Interpreting MDS Graphs
It’s convenient to think of the MDS graph in figure 16.3 as a sort of mental map. That is, it represents what the informant was thinking when he pile sorted those fruits. I say ‘‘sort of mental map’’ because MDS graphs of pile-sort data are not one-to-one maps of what’s going on inside people’s heads. We treat them, however, as a rough proxy for what people were thinking when they made piles of cards or words or whatever.
Looking at figure 16.3, it looks like there’s a citrus cluster, a berry cluster, a melon cluster, a fruit tree cluster, and a tropical cluster. We can check our intuition about these clusters by running a cluster analysis, shown in figure 16.6.
Cluster analysis of 18 fruits from one pile sort.
SOURCE: H. R. Bernard and G. W. Ryan, Analyzing Qualitative Data: Systematic Approaches. Los Angeles: Sage Publications. 2010. p. 177. Used by permission.
Read figure 16.6 as follows: At the first level of clustering, the informant put #7 (watermelon) and #12 (cantaloupe) together, and he put #6 (blueberry) and #10 (strawberry) together. These two clusters together form a cluster at the second level. And the same goes for the other clusters: They come together at the second level and all form one big cluster.
Because there is just one informant, there can only be two levels. The first level is the level at which the informant made the separate piles. The second is the entire set of fruits. I’ve taken you through what looks like a trivial exercise to show you how to read the cluster diagram (or dendrogram) and the MDS picture. As we’ll see next, things get more interesting when we add informants (box 16.4).