Electronic Textbook: "He-Said-She-Said:" Focus Groups Interviewing: Part One

"He-Said-She-Said:"
Focus Group Interviewing: Part One

We're right in the middle of our "interview trilogy," my friends! Last time out we looked at interviewing in general, and some ways to vary the interview situation. We also took a look at the broad spectrum (time- and topic-wise) of developing "good" interview questions.

This time out, I'd like to key in on a particular type of interview setting or situation: the focus group interview. Originally developed within the boundaries of marketing research, it has deservedly crossed over into the realm of educational and other types of social-behavioral research. It is a popular choice of data collection procedure for dissertations, and there are plenty of outstanding examples. Because of its relevance and popularity, I want us to take a good, closeup look at the technique.

We'll start with a bit of historical perspective on focus group interviews. Where did they originate? What were they used for? Then we'll re-examine the sampling issue: how do we select our focus group participants? This will give us the background on what a focus group interview is, and with whom it is done. Next time, we'll look at the how: planning the interview itself, and asking the questions.

All of our material for these next two chapters comes to us courtesy of Richard Krueger's outstanding and most readable 'primer,' Focus Groups: A Practical Guide for Applied Research. The second edition was published in 1994 by Sage Publications, Inc. Krueger's book, like Robert K. Yin's, was a 'breakaway success.' Krueger in essence 'bridged the transition' into showing educational researchers how to adapt this technique from marketing research and use it for their own purposes. It covers the gamut of topics and is a most useful reference!

I. A Brief History of Focus Groups

As we saw in the last topic, a relative advantage of interviewing over other methods is the opportunity for the subject to try and articulate his/her perceptions directly to the researcher. During the 1930's, social-behavioral researchers examined existing instrumentation and found it wanting. Specifically, surveys tended to be closed-ended in nature -- with the implication that all of the choices that respondents would make can be predetermined by the researchers -- and thus too "outside-researcher-driven."

As a result, nondirective and other types of open-ended interviewing quickly came into vogue in the 1930's and 1940's. Qualitative interviewing procedures began to be applied in organizational behavior research.

The seminal work regarding focus groups was published in 1956. This book is entitled: The Focused Interview, by Robert Merton, Marjorie Fisk and Patricia Kendall.

Up until recently, the primary domain for focus groups has been in the area of marketing research. You may have occasionally read or heard about a product being created or modified on the basis of "focus group interviews." This need arose from the realization that it is virtually impossible to create, advertise and sell a product that will have equal appeal to everyone. People's tastes and needs vary; furthermore, advertising is very costly. Rather than try to be "all things to all people," it's far more efficient for product developers and vendors to "shotgun" their product (including advertising for it) towards a smaller, more specialized "target market" of consumer for whom that product may have particular appeal - say, college-educated women, employed outside the home, living in an urban area, in the 25-39 year old age range. This in turn helps to similarly focus product features and advertising characteristics for this particular group. This attempt to 'match up' consumer demographics and other information with equally focused features of the product and its advertising strategy is known as market segmentation. Such a pre-identified subgroup (blast from the past: Intro to Research and Research Design fans will recognize this as a 'stratum!' from population and sampling!) It then becomes necessary to identify and 'understand' the characteristics and needs of that market segment in order to devise the optimum "fit" of product features, advertising strategy, channels of distribution of the product, and so forth. In sum, the focus group can reveal, "How do these people think and talk about Product X?" (or their needs as related to how a Product X might fill those needs)

For insight into how business uses focus groups, open the links below.

Focus Groups in Marketing

Well ... you might be thinking ... so far this sounds like a Marketing 101 class! I'm not here to earn an M.B.A.; I'm an educator! So - what's in it in focus group interviewing for me?! Why would a focus group fit the needs of what I want to research in the educational setting?

Here an example of using focus groups to get at an educational question.
University of North Carolina at Pembroke - Study of Undergraduate General Education

- > Ah, but that leads us very nicely into the main property of focus groups and how they function! And incidentally, why "crossover" books such as Krueger's have been a phenomenal success - why educational researchers believe they've 'discovered a goldmine' in the form of this particular method of conducting interviews!

 

 

II. The Nature of Focus Groups

At this point I'd like to share with you a visual that I use when I do workshops on focus groups. We'll then proceed to take a close-up look at some of those "pieces!"

 

Table 1.

WHAT IS A FOCUS GROUP?

 

*	A way to get information about ATTITUDES, FEELINGS & EMOTIONAL REACTIONS
*	Get BOTH individual and interactive opinions (how your subjects REACT TO ONE ANOTHER)
*	Record BOTH WORDS AND NONVERBAL BEHAVIORS
*	Done in a RELAXED SETTING
*	Done with 4-12 SUBJECTS AT A TIME
*	Interviewer asks up to 10 OPEN-ENDED QUESTIONS ("questioning route") with follow-ups ("probes") for clarification if necessary
*	One interview session lasts no longer than 1 1/2 hours
*	"There are NO RIGHT OR WRONG ANSWERS -- ALL IDEAS AND OPINIONS ARE IMPORTANT!"

Given the above key features, let's now take a more in-depth look to see: 1) what distinguishes focus group from other types of interviewing; and 2) why this particular form is so relevant to many educational research settings.

*	Attitudes, feelings and emotional reactions. These are indeed the "stuff" of many so-called cognitive decisions! How well a child supposedly 'performs' on an "academic" measure such as the ITBS test may be heavily influenced by such factors as his/her prior experience with (and possibly fear of) the test-taking situation, for instance. The very decision to become an educator may be laden with a 'passion' for the educational process - a desire to make a difference in children's lives. Administrators need not be reminded how the emotional reactions and attitudes of school board members, parents, community members, etc., can shape day-to-day policy and operating decisions! The affective component is critical in educational settings, and focus group interviews are centered on it.
*	Interaction in the group setting. As Krueger eloquently reminds us, not only is the affective element important - but also the group setting itself. Many decisions are made in a group setting. We look to our friends and colleagues for their opinions on anything from political election outcomes, to a shift in leadership at the work place, to a major personal life change that we may be contemplating. Furthermore, we often shift opinions and attitudes - change our minds - as a result of others' thoughts, feelings and inputs. Thus, the group setting, with its ebb and flow/changes in direction of opinion and attitude, may be a more realistic way to capture "how folks think and talk about an issue" than the isolated, one-on-one interview (where it's just a subject at a time, and you as the researcher). This is why Krueger has stated that focus groups inherently have high face validity. If you set a comfortable enough tone, you may indeed closely reproduce the 'give and take' of a group discussion as it might happen in real life.
*	Verbal and non-verbal data. Given the interactive nature of the group setting, as explained above, you as the researcher can gather a wealth of verbal and non-verbal information. He says something and she snickers. Another individual responds, and everyone falls silent. The subject in the corner makes a face and rolls his eyes. The woman sitting next to him makes a flippant remark, which appears to break the tension - at the end of her observation, most subjects relax their shoulders, and smile. This is, in essence, the "best marriage" of both observational and interview procedures. Not only can you gather such nonverbal cues, but you can gather them as subjects react to one another as well as to you, the interviewer.
*	Relaxed setting. We've spoken of the way in which group interaction is more realistic and thus 'face valid' in terms of how attitudes, feelings and opinions are formed and altered. On the other hand, all research is 'intrusive' to a degree - you run the risk of being perceived as an 'outsider' and thus getting somewhat artificial reactions from your subjects. In order to "reproduce this realism" as far as possible, you need to make the situation and setting as comfortable as possible for your participants. This means giving careful thought to where the interview will take place, and under what circumstances. A round table with comfortable chairs, where everyone can see one another, is far preferable to the traditional classroom desks-in-rows arrangement - with the attendant implication that you as the interviewer are "in charge" by being in front of the room, and those with desks up front are somehow "more important" than those seated in back. Remember: a sign of success here is if/when you reach the point where you are doing virtually none of the talking! But rather taking careful note of how they respond to one another! Another way of establishing a comfortable climate is to serve a meal or light refreshments: anything from "coffee-and" at breakfast time, to lunch at noon, to a fruit tray or similar refreshing snacks in the afternoon. Some of you may worry about that being perceived as a 'bribe.' Instead, it may be a practical necessity: for instance, to get that many traditionally busy professionals together, you may have to realize they are giving up their lunch or dinner hour to accommodate your request for the interview. Thus, it is seen as a courtesy - a gesture of appreciation for their valuable time - for you to provide the meal. Also, the social nature of providing food will allow you to sort of 'scope out' the participants if you set out the meal or snack beforehand. You may observe how people "meet and mingle" and identify those who seem shy, domineering, etc. This is valuable "human research!" and will help you later in "managing the traffic flow" of the interview itself!
*	Four to twelve subjects as the 'ideal' group size. This range has been confirmed time and again in all sorts of simulation studies. Get fewer than four and you don't really have enough participants to get a sense of 'group interaction.' That is: you might as well be doing individual interviewing because they probably won't 'play off' one another as much. Get more than twelve and - as every 'overloaded seminar' teacher knows! - they will no longer 'center back' on you as the central interviewer but will instead break off into several "subfactions." There probably won't be a single cohesive group, but rather, isolated cliques. If you have more than 12 subjects that you need or want to interview in a group setting, your best bet would be to plan for more than a single focus group interview session - i.e., break them up into 2 or more separate focus groups. One other comment: Krueger refers to fewer than six as a "mini-focus group." As we learned in Intermediate Statistics with "small" samples regarding inferential test statistics, we do run some risk of "atypicality" due to the "small" sample size. Nevertheless, it's accepted that 4-5 in a group can work reasonably well, especially if careful thought is given to subject selection.
*	kli,. There is nothing inherently "magical" about the number "10." It's more of a ballpark figure: perhaps a reminder to plan some general areas but also allow time for more specific follow-ups, called nhj, along with most of the general questions. There is a marked tendency (especially among novice, eager interviewers!) to "over-plan" and run out of time before they cover the gamut of their questions. (And as we'll see from the following point, the time frame itself is rather limited!) Such a set of planned interview questions is called a focus group questioning route, interview protocol, or interview guide. The last label itself is an important reminder, too, that these are exactly that: a general guide as to the main issues and follow-ups that you want to cover. It is important to learn to "go with the flow," and be prepared to take off in an unplanned direction as a result of a spontaneous comment, if you make an on-the-spot assessment that, "Hey, I never thought of that but it does relate! Let me pursue what he/she has just brought up" -- i.e., the serendipitous question. And that brings us back full circle to the very reason for being of focus groups identified earlier in : the dissatisfaction with the closed-end, overly preplanned nature of existing instrumentation in the 1930's. Not only is it important to keep them open-ended; but the interviewer must try and be confident and prepared enough to 'follow down an unexpected, but lucrative, path of the actual discussion.'
*	Maximum 90-minute interview session. This too has been repeatedly confirmed, empirically, from many different focus group simulations. There is inevitably a 'slow warm-up,' perhaps an initial shyness or hesitancy to respond to the first question or so. (Be prepared for this, so that you don't 'take it personally,' panic, etc.!) But trust me - once you get 'em going, and particularly as they play off one another, it's often hard to shut the session down! As with group meetings in general, there comes a point of fatigue, burnout, etc. Thus, it is recommended that you run a single session for no more than 90 minutes (and you may even need to keep it to an hour). A good way I've found, if you feel you'll run out of time before you've covered everything you want to, is this. Try and get the subjects' 'buy-in' to come back for a second session before the beginning of the interview, if possible. Explain that, due to the number of topics to be covered, and the fact that you know their current time is limited, would they please be willing to pull out their appointment books and schedule a 2nd session in case it is needed? It might turn out that you do indeed cover everything in that one session. But if you don't, you can proceed with confidence at the optimum pace, knowing that you've already gotten their prior OK to come back if you don't finish today.
*	"No right or wrong answers!" This one also comes full circle, to the very first point in this discussion. If you do hope to attain a 'true, close-up' look at subjects' inner attitudes, emotions, and feelings, you must indicate that it is indeed safe for them to share with you candidly. We are all aware of the "halo effect" and similar data collection biases, where subjects try to discern what answers or directions of discussion you, as the interviewer, "seem to like" - and then give you their responses accordingly, as opposed to their true feelings. Thus, it is advisable that you make this point clear at the outset of the interview - ideally, as part of your own introductory remarks - and be prepared to reiterate it as necessary. A secondary comment here is that you need to be careful of the 'supplementary cues' that you might be giving off in response to the comments. If you reinforce certain comments with remarks such as "good," excessive nodding, or similar verbal and nonverbal indicators of agreement, you run the risk of pre-biasing the subsequent direction of discussion. To be maximally effective, you need to try and adopt what some experts have called, "friendly neutrality:" pleasant attention to what is spoken; non-threatening eye contact; comments such as "I see," "thank you for sharing;" and the like. The idea here is that: you want to hear it all, whether "it" is perceived as "good, bad, positive, negative," etc. by the speaker. You are not there to "judge" those comments, but rather to receive them openly. Finally, you want to make clear that there will be no negative repercussions of any sort for sharing a given idea, attitude or opinion.

 

 

II. Selecting Focus Group Participants

We've talked about the desired size of group: 6-12, with 4-12 as the ideal "outer range" or limit. Now, what about the participants themselves?

Several rules of thumb, again from extensive empirical application of focus group interview procedures to a wide variety of settings, seem to apply:

 

*

Pre-target the general characteristics you want your subjects to have. This goes back to the idea of "market segmentation" we discussed at the outset of this module. Quite unlike the classic quantitatively based "random sample," you frankly are not looking for a 'broad mix of infinite numbers of variables.' Rather, your study may necessitate pre-targeting subjects with certain characteristics. These might be superintendents with 10 years of experience or more in urban school districts, for instance. In the majority of cases, the reason for your doing focus group interviews in the first place is that you are seeking a deeper understanding of these target variables or subject characteristics, as they relate to something else, say, a policy, procedure, method of recruitment or retention, etc. Therefore, it is quite all right and even desirable to limit your sampling scheme to subjects possessing those target characteristics. Back to our population and sampling terminology, then, this would imply judgment/purposive; criterion; and perhaps critical case sampling. The selection criteria, such as years of experience, gender, ethnicity, age group, etc., imply strata. The sampling plan might also involve 'subjects recruiting subjects' in cases of small, rare, specialized sub-populations - i.e., snowball or chain sampling. If you have large enough 'pools' of these subjects with the target characteristics - say, a large number of administrators with 10 years' experience in urban districts - it then builds in an extra measure of generalizability if you can then choose the actual focus group participants at random. But - there is absolutely nothing wrong with using a nonrandom, nonprobabilistic sampling procedure for selecting focus group interview subjects if this is not feasible! Furthermore, you want subjects who are reasonably articulate, likely to be comfortable sharing with you and one another in the group setting - an additional selection criterion, if you will. All of these factors need to be carefully considered in your sampling plan.

*

Homogeneous samples. This one, which we also learned about in our population and sampling Module #5, EDR 610 Intro to Research, is critical for focus groups to be successful! Lots of studies have been done that demonstrate 'too much diversity inhibits free exchange.' Again, while you hope to see some individual differences of opinion, keep in mind that too much of a 'subject mix' might inhibit comfort and spontaneity. Under some circumstances, even combining men and women in the same interview session has led to biased and distorted results. Similar findings have resulted in combining age groups, ethnic groups, superior-subordinate teams in work settings, and so forth. It is virtually impossible to state a general rule here; it depends almost entirely on the nature of the topic of your study. You need to be aware of it and see if you might have a built-in conflict or inhibition due to such heterogeneity of the sample. A related side issue is your own role as the interviewer vis-a-vis the subjects. I once participated in transcribing a focus group interview session that was conducted by a young Anglo man from NAU with a group of Navajo and Hopi teenagers on the reservation. I about wept when I listened to the tape. The interviewer "made all the right moves," said all the 'suggested' things, to try to draw out the subjects - but they simply totally clammed up on him. He finally, reluctantly, chose to terminate the interview. My heart went out to him: he 'failed' not because of what he did or did not do - he was simply 'too different' to succeed in conveying that comfort zone. In cases of great differences on any such background demographic characteristic, the researcher needs to consider either a) letting someone of a similar background to the subjects conduct the interview; or b) at the very least, having someone of a similar background serve as assistant moderator.

 

Go back to Focus Group Interviewing Part I

E-mail M. Dereshiwsky at statcatmd@aol.com
Call M. Dereshiwsky at (520) 523-1892

Copyright © 1999 Northern Arizona University
ALL RIGHTS RESERVED