview is built
but one that allows creativity and flexibility to ensure
that each participant’s story is fully uncovered.
Finally, at the other end of the continuum are
survey or standardized interviews, in which the goal
is to expose each participant to exactly the same
interview experience (Fontana & Frey, 2005) so that
any differences are assumed to be due to variations
among participants rather than to differences in the
interview process itself (Singleton & Straits, 2002).
To this end, such interviews follow a highly struc-
tured protocol consisting most often of closed
questions (those that seek a definitive one- to two-
word answer such as ‘‘yes’’ or ‘‘no’’ and are often
used to ascertain facts) presented to respondents in
the same order. Furthermore, the interview process
itself is highly regulated (e.g., questions are read
exactly as written, standard probes are used, no
interviewer disclosure is to occur), such that re-
searchers are neutral and consistent throughout all
interviews (Fontana & Frey, 2005). In effect, then,
‘‘the goal is nothing less than the elimination of the
interviewer as a source of measurement error’’
(Groves, 1989, p. 358). Wholly standardized inter-
views have the potential advantage of greater uni-
formity across respondents but inhibit the
uncovering of participants’ rich and unique experi-
ences, especially those that lie outside the bounds of
the interview questions themselves.
Phone Versus In-person Interviews
Another decision that qualitative interviewers face
involves the actual means of completing the inter-
view: Should participants be interviewed by phone
or in person (i.e., face-to-face)? Little research has
compared the benefits of these means of data
collection, likely because, according to Shuy
(2003), such studies are expensive and difficult to
carry out, and few researchers have been motivated
to examine the relative merits of the differing
approaches. Two studies that did examine phone
versus in-person interviews found a slight advantage
for the latter in yielding better quality data (de
Leeuw & van der Zouwen, 1988; Jordan, Marcus, &
Reeder, 1980). In a third study, a meta-analysis
focusing on participants’ responses to sensitive
topics in surveys, Tourangeau and Yan (2007) found
that interviewers contribute to participants’ misre-
porting because respondents have to share their
answers with another person (vs. with a computer
or only with themselves [as in a written survey]), and
that social desirability bias is worse in phone than in
face-to-face interviews.
Despite the potential for such bias, phone inter-
views are quite common. First, they enable research-
ers to include participants from virtually any
geographic region; no one is required to travel for
the interview. The ability to cast this broader net may
be quite attractive to researchers who seek an efficient
Qualitative interviews 567