Editor’s note: There has been considerable controversy surrounding the collection of quantitative data while conducting focus groups. One individual who believes it is possible is Dr. James M. Leiman, who reports on his methodology below in a TrendFacts Research copyrighted article. James M. Leiman, Ph.D. is vice president of TrendFacts Research/The Creative Group, Inc., a full-service marketing research firm located in Farmington Hills, Michigan. He received his Ph.D. in Cognitive Psychology and Measurement from Wayne State University. Besides being an experienced focus group moderator, Dr. Leiman's areas of expertise include survey sample design, multivariate statistical analysis, and the measurement and modeling of consumer preference and choice. Prior to joining TrendFacts Research he worked for Market Opinion Research and Wayne State University in Detroit.
Because focus groups do not employ rigorous sampling plans and do not interview large numbers of respondents, they generally do not present an appropriate setting for collecting quantitative data. However this does not mean that it is never appropriate to collect quantitative data as part of a focus group. One can collect quantitative data in a focus group setting in support of a number of different quantitative research objectives. This is especially true when the focus group research is examining the attitudes and opinions of members of low incidence groups which are too difficult or costly to study using traditional quantitative research designs.
From one perspective the focus group setting is ideal for collecting quantitative data. Complex stimulus materials can be easily presented in a focus group setting. Also, data collection tasks which are difficult to do over the telephone can be easily done in a focus group. Perceptual mapping studies involving several attribute ratings on a number of objects or products are a good example. Multidimensional scaling projects requiring a large number of similarity or dissimilarity judgments are another.
The collection of quantitative data can also actually serve to facilitate group discussions. Many moderators use facilitation techniques involving paper and pencil tasks. Quantitative data collection tasks can be used in the same way. For example, by having participants make product attribute importance ratings they are forced to reflect on the dimensions of a particular product or service. A skillful moderator can stimulate discussion by having participants share their ratings. The use of focus groups for collecting quantitative data, therefore, should not be rejected out of hand. Rather, we need to examine what the requirements and limitations are for using focus groups to collect such data.
Two areas of concern must be addressed if one wants to use a focus group to collect quantitative data. The first is the lack of a random sampling plan used for the recruitment of participants. The second is the relatively small sample sizes involved in focus group research.
The fact that participants are not randomly sampled is a serious problem when working with many consumer groups. However, much focus group research is conducted with members of relatively small subgroups. For example, cardiologists may be recruited to discuss a new medication or architects may be recruited to discuss certain types of building materials. Such groups are nearly impossible to randomly sample. Any research conducted with them always involves non-random sampling. However, through the use of quota and controlled sampling techniques one can at least recruit participants with profiles similar to known group characteristics.
The problems presented by the lack of a random sampling procedure for such groups is offset by the fact that many projects focus on a limited set of issues of consequence only to members of the subgroup. Such groups tend to exhibit a greater homogeneity of response compared to consumer groups recruited to discuss more general types of issues. The greater homogeneity of response of professional groups has direct implications for the second area of concern: sample size. As the homogeneity of response increases, sampling variance decreases and the sample sizes required to establish acceptable levels of reliability get smaller.
In order to illustrate the potential focus groups offer for collecting quantitative data, let us discuss a recently completed project. Ten groups of professionals from the office furniture and design community were conducted in two cities. After the normal introductions the moderator had participants make importance ratings on a questionnaire for a set of attributes. Following the paper and pencil ratings, the moderator used the rating as a vehicle for getting participants into a general discussion of the product area. Next, some actual product was presented including prototypes as well as competitive product. Participants were required to rate the products using the same set of attributes as before.
Following the product ratings, a "standard" focus group discussion took place for approximately one hour. Then, the participants were asked to consider just one of the products on display (the prototype). Various product "packages" were presented consisting of different options and prices. The participants were asked to say whether they would consider purchasing the prototype given the package. The different packages happened to be factorially generated as part of a conjoint analysis design. The quantitative analyses yielded by the ten groups included tabulations of importance and product ratings, as well as a modeling of purchase consideration as a function of product attributes and price. Price elasticities were easily calculated using these models.
The outline above could be modified to accommodate a variety of quantitative objectives. The important point is that the introduction of the quantitative tasks did not detract from the groups and, in fact, added an additional dimension to the analysis. The quantitative objectives of the research will dictate the number of groups required when focus groups are used to collect quantitative data. At times, more groups will need to be conducted than one would normally conduct if they only had qualitative objectives. In such cases, abbreviated "mini-groups" can be conducted along with other groups. The mini-groups would typically take less time and would consist of all of the quantitative data collection tasks as the other groups, but the moderator would use an abbreviated discussion outline.
Focus group research will never serve as a substitute for well-designed survey research, not when the primary goal of the research is the collection of certain types of quantitative data. However, there are situations where survey research becomes too difficult or costly to justify. This is especially true when a low incidence group of highly homogeneous respondents needs to be studied. The scope of focus group research should be expanded to allow for the collection of quantitative data in such situations.