The other day, a prospective client asked me how many focus groups I would recommend for a particular project. It was an introductory conversation, so there was a lot I didn’t know yet about the project scope. In my experience, when you are working with colleges and schools, it can take several conversations before you nail down the number that is most appropriate.

I did tell her that when you commission a research project you rely on the your consultant to answer a question such as this. That’s why you’re hiring him or her. It is the consultant’s roll to design the research protocol based on his or her experience.

There are some general rules of thumb regarding focus groups that apply in most cases. Before I get to these, let me give the standard caveat: please be aware that there is no number of focus groups by which you will arrive at statistically valid findings. Focus groups are qualitative research: their purpose is not to achieve the same sort of reliability you get with quantitative research. In this light,  it doesn’t matter whether you conduct two, four, ten, or twenty focus groups. The research will continue to be directional, not projectable over a population with a statistically valid level of certainty.

Focus groups do have their own sort of reliability. If they are conducted in a professional manner at sufficient quantity and with appropriate safeguards, the results are can be extraordinarily helpful in developing marketing strategy. Quantitative research and qualitative research play separate roles, and each yields a different kind of information.

Now to get to the question at hand: the minimum number of focus groups you ever want to be conducting on any topic is two. As anyone who has observed them can tell you,  focus groups tend to converge on a particular point of view. If you conduct two focus groups on the same topic you can often walk away with very different impressions from the two groups. You conduct a minimum of two focus groups to check against this kind of  skew and develop a broader picture of audience attitudes.

If the basic number is two, you can arrive at the quantity needed for a project by multiplying two by the number of segments you want to compare and analyze. For example, if you want to study female responses versus male responses on a particular question, you should conduct a minimum of four groups: two with men and two with women. If you want to study two separate variables in a single project, for example attitudes of in-state versus out-of-state prospects and men versus women, you would conduct a total of eight focus groups, two devoted to each of the four cells of information created by the two variables. You can see how the number can quickly multiply: An advancement research project that studied non-donor versus donor attitudes across three different generational segments would require 12 focus groups.

The number two is the minimum for any particular segment. But this assumes that you are conducting focus groups with full professional standards – that you are conducting the groups in a dedicated focus group facility with a professional moderator, professional recruiting, and other safeguards against bias. We all understand that in academic settings, it is sometimes not possible to achieve this level of professionalism. Budgets dictate that you cut corners.

It can still be worthwhile to conduct focus groups, even if you cannot afford to do them in the correct professional manner.  To safeguard against the increased unreliability that comes with less professional groups you should increase the quantity. Rather than conducting two focus groups per segment on any given topic, you might conduct three or four per segment.  I worked in a volunteer capacity on a research project for my synagogue in which the entire moderating staff, with the exception of me, were untrained and we recruited the subjects ourselves.  You can be sure we worked very hard; we conducted fifteen focus groups on a single segment in the course of the research. The results were adequate for the question we were researching.

So, the basic rule is this: multiply each segment by a minimum of two. If you are compromising the professionalism of the groups, increase that multiplier. That’s  the answer in a nutshell. Any other questions?