Tải bản đầy đủ - 251trang
5.2.3: Statement of the Basic Rules or
Guidelines for the Interview
Although you do not want to simply list a bunch of rigid
rules of conduct, you do want to establish some ground
rules around the interactions during the focus group.
You need to explain that you expect an open, polite, and
orderly environment where everyone in the group will be
encouraged to participate. If you plan to toss questions
out to the full group, to be answered by anyone, tell the
group that this will be your procedure. If you intend to
ask each subject a question in turn, obtain a quick answer,
and then open it up for discussion by the group—tell
them that this will be how the interview will proceed.
Subjects need to know what to expect. The moderator
should also reinforce to participants that everyone may
have a different opinion or answer to the questions and
that you want to hear all of these opinions. Emphasize
that no one has to agree with anyone else, or pretend
to agree with them, but that they have to let each other
speak before voicing any disagreement. It is also a good
idea to point out any recording device and its purpose, if
one is present in the room. If the session is being recorded
by a hidden camera, this too should be indicated to the
group, as well as why the camera is not in the room. For
example, you might tell the group that the camera is hidden to avoid making them feel self-conscious, but that
they should be aware of its existence.
5.2.4: Short Question-and-Answer
Most focus groups operate with a short series of discussions, sparked by questions asked by the moderator
(Krueger, 1997; Krueger & Casey, 2000). These questions
could be written out and listed in a similar manner to
a semistructured interview guideline document (see
Chapter 4). You may even plan out intentional probes to
be used to facilitate more information in the event that
there is little discussion after asking the initial questions.
Experienced moderators are likely to deviate from such a
schedule, as the dynamics of the group begin to animate
the focus group experience, giving it a kind of life of its
own. Less experienced moderators, however, may feel
more secure having a script of questions to ask, which is
fine as long as the script does not inhibit participation from
5.2.5: Special Activities or Exercises
Although many focus groups restrict their data collection
to responses from a series of questions, some, especially
those undertaken with children, may include drawing or
role-playing exercises so that subjects may better express
Focus Group Interviewing 97
their views (Wright, 1994). It may also be helpful to the
researcher to have a pencil-and-paper exercise to help validate the verbal responses that children are likely to offer
(Wright, 1994). The major consideration for thinking about
the inclusion of various additional exercises is age and
maturity of subjects.
Exercises and activities also allow the moderator to
determine what subjects individually know or believe
without the influence of others in the group. One useful
strategy is to have the subjects fill out a brief pencil-andpaper survey that is administered before the actual question-and-answer/discussion segment of the focus group
begins (discussed later as an extended focus group). This
prefocus group activity allows participants to think about
and perhaps commit to certain ideas and attitudes about
topics to be discussed during the group session (Wimmer
& Dominick, 2006).
Focus groups invite people to sit together and discuss
issues, events, and ideas. But not everyone in or out of the
group will be equally informed about the issues or events,
and it is difficult to assess or make use of different degrees
of knowledge beforehand. On the other hand, if the discussion is to focus on something relatively tangible, such as
issues raised in a particular film, or something that can be
made tangible, such as forms of advertising for which the
researcher can provide examples, then it is often useful to
start the session by watching parts of the film or reviewing other materials together. This shared experience gives
everyone a common frame of reference from which to discuss the research questions.
5.2.6: Guidance for Dealing with
As in any interviewing session, focus groups require the
moderator to use sensitivity when dealing with certain
subject matters. These typically include questions concerning alcohol or drug use, stigmatized behaviors, grief and
loss, and certain mental health issues. In the focus group,
one way to approach such sensitive issues is to begin with
a general question for discussion that deals with the subject matter. For instance, let’s assume you are interested in
knowing about cigarette use by Asian American teenagers.
Rather than saying immediately, “Tell me about your cigarette smoking habits,” you might begin with a question
such as, “What do you think about cigarette smoking?” In
some cases, this slightly broader question may open the
door for discussion in the group about individual participants’ smoking habits—but without having placed anyone
on the spot. If this does not occur, the more specific question may subsequently need to be asked.
Sensitive topics may also appear less threatening to
participants when activities and tasks are incorporated
into the focus group session. Among the activities one
98 Chapter 5
might use are free listing, rating or ranking of things being
discussed, pictures to stimulate conversation, storytelling, projective techniques, and even some role-playing
(Colucci, 2007). Bloor, Frankland, Thomas, and Robson
(2002) suggest that focus groups may, in fact, be ideal situations for discussing sensitive topics, particularly when in
the presence of friends, and colleagues, or with others similarly situated or involved in sensitive activities to themselves. Among people who share some challenge, stigma,
or loss, the group discussion can become welcoming and
even therapeutic, even when the moderator lacks the personal experience or deep knowledge shared by the group.
Thus, focus groups may work particularly well, under certain circumstances, for addressing sensitive topics.
It bears repeating that no sensitive or threatening topics may be dropped on an unsuspecting group. Informed
consent procedures require researchers to alert participants
of the kinds of topics that will be discussed before they
agree to join, and moderators must remind participants
that they can withdraw if the process becomes uncomfortable. These are important technical procedures. As well,
a prepared researcher will try to ensure that the group
composition does not itself invite threatening conditions
(though disagreements are fine). For example, the suggestion above that group members might feel supported if
the other participants shared their particular challenge or
stigma only works if the whole group is more or less on the
same side of that issue. You can combine people who support “enhanced interrogation” (torture) techniques with
those who oppose them. That’s opinion research. You cannot combine people who support these techniques with
those who have been tortured.
5.3: Focus Group Data
Explain how moderators use the nonverbal data
Observations give you data about people’s actions. With
focus groups or interviews, you can only ask about actions.
Such self-report data gives you stories about behaviors,
not actual behaviors. If you are interested in observing
behaviors and meanings as they emerge in their natural
setting, you may find that the simulated conversations of
focus groups are insufficient compared with traditional
forms of participant observations and various sorts of field
ethnography. More powerful still is the two techniques
in conjunction. One can observe natural interactions in a
formal or informal setting, and then bring the participants
together to discuss that setting and their roles in it. We
assume that the researcher and the participants will not
interpret the participants’ actions in exactly the same way,
but we rarely have the opportunity to discuss this behavior with them unless we use multiple methodologies.
Focus group data reflects the collective notions shared
and negotiated by the group. This is very different from
individual interview data, which reflects only the views
and opinions of the individual, shaped by the social process of living in a culture. The group data is based on interactions, cross-conversation, negotiation, confrontation,
and collective decision processes. Participants, one hopes,
do not merely answer questions when asked, but actively
explain themselves to each other. It is the desire among
group members to make themselves understood by others
in the group that yields the richest data.
Focus groups are therefore appropriate for measuring meanings, which are otherwise difficult to get at.
Focus groups are also well suited for studying attitudes,
preferences and priorities, and beliefs. And more than
any other method, they enable us to study participants’
rationalizations and justifications. Many methods can
be used to measure people’s opinions. But focus group
research challenges study subjects to explore and even to
defend these opinions.
The information obtained from focus groups provides
elements of data similar to those of traditional interviewing, direct observation, and even certain unobtrusive measures commonly used in qualitative research. Yet, I caution
that focus group data does not actually offer the same
depth of information as, for example, a long semistructured interview. Nor does it provide as much rich observational data as one might obtain, for instance, by observing
a class of sixth graders on the playground over a period of
Sussman et al. (1991) found that subjects’ responses
tended to be more extreme in focus groups when compared to responses offered in survey questionnaires. Taken
together with Fern’s (1982) earlier work, this suggests that
an interviewer must be willing to give up some degree of
data precision in exchange for the interaction experience.
Another interpretation is that people’s actual opinions
might be more extreme than they normally want to let on,
but when challenged, they will let loose.
The information collected during the course of a focus
group, like that collected during the course of a face-to-face
interview, is raw data. The researcher’s job is to prepare an
analytic statement based on this collected raw data. Ideally,
this assessment should be thoroughly grounded in the
data. The first step is to transcribe the entire discussion.
This should be a verbatim transcription of each question
asked by the moderator and each individual answer given
by the focus group participants. It should include all probes
asked by the moderator and various group members. It
should also include any slang, dialects, or pauses offered
by focus group members as they respond to the moderator and each other. Transcripts can also be annotated with
the researcher’s notes concerning the participants’ behaviors during the discussion. At what point did voices rise?
When did members try to interrupt one another? Were
there physically aggressive or intimidating acts? What
about nodding and other nonverbal indicators?
Subjects in focus groups may use body language,
gestures, or other nonverbal clues to encourage, or intimidate, others while they are speaking. When a participant
suddenly breaks off a comment, or shifts into a more
confessional tone, your notes should indicate anything
significant happening in the room that might have precipitated the change. As with all forms of interviewing,
of course, body language is not our data. However, it is
always worth noting if something nonverbal impacts the
flow of conversation. Even with such observations, your
notes will represent only a small portion of the basic verbal
data typically collected during a focus group interview.
Videotaping helps, when possible.
During the course of the focus group, either the moderator or a second observer working with the researcher
should take copious notes. Taken together, the transcription and the observer notes provide a complete record
of the discussion that unfolded during the focus group
interview and will assist in analysis of this data. The
next step is to analyze the content of the discussion to
identify trends and patterns that reappear either within
a single focus group or among a series of focus groups.
Thus, the researcher undertakes a variation of content
analysis (discussed in detail in Chapter 11), which begins
by examining the text for similarly used words, themes,
or answers to questions. Some system of indexing and
retrieval of these terms and patterns must be used (see
Chapter 11 again). The researcher should additionally consider the emphasis or intensity of respondents’ comments
(sometimes illustrated in observer notes). As well, the
researcher should consider the consistency of comments
and responses to probes both within a given focus group
and across a series of focus groups.
Traditional interviewing approaches sacrifice the ability to observe interaction for greater amounts of detail on
various attitudes, opinions, and experiences. In many ways,
it is the very give-and-take interactions characteristic of
focus group interviews that lead to spontaneous responses
from session participants. Hearing how one group member
responds to another provides insights without disrupting underlying normative group assumptions. Meanings
and answers arising during focus group interviews are
socially constructed rather than individually created. They
also emerge from the participants’ interests rather than the
researchers, which improves the validity of the data. As
Rubin and Rubin (1995, p. 140) explained:
In focus groups, the goal is to let people spark off one
another, suggesting dimensions and nuances of the original problem that any one individual might not have
thought of. Sometimes a totally different understanding of
a problem emerges from the group discussion.
Focus Group Interviewing 99
Because interactions between group members largely
replace the usual interaction between interviewer and subject, greater emphasis may be given to the subjects’ viewpoints. As with informal interviewing, focus groups can
sometimes be undertaken without preconceived questions,
focus questions, or guidelines (Morgan, 1997). This can
effectively eliminate the researcher’s perspective from the
resultant data. Conversely, should more guided responses
be desired, focus group interviews, like individual ones,
can be made more formal and structured.
David Morgan (2002, p. 148) offers a description of
what he terms his ideal focus group, which demonstrates the
difference in flow between a face-to-face interview and a
focus group session:
The ideal group would start with an opening question
that was designed to capture the participants’ interest,
so that they themselves would explore nearly all the
issues that a moderator might have probed. Then, just
as the allocated time for that question was running out,
one of the participants in the ideal group would spontaneously direct the others’ attention to the topic for the
second question by saying something like, “You know
what really strikes me is how many of the things we’re
saying are connected to….”
I would add that my ideal group would also contain just
enough disagreement among participants that they constantly
feel the need to explain themselves very carefully. Naturally,
such an ideal type of focus group is unlikely to unfold.
However, Morgan’s illustration should serve as the model to
strive toward in undertaking focus group interviews.
The analysis of focus group data must take into
account both the individual responses and the group
interaction. As with any content analysis, we look at patterns in terms used, ideas expressed, associations among
ideas, justifications, and explanations. With focus groups,
however, we also need to examine the flow of ideas
throughout the group. The analysis needs to attend to consensus, dissensus, and resonance (Lune, Enrique, & Koppel,
2009). Consensus refers to points of agreement within the
group. Did certain suggestions, ideas, or explanations go
entirely unchallenged? Did the ideas recur among different speakers’ responses? These ideas represent general
points of agreement within your study sample. Different
participants may have entered into the discussion with
different ideas, but they have come to agreement around
specific and therefore important ideas or statements.
What about ideas or suggestions on which the group
could not come to agreement? Are there points of disagreement where compromise or flexibility seems impossible?
These ideas have stronger weight with the respondents
than ideas around which they are willing to shift their
positions. Within these points of dissensus or disagreement, were just a few people arguing or did most of the
group divide into opposing camps? A careful researcher
100 Chapter 5
must also distinguish between topics of disagreement (for
or against, for example) and reasons for the disagreement.
Often what appears to be a dispute about an event or idea
turns out to be a dispute about the definition of the event
or idea. One person may support some new legislation
because they agree with its goals while another opposes it
because they worry about the legal precedent in its wording. Are these people really on opposite sides?
And finally, resonance. Do certain expressions seem
to “catch fire” within the group? Are there moments in the
discussion where one participant expresses an idea that
suddenly unites all (or nearly all) of the members? These
are ideas or expressions that resonate within the study
group and which may have a powerful influence on the
thoughts or feelings of your study population in general.
Points of agreement and disagreement stand out for
obvious reasons. But we also want to measure the intensity of the feelings or expressions associated with them.
Disagreement can be polite, tense, threatening, angry, or
even abusive. Where and how do participants fight, if
it comes to that? Related to this, though it is difficult
to record the absence of something happening, we need
to note when and why some participants went silent.
If something in the group process leads to someone or
several group members choosing to withdraw, that must
have been a significant moment. Yet, since no one has mentioned it, one can easily overlook such moments.
There are several important rules of thumb for analyzing
focus group data, which are quite different from the analysis
of other textual data such as field notes or interview data:
• Avoid quantifying results or offering magnitudes; just
because four of seven group members made a statement
does not mean that 57 percent of the subjects agree on
that statement. Such an assumption is meaningless and
is not a finding in itself.
• Provide quotations to support your assessment of what
the various trends and patterns of discussion are. It is
not necessary for all participants to signal agreement
with a statement for us to observe a tendency toward
support for that statement.
• Offer relevant characteristics of each group member prior to offering their quoted responses in order
to provide a sense of three-dimensionality to group
members (e.g., a 26-year-old single Latina mother of
two stated, “…”). The operative phrase is relevant
characteristics, not random demographics.
• Make a point, or state a specific pattern, before offering
quoted materials intended to demonstrate the point or
• Use quotes to illustrate, not to prove. It does matter that
everyone in your group prefers the yellow box over the
green one, but it does not prove anything.
5.4: Selecting Focus
Groups as a Method
Evaluate the effectiveness of using focus group
techniques in research
Focus groups may be used either alone as a data-collection strategy or in combination with other techniques.
In their simplest form, focus group interviews can be
used as a sort of stand-alone data or primary data. As
suggested in the first chapter of this book, triangulation in qualitative research can be important to issues
of validity. Whenever you can demonstrate corroboration of information you have obtained, you are on solid
ground. One excellent use of focus group research, then,
is to probe the validity of results that were suggested by
other means. Although this is not its primary purpose,
this additional line of action may, in fact, offer either corroboration of other data or insights into areas other data
fail to illuminate.
You might consider using a version of focus group
interviews to pilot an interview schedule. In this instance,
you could have members of the focus group read through
the instrument under consideration. Next, the group
would discuss the usual concerns researchers have about
such research instruments: the level of language, comprehensibility of the questions, question order, affected
wording of questions, and so forth. A more interesting and
promising test would be to ask the group the questions
from your interview guideline, exactly as written, and see
whether everyone in the group understood the question in
the same manner or not.
As a primary means of data collection, focus groups
offer exceptional opportunities for the deep study of difficult subject matters. The interactive element allows participants to reflect on each other’s words, add to them, pick
up loose threads of their ideas, or even challenge them
directly. In this fashion, what might have been a passing
comment in an individual interview may be recognized
by other participants as significant to their shared experiences, and drawn out. Even a highly skilled interviewer
is likely to be surprised by the nuances of what his or her
subjects perceive that we do not.
Focus groups comprised of individuals with shared
experiences may, under the best of circumstances, become
supportive and empowering to the participants. The stories
that they choose to share with each other would be different from those that they would be likely simply to offer to
a researcher. Denzin (1989, p. 39) has suggested that biographical experiences have effects at two levels in a person’s
life: the surface level and the deep level. On the surface level,
effects may be barely felt or noticed. They are often taken
for granted and are nondisruptive. Picking up a container
of milk on the way home from work might be an example.
Effects at the deep level, however, strike at the core of an
individual’s life. They have a strong hold over us as individuals and affect how we behave, think, and understand
things. Acceptance of our sexuality, self-hate, grief, and other
deep-rooted epiphanies serve to illustrate deep-level life
structures. Focus groups, in some cases, provide avenues to
understand a variety of deep structural elements.
5.5: Selecting Groups
Identify the challenges of assigning subjects to
different groups according to conceptual categories
Most researchers who use focus group techniques
acknowledge that group influences can distort individual
opinion. Some opinions may be more extreme and some
may be less verbalized than others because of the group
effects (Sussman et al., 1991). Having some idea about
how individuals thought about certain topics before the
group sessions start allows the investigator to gauge this
group effect. This is not to say that material obtained during the group session is false. Quite the contrary; the opinions voiced during the session, even those that contradict
pregroup questionnaires, merely demonstrate the impact
of group dynamics. Additional information, confirmation
or refutation of beliefs, arguments, discussion, and solutions heard during the group session shape participants’
thinking. What results is a collective understanding about
issues discussed by the group. If only for that reason,
selecting and recruiting your group participants has all of
the normal challenges of sampling, and more.
When you design a focus group interview study,
your plans for participant selection must be undertaken
very carefully. Even among marketing researchers, care
is required to create samples that include subjects with
necessary product user characteristics (Tynan & Dryton,
1989). For the more traditional social sciences, one should
begin using standard strategies for sampling to create a
theoretically and experientially appropriate sample pool.
From this pool, the smaller focus groups may be formed.
Focus Group Interviewing 101
For example, let us say you are interested in studying some aspects of the lives of incarcerated women.
Perhaps, you want to know how these women perceive
their family role as mother, even though they are separated from their children (Moloney, 1997). In most states,
there are few women’s correctional facilities, often only
one or two for the entire state. Thus, you easily can begin
with a census sample2 of women in prison to form the initial pool. Next, you might stratify this group into those
who have children currently of juvenile status (under the
age of juvenile jurisdiction) and those who do not. Using
the group with children, you might now have a sample
of 50 or 60 women. Assuming no rejections, you could
randomly assign women in this group to five or six focus
groups and conduct sessions in a fairly brief amount
Of course, when your question requires such a particular background, your options for finding subjects are
highly constrained. But what if you want to research a
question with a broader applicability? For example, we
might want to study something about attitudes toward
the criminal justice policies that directly impact incarcerated mothers. In that case, we would want a sample pool
of people who are not incarcerated, and possibly only
those who never have been. This raises a number of further questions about who should be included or excluded
to represent the study population. Should we include
those who have friends or family in prison or not? Should
we separate the men and women into different groups?
Should parents and nonparents be interviewed separately?
These are questions of research design. For purposes of
assembling an appropriate study group, we are guided
by two overall decisions: what identifiable factors might
shape a subjects’ participation in the discussion; and when
do these factors have to be held constant within a group?
Figure 5.1 demonstrates how participants may be recruited
into a study through relevant characteristics or experiences, and then stratified into different groups for a more
focused discussion among subgroups.
You can develop focus groups using other strategies
to create the initial sample pool. This is particularly true
Figure 5.1 Berg Sampling Strategy