Design Options
Single-Category Design
The traditional design for a focus group study is to conduct focus groups until you have reached the point of theoretical saturation-the point when you are not gaining new insights. The number of groups needed to reach saturation can vary, but usually the researcher will initially plan for three or four focus groups with a target audience and then decide if adequate saturation has been reached or if additional groups should be conducted.
Although theoretical saturation is a great concept and useful in academic work, as a consultant you won't land many contracts if you say you plan to conduct groups until you reach theoretical saturation.
EXAMPLE
Example of Single-Category Design
Let's say you wanted to use focus groups to evaluate a leadership development program for youth. You decide the information-rich people are youth who have completed the program in the past two years. You don't want to compare or contrast based on any other features. So you use a single-category design. You conduct three or four groups with your target audience and determine If saturation is achieved. If saturation is achieved, then the study is completed and reports are prepared. However, if new information is being obtained after the third or fourth group, the researcher will need to decide if additional groups should be held or if the report should indicate that saturation was not achieved. Perhaps it is enough to know that people hold widely different views on the topic.
Number of Groups (0 = 1 group)
Audience (youth who completed the program in the past 2 years)Saturation? o o o o
Figure 2.1. Single-Category Design
Clients want to know how many groups are needed, how long the study will take, and how much it will cost. Few clients will give you a blank check to conduct as many groups as needed until saturation is achieved. Therefore, we typically plan at least three or four groups for the most important groups we want to listen to.
Multiple-Category Design
A variation of the traditional design is to conduct groups with several audiences, either sequentially or simultaneously. This design allows the researcher to make comparisons in two ways-from one group to another within a category (youth) and from one category to another category (e.g., comparing what youth said to what parents said).
EXAMPLE
Example of a Multiple-Category Design
Let's continue the above example but make it more complex. Let's say it is Important to hear the reactions of youth who have completed the program, as well as parents of youth who have completed the program, mentors involved in the program, and staff members. It is important to be able to compare and contrast the reactions of these different groups. Let's suppose that you have enough resources to conduct only ten groups.
You believe feedback from youth and parents will be most useful for your study, so you place more emphasis on getting information from them. You conduct four groups with youth because their feedback is most important. But you divide the youth groups by age, knowing that younger youth are often intimidated by Older youth or simply defer to them. You cons duct three groups with parents. It is important"* but not critical to get feedback from bothji mentors and staff. You decide to conduct two groups with mentors. There are so few staff members that you are able to get their inpufjs in one focus group.
Number of Groups (0 - 1 group)
Audience 1 (youth, ages 14-15) (youth, ages 16-18) Audience 2 (parents) Audience 3 (mentors) Audience 4 (staff) o oo oo o oo oo
Figure 2.2. Multiple-Category Design
Double-Layer Design
Another version of the traditional design involves multiple layers, which might involve geographic areas as the first layer and different audiences as the second layer. In this design, the researchers can make comparisons between any of the layers in the design.
EXAMPLE
Example of a Double-Layer Design
Suppose that a nationwide health service wants to do a study to better understand what patients suffering from severe depression consider to be good health care. The results of the study are to be used, along with numerous other sources of data, to develop clinical guidelines for care of patients who suffer from depression. The study team decides it is most important to talk with patients who are being
treated for depression but are not currently hospitalized. They also want to listen to family members who are close to these patients. This might include spouses, parents, or adult children of patients. The health service has four geographic regions, and studies traditionally have included all four regions. One facility is selected to participate in each region.
Layer 1 Layer 2 Number of Groups (0 = 1 group)
East
Audience 1 (patients) o o o
Audience 2 (family members) o
West
Audience 1 (patients) o o o
Audience 2 (family members) o
South
Audience 1 (patients) o o o
Audience 2 (family members) o
North
Audience 1 (patients) o o o
Audience 2 (family members) o
Figure 2.3. Double-Layer Design
Broad-Involvement Design
Occasionally there are studies with widespread public interest. In these studies, some people may be concerned if they are left out of the study. A potential sampling strategy is to identify a primary audience who is considered to be your greatest information-rich source. We'll call this source your "target audience," whereas other sources represent
EXAMPLE
Example of a Broad-Involvement Design
Let's say a state department of education is proposing policy changes in special education. They want to know what kind of implications the proposed changes may have for what takes place in the classroom. The proposed changes directly affect special education teachers; therefore, the study team believes special education teachers are the people who can best provide the needed information-how the changes will affect the classroom. Yet, the team knows that other people can also provide feedback: regular education teachers, parents, students, advocates, and school administrators. In some studies, particularly in the public sector, there are groups that feel they must be listened to before decisions are made or policy is established. These groups may feel that the study would be incomplete unless their views are heard and recorded. The study team may agree that listening to these groups would enhance findings, but they have limited time and resources. In these situations, the strategy begins with your primary audience (special education teachers), and the team listens broadly to establish a clear baseline. Perhaps the study team conducts one group with special education teachers in each of seven education districts in the state. Then one or two focus groups are conducted with each of the additional audiences (regular education teachers, parents, students, advocates, and administrators). Results of these groups are compared back to the baseline established by special education teachers. Occasionally, unique and crucial information may emerge from the focus groups with these secondary audiences. Then additional focus groups can be conducted with that audience.
Audience I (special education teachers) Audience 2 (regular education teachers) Audience 3 (parents) Audience 4 (students) Audience 5 (advocates) Audience 6 (administrators) Number of Groups (0 = 1 group)o o o o o o oo oo oo
Figure 2.4. Broad-Involvement Design
secondary but important perspectives. For example, in a program to prevent teen violence, researchers decided that their target audience was male teenagers, ages thirteen to seventeen, living in high-crime neighborhoods. Other audiences (police, teachers, parents) may have had valuable perspectives on violence prevention, but resources prevented an in-depth study of all audience categories. However, if the study listened only to the target audience, other audiences may not have cooperated in developing solution strategies. In some environments, there is the attitude that if my "group" isn't involved, then we're not considered important when it comes time to work toward a solution. Or, "I can't believe it because you never listened to people like me!" In this design, you are concerned not only about what is good research but also about what it will take to get people to buy into your project and support your efforts. You start by making sure you have good research, then you add groups to make sure the study is practical in a political sense.
The design strategy consists of first anchoring the study with the groups of the target audience. In fact, this audience is often purposefully oversampled and exceeds the point of saturation. This is often done to include geographic representation (e.g., a state agency wants one group done in each of its seven regions). After the patterns are detected within the target audience, later analysis compares each additional audience type back to the target audience. If a particular audience provides key information that is not included in the analysis of the target audience, then the researcher might add a second or third group to that particular secondary audience to determine if that pattern appears again in similar groups.