|
Question
27.2
The linkages
near the bottom of the dendogram suggest that ratings of similarity
likely varied as a function of:
 | number
of letters in the words |
 | typical
habitat |
 | type
(i.e. class) of animal |
A technical problem
with uncovering the underlying structure of ratings involves the
general lack of criteria for determining the optimum number of
clusters (i.e. a stopping rule). Comparisons between cluster
solutions obtained in patients and in healthy adults suggest that
between 8 and 10 categories of sensory experience serve to adequately
differentiate among various states of breathing discomfort.
| Figure
27.2 Hierarchical Clustering Solution for Dissimilarity
Judgments (n=100) Obtained between Pairs of Descriptors
of Breathlessness
|
|---|
|
|
| Reprinted
by permission of Chest. http://www.chestjournal.org |
|
Work
will likely continue to elucidate the precise set of phrases and
combinations that best reveal patient experiences of breathing
discomfort. Future applications of clustering algorithms to explore
not only the sensory but also the affective, evaluative, and temporal
characteristics of descriptors may affect the number of clusters
in any final solution.
|