1.3 – Analyzing and Interpreting the Wisdom of the Group

A Look at the Last Three Steps of the Process

We explored the first few steps of the GCM process in the previous module.

Step 1 – General Planning: The researcher identifies participants and develops the focus prompt.
Step 2 – Idea generation: Participants respond to focus prompt by contributing ideas.
Step 3 – Data Development and Phase II Planning: The researcher collects and synthesizes data; invites participants to organize it.
Step 4 – Organizing: Participants sort and rate statements.

By the end of step 4, the researcher has a data set to analyze in step 5.

Let’s look now at steps 5, 6, and 7.

Step 5 of the GCM Process

In step 5, the researcher uses statistics to analyze the sets of sorted responses to the focus prompt. She produces a similarity matrix, a table of values that indicate how often each statement was sorted with each of the other statements. Then she uses those values and a process called multidimensional scaling to create a “map” of the statements.

This first concept map is the point map. In this example, we use only 10 ideas, but in most projects there are usually between 80 and 100 unique ideas.

Point Map

Each point on the point map represents one specific idea from a participant in response to the focus prompt. Responses that are closely related, according to how participants saw their similarity, are close together on the map; responses that are different are farther away from each other.

On the point map, the distance between points is what is important.

Once the point map has been created, it is analyzed further using a process called hierarchical cluster analysis, to create the cluster map.

Cluster Map

The researcher creates the cluster map using hierarchical cluster analysis. This map shows the clusters overlaid on the point map.

The cluster map is a way of sharing a structure that all participants helped build, at a higher level. Concepts or themes emerge from the specific idea responses.

How many clusters?

Given a particular point map—in this case, one that contains 66 ideas, how many clusters is the right number of clusters to choose for the solution?

For example, in the image above, the points in the point map have been grouped into six clusters. Is six the only possible number of clusters? View the animation below to see how the number of clusters can vary.

As you can see, it’s possible to create a wide range of cluster maps, some with many clusters and some with fewer. Which cluster map is “correct”? This is a question that the researcher must consider carefully. She must determine which cluster solution best expresses the way the group thinks about the issue at hand. Her choice of the best solution will depend on the purpose of the project.

After the cluster map is created, the researcher can overlay the ratings data.

Knowledge Check

Step 6 of the GCM Process

In step 6, the researcher shares the maps and other outputs with the participants. Participants review the results, give feedback, and use their understanding of the context and issue at hand to interpret the maps and outputs. This is a critical part of the process, bringing fuller meaning to the results, and deepening the understanding of the collective mental model.

Step 7 of the GCM Process

The process usually concludes with collaboration between the participants or a representative group and the researcher in step 7.

When the researcher initiated this project, she did so with a particular purpose – to use GCM to help her make a decision, develop a theory, or create a plan. During step 7, the researcher and participants work together to craft a plan for utilization, develop an assessment plan, and then to use the results for the purposes of the project.

An Example – The Future Foundation

Let’s see how analysis is going for the team in our example, the Future Foundation. The researcher has created a point map to show the relationships between the responses in the data set.

Ideas that were more frequently sorted together are closer together on the point map, like these three:

• help employees understand when to access care and how to be responsible consumers of care (52);
• review and share how increased health positively affects productivity, profit, absenteeism, and other business functions (62); and
• support work-life balance to positively affect health of employees and their families (70).

These three statements appear close to one another on the map because the participants felt they were similar in meaning.

Responses that are quite different appear far away from each other on the map, like these three:

• review and share how increased health positively affects productivity, profit, absenteeism, and other business functions (62);
• support access to quality healthcare for all, regardless of income or immigration status (85); and
• be able to frame health as a form of lifelong learning through interaction and community resources (53).

These appear far away from each other on the map.

The researcher studied the point map and decided that the data is well described using eight clusters. They are labeled below as Families, Corporate wellness, Rural health, Poverty, etc.

Each cluster represents a set of ideas that together reflect a common theme. The Families cluster included these statements:

• fund early literacy programs for children and families, and track early-childhood literacy changes (7);
• screen students for not just physical issues, but also emotional, social, and behavioral delays (36);
• focus on better family functioning and parental mental health and wellness as foundations of healthy development of the whole child and family (39); and
• expand efforts on wellness education programs in the community that involve pre-k children (91).

After the cluster map is explored, it is time to overlay the ratings data. The researcher used “importance” to understand the current value of the statements contributed using a scale from 1 to 4. The ratings data can be added to the cluster map using layers to indicate ratings values, as shown here.

The cluster rating is the average for that cluster, an average of all statements in the cluster based on all participants’ ratings. The Cluster Legend below indicates that participants rated importance on the higher end: cluster ratings are between 2.69 and 3.01. The cluster rating map shows differences with layers, one for each of the 5 segments in the legend.

Cluster Rating Map

This is a Cluster Rating Map. Using layers, it shows the average rating for each cluster. Families has the highest rating. Three of the clusters have the lowest ratings on importance.

Take this quiz to check your knowledge. Then close this lesson and make your recommendation for Sally.

Knowledge Check