Purchase this article with an account.
Trefford Simpson, Yunwei Feng; Multidimensional scaling of words used to characterize ocular surface discomfort. Invest. Ophthalmol. Vis. Sci. 2013;54(15):6039.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To use multidimensional scaling (MDS) to determine the number of dimensions required to account for the dissimilarity between typical adverbs used to characterize how the eyes feel and then position these words in this multidimensional space.
20 subjects rated the dissimilarity of all pairs of 22 “typical” adverbs that completed the phrase "My eyes are/feel...", For example, these adverbs included burning, scratchy, cold and OK. All 231 pairwise combinations of these words were rated by subjects using magnitude estimation on a scale of “exactly the same” (0) to “completely different” (100). For each subject the order of pairs was randomized and from these ratings, dissimilarity matrices were formed. These matrices were in turn analysed using a number of MDS algorithms, in particular SMACOF because it also provided a jackknife error estimation for the solution. Data were analysed using R.
Scree plots showed that 2 dimensions account for most of the data variability. Regardless of the method used the results were consistent and are illustrated in the Figure. The first dimension is a "discomfort" dimension with most of the words connoting unpleasantness generally clustered in the same region of the solution space (circles/words on the left in the fig.) and the "comfortable" words at the other end of this dimension (on the right). The second dimension is thermal, with warm (highest in the fig.) and cool (lowest in the fig.) at opposite ends.
The results are somewhat surprising since current physiological and psychophysical evidence would suggest at least 3 dimensions, one based on words characterizing mechanical (painful) attributes, one for chemical (also painful) descriptors and one based on mild temperature (particularly cool). Since subjects appear, rather, to collapse the 2 physiologically plausible pain dimensions into a single discomfort dimension, perhaps more attention could be paid to understanding subjects' interpretations of the actual terms used when measuring ocular surface discomfort before these words are used, untested, in scales.
This PDF is available to Subscribers Only