Emergent Focus Group Design

When we think about doing a focus group study we usually design the study, collect the data, and do the analysis—in that order. We recruit all of the participants with the same pre-designed criteria, ask them the same set of questions, and produce one big report at the end. This approach works well in most cases, especially when studying well-defined issues that occur in familiar study populations.

But when studying ill-defined or complex issues, an unfamiliar study population, or both, this predetermined design-collect-analyze sequence has potential drawbacks: we miss out on opportunities along the way to reflect on and learn from data revealed between groups. Consequently, we miss the chance to adjust the design in ways that can maximize data quality and quantity.

Opportunities Afforded by Emergent Design

For ill-defined or complex studies, nationally recognized focus group expert, Dr. David Morgan, proposes an alternative called “emergent design.” Emergent designs allow us to stop and reflect between groups or sets of groups, creating key opportunities for making insightful design changes along the way.

The between-group breaks afford us opportunities to move beyond study questions that quickly reach saturation or turn out to lack imagination. Emergent design also allows us to further hone recruitment criteria for later groups as well as identify information-rich participants from initial groups. Ultimately, emergent design allows us to customize the study in ways that improve data quality and use through the continual process of realigning data collection techniques with stated project goals.

Recently David met with me to outline the fundamental concepts and applications of emergent design, which I’m happy to share with you in in this blog.

Shifting the Funnel from Within to Between

As David explained to me, many of us already (perhaps unknowingly) incorporate emergence into focus groups when we sequence questions from general to specific in question guides—the so-called “funnel approach.” Broad, generic questions form the wide end of the funnel while more specific, targeted questions form the narrow end. If the same set of questions is used for all groups, emergence is said to occur within groups.

However, if participants generate an unexpectedly large number of new topics or the group facilitator is unable to facilely work the topics into the guide, the discussion can be stifled and important questions left unanswered. This is a perfect situation in which to switch the funnel from within groups to between groups. Broader questions get asked in earlier groups and more specific ones saved for later groups. And, instead of the usual uninterrupted data collection process with a single analysis at the end, data collection and analysis occur simultaneously throughout the study.

Application of Emergent Design

The flexibility of inter-group emergent design has many applications. Specific project goals and data needs must always be primary considerations when designing focus group studies, but some situations lend themselves particularly well to emergent design. Here is a partial list:

• A poorly formed study topic
• An unfamiliar population group or service area
• A limited or tight budget
• A primary interest in participant generated topics
• An anticipated early saturation of some topics
• A desire to generate new hypotheses
• Recognized poor understanding of the issues
• Ill-defined population segments

Configurations and Characteristics of Emergent Groups

Emergent group designs can assume a number of configurations. The more you understand about your study population and research topic at the outset, the more your design will resemble a traditional one, i.e. several groups in one set. The more preliminary your study, the more you will want to separate the set into focus group subsets to provide opportunity for adjustment between each subset.

Regardless of the number or arrangement of groups into subsets, all emergent designs share basic methodological traits. Early groups are more exploratory, less-structured, and ask fewer and broader questions. Subsequent groups are more structured, focused on specific questions, and more likely aligned with study objectives than participant-generated topics. The chart below outlines how traits change as focus groups emerge.

	Early Groups	Later Groups
Discussion	Exploratory	Targeted
Participant Recruitment	Broad selection criteria	Selection criteria determined by insights from earlier groups
Question Guide*	Broad questions	Specific questions derived from responses to earlier questions
Breadth Depth	Many topics	Fewer topics
Facilitation	Less structured	More structured
Topics	Participant generated	Researcher generated
* Could also decide to use the same question guide for both early and subsequent groups.

Two Scenarios, Six Groups

Included here are two generic scenarios (of many possible options) that David outlined to demonstrate how one might operationalize emergent design with a set of six focus groups.

• Three and three

Suppose you decide to start with one subset of three groups and hold the other subset of three back until you see what you learn from the first three.

Group 1

Group 2

Group 3

Analyze Report Redesign

Group 4

Group 5

Group 6

You conduct the first three groups using the kinds of questions that would typically appear at the wide end of the funnel. Questions in the subsequent three groups are likely to resemble those at the narrow end of the funnel. Your participants have been recruited using generalized selection criteria that match the goals of your study.

Midway, you stop to analyze and informally report on what you learned in groups 1, 2, and 3. Five questions you will want to answer before proceeding with groups 4, 5, and 6 include:

1. Have we reached saturation on some topics? If so, do we really need to collect any more data on those topics?
2. Do we need to hear more about certain topics? Have new topics been raised or triggered by participants?
3. What topics or subtopics might we want to add that either were not included initially, or emerged from the original topics?
4. Do we have the right types of participants? Should we include more population segments or collapse some segments?
5. Overall: Do we need to change anything or should we proceed with the second set of groups just as we did with the first?

Based on the answers to these five questions you may want to:

1. Delete unproductive questions.
2. Add newly generated questions.
3. Make questions more specific or provocative.
4. Revise recruitment criteria for groups 4, 5, and 6.
5. Create additional population segments or collapse those with similar responses.
6. Continue with groups 4, 5, and 6 in the same way you conducted groups 1, 2, and 3 using the same question guide and recruitment criteria.
7. Stop the study if enough high quality information has been generated on all topics of interest.

• Two, two, and two

If you are less sure of your understanding about a certain topic, you may want to break the six groups into three subsets of two groups each, mimicking a more elongated funnel that moves from broader exploratory questions to more narrowly defined formats.

Group 1

Group 2

Analyze Report Redesign

Group 3

Group 4

Analyze Report Redesign

Group 5

Group 6

Let’s assume you will use a progression of three sets of two groups each. Maybe you have a situation in which you are trying to create a workable design out of a difficult set of choices and want to focus the questioning in a way that is most likely to produce a high value for your effort and maximize your limited budget.

In this scenario, Groups 1 and 2 would be exploratory. Participants would come from all possible segments and be asked to respond to four or five very broad questions.

Groups 3 and 4 would be semi-structured. Participants would be selected from segments that emerged as important in groups 1 and 2. Groups 3 and 4 may even include some of the same participants from groups 1 and 2.

At this point, non-useful or saturated topics would be been eliminated, questions redesigned to be more directive, and participant generated topics blended in with study topics.

Your target is to use groups 5 and 6 to tie up loose ends.  At this stage, questions are specifically focused on the fixed agenda that emerged from groups 3 and 4. Participants are recruited according to potentially re-revised criteria. And, the focus is narrow and  overall purpose to generate specific depth and detail.

Resources

If you would like to learn more, you can find a chapter on emergent focus group design authored by Dr. Morgan in The Handbook of Emergent Methods edited by Sharlene Nagy Hesse-Berber and Patricia Levy. David and I would also be happy to consult with you on your emergent focus group project. Or email us about how you’ve used emergent design to tell us how it went.

Please subscribe to my blog. It’s easy!