The University of Adelaide CENTRE FOR AUTOMOTIVE SAFETY RESEARCH

home   /   centre for automotive safety research   /   Publications   /   List   /   Details

Publication Details

TitleConsensus versus polarisation: measuring ordinal variability
AuthorsHutchinson TP
Year2005
TypeJournal Article
AbstractProposals have been made recently by Blair and Lacy and by Yager concerning how variability on an ordinal scale should be defined. The present article criticises these proposals for having some characteristics appropriate when the level of measurement is nominal rather than ordinal. It also shows that the use of numerical scores can easily lead to incorrect conclusions. An alternative idea is to compare distributions by notionally taking two random observations from one and two from the other, and saying there is evidence that the first is more variable if it is the source of both the smallest and largest of the four observations. Another alternative, a graphical method called the P-P plot, shows one cumulative distribution in terms of the other. Two datasets are used as examples. One concerns consensus in attitudes to homosexuality, and the other in the perceived location (left vs. right) of political parties.
Journal TitleAustralasian Journal of Market & Social Research
Journal Volume (Issue)13(1)
Page Range27-35
Page Count9

Reference
Hutchinson TP (2005). Consensus versus polarisation: measuring ordinal variability. Australasian Journal of Market & Social Research, 13(1), 27-35.