Using Qualitative Focus Groups in Generating Hypotheses For Subsequent Quantitative Validation and Strategy Development

Larry Percy, CREAMER INC
ABSTRACT - This paper follows Calder's notion of construct operation in focus group application reporting the results of an explanatory approach to qualitative research where first degree constructs are generated in three areas, with subsequent quantitative measures that produce their corresponding second-degree constructs. These findings are reported in sufficient detail to track the hypothesis generation-validation procedure, along with the resulting strategic decisions following from the results.
[ to cite ]:
Larry Percy (1982) ,"Using Qualitative Focus Groups in Generating Hypotheses For Subsequent Quantitative Validation and Strategy Development", in NA - Advances in Consumer Research Volume 09, eds. Andrew Mitchell, Ann Abor, MI : Association for Consumer Research, Pages: 57-61.

Advances in Consumer Research Volume 9, 1982      Pages 57-61


Larry Percy, CREAMER INC

[The author is in obvious debt to John R. Rossiter, who conducted the actual exploratory focus groups discussed in this paper and who was in large part responsible for the hypotheses generated.]


This paper follows Calder's notion of construct operation in focus group application reporting the results of an explanatory approach to qualitative research where first degree constructs are generated in three areas, with subsequent quantitative measures that produce their corresponding second-degree constructs. These findings are reported in sufficient detail to track the hypothesis generation-validation procedure, along with the resulting strategic decisions following from the results.


Too often the use of qualitative research in general, and focus groups in particular, follows no systematic or purposeful plan. It is seen as a (relatively) inexpensive and easily interpreted means of gathering some information about a question; one where answers are readily and quickly forthcoming. While this rather serendipitous approach may indeed be occasionally useful, or more to the point thought to be useful, it does not begin to realize the potential power of focus grouPs in an integrated research program.

Borrowing from Calder (1977), qualitative research should be utilized to generate constructs suitable for measurement. He distinguishes three approaches to qualitative research. exploratory, clinical, and "phenomenological". The latter are groups conducted to familiarize oneself with the everyday language and experience of consumers. This type of approach does not lend itself to subsequent quantitative inquiry, yet more often than not such qualitative groups are used as a basis for further quantitative measure,

If one is using qualitative research as a springboard to further inquiry, an "exploratory" mode is more appropriate. Such an approach treats reality as it is understood by the consumer (something Calder has called "first degree constructs), and juxtaposes it against the everyday experiences of the consumer. Hypotheses emerging from this type of qualitative focus group research are then ready to be validated through explicit numerical measurement to test the hypotheses generated, not merely enumerate results.


What one is actually seeking in conducting exploratory focus groups is insight from the everyday knowledge and experience of consumers. While it is not expected that people are .necessarily introspective about their behavior, by listening to them discuss their attitudes and behavior in the area of interest, one hopes to infer something of the decision process involved in choice behavior. This is by no means a simple process, and one is particularly cautioned against the assumption that the hypotheses generated are in fact "findings." This is not to imply that what you can learn from the exploratory focus group is therefore useless. It should shape one's thinking and suggest appropriate methodologies for further quantitative development. The real mistake is to simply assume answers have been forthcoming from the groups, subject only to the limitation of the small samples involved.

As Calder (1977) has pointed out, perhaps the single question most often raised in utilizing a focus group methodology is the generalizability of the findings. Generally, this is discussed in the usual way one talks about the generalizability of survey data: how "reliable" is the information; can we "project" it to the population at large? But this is clearly not the way to look at qualitative research, especially in the exploratory mode. In fact, it is not even particularly meaningful.

Our goal in conducting focus groups in an exploratory mode is to outline a hypothesis or model that must be confirmed by subsequent quantitative research, and at that point the question of sample generalizability becomes germane. From this perspective, it is simply not appropriate to think about the generalizability of exploratory focus group results. (While somewhat more applicable to a phenomenological or clinical approach to focus group research, that is not our focus here, nor, in the main, should this be the focus of most consumer behavior research.)

It follows, then, that the hypotheses generated by exploratory focus groups require quantitative development; not because of problems with sample generalization; but for the more important requirement of scientific discipline. To illustrate, extensive exploratory focus group work was conducted some years ago in an effort to develop some understanding of how consumers made choice decisions in purchasing yogurt. The resulting "model" inferred from the research is shown in Figure 1. It postulates a two-step model; one for first time purchases, and one for subsequent purchases. The key attributes effecting choice were flavor, brand, and price. Initially, brand was not considered, only flavor (keyed by a visual cue), and price. Assuming the experience was favorable, brand then became highly salient, acting as an anchor for subsequent purchase (providing stability in choice). The model implies a definite sequence in which potential buyers first seek out the familiar (i.e. "safe") brand, verify the desired flavor, then check price.



However, for this model to be useful, not only must we determine its efficacy through quantitative measures, but for purposes of advertising and marketing strategy, it is necessary to determine which of the stages of the model contributes most to ultimate choice. To evaluate if the brand - flavor - price hierarchy implied by the model was correct, a conjoint measurement analysis was used. The part-worth values revealed price was actually the most important consideration, followed by brand, with flavor least important.

An interesting measurement consideration here underscores the importance of selecting the most appropriate quantitative technique for hypotheses testing. To have simply asked consumer which of the three factors was the most, and which the least, important consideration in choosing the product would have ignored the potential of interaction effects, and provided misleading results. Response to that question indicated that flavor was the most important consideration, and brand least (Percy, 1979).


A particularly effective example of how a well-conceived application of exploratory focus groups to generate hypotheses for subsequent validation through quantitative analysis arose from a concern a number of years ago among potato growers and marketers over a long and continuing decline in per capita potato consumption. In addition, there was a feeling that the image of potatoes was suffering from a growing negative attitude toward its food value; a parallel problem to the public policy concern over America's food habits, especially nutrition.

Specifically, exploratory focus group sessions were conducted among homemakers in three geographically diverse cities. The groups were structured to treat reality as it was understood by the consumer, following what Calder (1977) has described as "first degree constructs," and juxtaposing it with the everyday experiences of the consumer hypotheses generated by these groups were then subjected for validation through a "scientific" approach (again, following Calder); i.e. using the quantitative measures to test the hypotheses, not merely enumerate results. These hypotheses were in fact pre-tested among 120 homemakers in three additional cities to optimize the measurement instrument, and validated among a nationally representative sample of over 1,200 homemakers.

Construct Generalization and Measurement

Results of the exploratory focus groups generated a number of first-degree constructs from which hypotheses evolved; three are discussed below in some detail in order to illustrate the procedure. Family Influence on Consumption; Attribute Importance in Selection; Images and Beliefs Associated with Consumption.

Family Influences on Consumption. Family influences on serving potatoes were found to center around a general feeling that husbands, and to a certain extent children (especially with french fries), were much more likely than the homemaker to enjoy potatoes and feel they are important to a meal. Women in the exploratory groups consistently voiced their husbands' stated belief that (referring to dinner): It's not a meal without potatoes. This was in fact found to be the case when measured in the phases. Nearly two-thirds of all homemakers felt their husbands "liked" potatoes, and over one-half said children like potatoes. In terms of the importance of serving potatoes at the main meal, women perceive their husbands (or men in general if not married) to be significantly more likely to find serving potatoes with the meal very important; and in addition, nearly all agreed strongly, with the statement: Most men feel a meal is not a meal without potatoes. So despite the fact that serving a potato at the main meal of the day is not considered very important to most women, they do feel it is important to their family.

This potential dissonance relationship between personal vs. family preferences in shaping potential use also extends to the ways in which potatoes are prepared. Clear favorites for different family members were measured in the quantitative phase: the husband's parAlleling the more all purpose and frequently served potato dishes; the children's quite different from the husband's; and the homemaker's paralleling those she perceives as best tasting. One might conclude from these results that the husband's taste, rather than the homemaker's, tends to be dominate use. Also apparent were a number of intra-potato dish differences and similarities.

Attribute Importance in Selection. As a result of the exploratory quantitative research; it was hypothesized that the potato selection process could be modeled as follows: first, a shopper searches for the general type of potato that will best meet their end-use meal planning needs. Here they would be looking for attributes of potatoes, or of various types of potatoes, in a positive sense. Then, once the general type decision is made, a Particular potato within that type category is selected from the loose or bagged potatoes of that type which are available at her supermarket. This is achieved by a process of negative attribute avoidance.

Essentially, it is the particular set of attributes associated with a particular type of potatoes which determines selection, Always with end-use in mind. The functional typology suggested appears to be as follows: at for baking thicker skinned potatoes, principally Idahoes; b) for boiling, thin skinned or skinless "regular" (non-Idaho) potatoes, red potatoes, or whole "baby" potatoes; c) for mashing, almost any potato, but more often the regular; d) for french fries the longer potatoes (for cutting) such as Idahoes or large regular whites; e) for home-fries, regular whites or canned baby potatoes; and f) for potato salad any regular white or red. The hypothesized relationship in almost all cases was specifically upheld, with only minor variations, in the quantitative phase.

Additional measures were gathered in the quantitative phases in an effort to establish the cognitive links between particular types of potatoes and main meal end-use. Each of the original functional types, plus two additional "special occasion" types (creamed potatoes and escalloped potatoes), were ranked in terms of their appropriateness for seven selected main dishes.



The resulting data were mapped into measurement via a multi-dimensional unfolding analysis. Looking at iso-preference contours for each main dish, the closer any particular potato dish lies to the main dish, the more likely that particular potato dish is considered appropriate to serve with that main dish. In the solution space shown in Figure 2, for example, one can see that roasts and pork are seen as being very much alike in terms of potato dishes most likely to be served with them. Another interesting finding is the remarkable similarity (in terms of main dish serving likelihood) of boiled, creamed, and escalloped potatoes. They seem to derive their commonality from the fact that unlike the other potato dishes, they are neither very likely nor unlikely to be served with any of the main dishes studied.

Images and Beliefs Associated with Consumption. While a number of associative attributes are generated in the exploratory phase of the research, the resulting image of potatoes was hypothesized to center around health and nutrition beliefs, with a basic positive side and a basic negative side. Questions of "value," although in some ways related to each of the previously discussed first degree constructs, also tended to he wrapped up in health and nutrition based beliefs.

On the positive side, potatoes were hypothesized as a food staple, and a major source of vital energy-producing carbohydrates. It was found, however, that it was not the vita mins in potatoes that were valued, nor even the vague term "minerals" (although the latter was a more common association). When people refer to potatoes as "nutritious" they do not mean the word in its colloquial sense of "vitamin rich" but rather in its literal sense of "nuturant" or "sustenant." This, of course, is an important distinction Additionally, potatoes (or more precisely fresh potatoes) are hypothesized as being regarded as a natural food - a product of nature with minimal interference from outside influences.

Measurement of these beliefs confirmed that indeed most homemakers would apply the vague attribute of "nutritious" to potatoes, but few would ascribe the more specific attributes of "rich in vitamins and minerals" or "Vitamin C." Coincidentally, the attribute "natural" factored with "nutritious" (along with such other general attributes as "fresh" and "food staple"), further underscoring the hypothesized literal sense of "nutritious" rather than its colloquial sense of vitamin and mineral rich.

The negative image of potatoes drawn from the qualitative research clearly centered on health fears associated with potato consumption: overweight and attendant medical complications, aesthetic loss of bodily trimness, and cosmetic complexion problems. A complicated set of beliefs appeared to be involved.

When pressed, people will logically admit that it's not the potato that causes problems but (a) what you put on it and (b) how you prepare it. The idea of "what you put on it" may seem to exonerate potatoes, but it doesn't. People very very rarely eat potatoes plain. Baked potatoes are eaten with butter, margarine or sour cream; mashed, with gravy or cheese topping; boiled or roasted, with gravy; french fries, ketchup; potato salad, mayonnaise or dressing. Public knowledge of vitamins and balanced diets may be vague or ill-formed, but people readily identify foods high in "fats" (saturated fats which produce cholesterol). Note how many of the typical potato add-ons are in the high fat group: butter, cream, cheese, egg yolks in mayonnaise, oils in salad dressings, meat fat and sugar in gravy.

The second idea which potentially shifts blame away from potatoes per se lies in the way they are prepared. Quite aside from whatever is added later, it was hypothesized that people perceive a kind of "health hierarchy" based on mode of preparation. It begins with the likely healthiest form, potato salad, and descends through the hot varieties.

The hierarchy introduces another common belief about potatoes, namely that most of the vitamins and minerals are contained in the skin. Removal of skin thus reduces health and nutrition (colloquial sense this tame) connotations.

Measurement in the quantitative research provided ample support for these hypotheses. A significant number of women believe you can't eat a potato without something on it, and almost all strongly agree that seasoning is very important in preparing fresh potatoes; and nearly everyone agreed that the skin is the most nutritious part of a potato.

Finally, the third important negative aspect of potatoes is a popular belief in their high caloric content. Again, the psychological basis of this belief appears to be fairly complex. First of all, potatoes just seem to be heavy by their very nature. Factual caloric evidence to the contrary, other starches such as rice or bread- even with topping or butter added, simply taste and look lighter than potatoes.

Supporting the high calorie perception is the mental picture of heavy potato eaters as being heavy people. Subjects were asked to descriptively profile their image of a frequent eater of potatoes, then to profile a non-eater of potatoes. The strongest frequent eater image was that of a large, heavy or fat man; if not fat, a large "plain eating" manual worker. Non eaters, in contrast, suggested a slim woman, a weight watcher or dieter.

Quantitative measurement reveals that fewer than one-half would say a potato is "healthy," while over one-third felt they had too many calories. Asked to describe someone who eats a lot of potatoes, overwhelmingly a person's first image is one of someone who is fat; this contrasts with the image of someone who never eats potatoes as thin or dieting. It is no wonder that inhibitions are focused on the potato itself, regardless of how it is prepared or served.

Strategy Derived from Quantification

Because of the potential instrumentality between behavior and the negative salience of many beliefs, initial strategy development is drawn from those constructs; family influences or consumption and attribute importance in selection are utilized only as secondary or mediating constructs. Assuming the dominance of the image dimension, our first task in developing a successful advertising strategy from the validated hypotheses is to construct a set of values for the target receivers, distinguishing positive and negative. Major values are derived from the foregoing discussion, and arrayed in a value hierarchy from strongest to weakest (see Table 1).



When a communicator is presenting a communication in which the conclusion is one with which the audience already agrees (e.g. the positive value links described above) he need only present a favorable one-sided argument. However, when the communication involves a conclusion with which the audience is predisposed to disagree (e.g. the negative potato value linkages) he has a choice of either ignoring the audiences natural counter-arguments or refuting them with a two-sided argument of his ownCi.e. one which attempts to address and refute the audience's side and persuade them over to the communicator's side. The ignoring strategy works best when the counter-arguments are either weak or nonsalient. When the counter-arguments are both strong and salient, too much so to be ignored, a refutation strategy should be adopted.

Some of the earliest work on the efficacy of refutational appeals came out of research during the second World War by Hovland and his associates (cf. Hovland, Lumsdaine, and Sheffield, 1949), which showed that neither ignoring opposition arguments nor explicitly refuting them has greater efficacy under all conditions. In general, ignoring the opposition was found to be more desirable if the receivers were initially favorable towards one's conclusion, but the mentioning and then refuting the opposition is somewhat more desirable if the reviewers were initially opposed to one's position (as is the case in this problem). Work by others (mentioned by McGuire, 1969) has likewise indicated that refuting or ignoring opposition arguments are about equally effective overall in producing direct attitude change, although earlier studies by McGuire (1963) found that explicitly refuting opposition arguments, rather than ignoring them and presenting only one's own arguments, does produce more direct attitude change.

Since any communication effort would involve a message conclusion which the target segments were predisposed to disagree with, it was quite clear that a major reversal of diminishing homemaker favorability toward potatoes could only be accomplished by successfully refuting the increasingly strong negative health aspects surrounding the image of potatoes. It is a well established research finding of social psychology that when both positive and negative attitudes are held toward a common entity, negative characteristics subjectively outweigh the positive in determining outcomes (Wyer, 1974).

One is forced to adopt a refutation strategy in order to break down the existing beliefs and replace them with new beliefs which could be subsequently associated in a positive fashion with potato usage. e general refutation format suggested was:

(1) Forewarn the receiver of the intended conclusion. This places the source bias out in the open at the very beginning, maximizing potential reception of the point of the message. It is, of course, no guarantee of persuasion.

(2) Present receiver-beliefs first. This has the effect of preventing an early tune out to the message; and importantly, tends to reduce the likelihood of the receiver raising those same beliefs in counter-arguments after the advertiser's side of the message.

(3) Present message points which refute receiver's negative beliefs and support advertiser. This is the start of the advertiser's message.

(4) Draw the conclusion explicitly. This serves three functions: (a) it overcomes any receiver reluctance to draw it himself; (b) it provides a sense of closure and completeness; and (c) it underscores the confidence of the source in his position.

One successful print execution of this strategy utilized a headline reading "Lies, Lies, Lies" over a picture of a large baked potato surrounded by negative captions that mirror many of the target segments' negative beliefs. Body copy refuted those beliefs, and closure was provided by the tagline: "The Potato. Something good that's good for you" (see Table 2).



Results of the advertising campaign developed from this communications oriented segmentation of the potato market based upon homemaker attitude has been significant over the last several years. As reported in the Marketing News (1976) the general level of attitude had changed from a "not-very-nutritious, fattening substance, to a nutritious vegetable that isn't too fattening." They quote data from a research study that showed the number of people who believe potatoes have too many calories had dropped from about 1/3 to 1/4 over a two year period, while the percent who believe they are basically nutritious grew from 55% to 81%. And most importantly, consumption of fresh potatoes by the average household rose 17.7%; the first major rise in usage since 1950.


In summary, when using focus groups in the study of consumer behavior, it is important to bear in mind that these groups are used only for generating hypotheses, and do not represent findings. These hypotheses must be validated through quantitative research. It is essential to the development of effective marketing or communication strategy that this two step process be followed. The first is necessary to ensure the domain under study is adequately addressed from the reality of the consumer; the second that it is valid.


Calder, Bobby J. (August, 1977), "Focus Groups and The Nature of Qualitative Research," Journal of Marketing Research, 14, 353-364.

Hovland, C.L., Lumsdaine, A.AL, and Sheffield, F.D., (1949), Experiments and Mass Communications, Princeton: Princeton University Press.

McGuire, William J. (1963), Immunization Against Persuasion, New York: Department of Social Psychology, Columbia University.

McGuire, William J. (1969), "the Nature of Attitudes and Attitude Change," in G. Lindzey and E. Aronson (Eds.), The Handbook of Social Psychology, Preading, Mass.: Addison Wesley Publishing.

Marketing News, July 6, 1976.

Percy, Larry (1979), "The Importance of Understanding Consumer Decision Processes When Applying Analytic Product Models," in Allan D. Shocker (Ed.), Analytic Approaches to Product and Marketing Planning, Cambridge, Mass: Marketing Science Institute, 360-367.

Wyer, R.S. (1974), Cognitive Organization and Change: An Information Processing Approach, Potomac, Maryland: Lawrence Erlbaum Associates.