Research, evaluation, analysis

Focus Groups: A User's Guide

Focus groups are popular as a research method these days. They are simply discussion groups composed of members of the population you want to study. They are usually an exploratory research method that is, they are useful in generating hypotheses to test in further quantitative research.

Focus groups can be used as a quantitative research method rather than as a qualitative one. Instead of using focus groups to generate hypotheses, the researcher treats them as a survey and uses them to test statistical hypotheses which have already been formulated. In general, using focus groups for this purpose is very difficult. Interpreting the results requires some analytical sophistication, especially as you usually do not know beforehand what your dependent variables are going to be. And because sample sizes are usually small, any statistically significant result will be of the trout-in-the-milk variety. If you do find a trout in the milk, that's an important discovery, but you certainly can't count on finding one.

There are several other drawbacks to using focus groups in this way:

  • The chances of getting a representative sample in focus groups are low.
  • Focus groups take a relatively long time to conduct, so sample sizes are necessarily smaller than with other methods.
  • It's difficult enough to get people to answer a questionnaire that's been delivered to their homes; getting them to turn out in the evening and sit around for a couple of hours talking to strangers is even more difficult (and usually involves paying them money). Participants in focus groups therefore probably represent a much smaller segment of the population of interest than do participants in other types of survey.
  • Payment creates another problem. Payment is an additional incentive for participants to say what they think the person paying them wants to hear. This is a variety of experimenter effect participants in research often try to be helpful and do what they think the experimenter wants to see them do. After all, most of us want to be helpful and useful. Payment is an additional reason for participants to succumb to this temptation. This problem is attenuated in mail surveys because the experimenter is absent. In a focus group there is always a moderator to serve as a stimulus for good intentions.
  • Then there are conformity effects. The pressure in groups to conform to the norm the group establishes has been well documented. Some important opinions may therefore not be expressed.
Overcoming these problems in a quantitative focus group study requires a highly skilled moderator and a sophisticated analyst. For the typical consumer of research, the better approach usually is to analyze the discussion thematically and then test the analysis with a standard quantitative method and a new sample of consumers. This approach can produce very powerful results.

The urge to use focus groups as a quantitative method sometimes leads to their being conducted as if they were questionnaire surveys. The moderator simply asks each participant the same questions in rotation. Apart from encouraging experimenter and conformity effects, this approach is a waste of everyone's time and money. If you know what questions you want to ask, give people printed questionnaires they can fill out in a few minutes rather than have them take a couple of hours out of their lives to listen to other people answer questions. You can draw a larger sample that way, too, and attenuate experimenter effect.

For most consumers of research the best approach is probably to use focus groups when you're not sure what questions you want to ask the sample you intend to draw. Let the participants in the group decide what to talk about, and let them discuss it without any intervention from you except when necessary to maintain order or to make sure that you understand the points people are making. Analyze the discussion thematically rather than quantitatively. If you're paying attention, you'll learn something useful and interesting.

Focus Groups: A User's Guide © 2000, 2005 John FitzGerald
Home page | Decisionmakers' index | E-mail