Re: Re; Grid Analysis

Tue, 26 Mar 1996 18:35:53 +1100 (EST)

Dear Hemant

I am working on the representatiomn of grids in terms of fuzzy sets. This
allows one to be more realistic about ratings which are after all
'neighbourhood' values. So a rating of 5 in terms of fuzzy sets can be seen
as equivalent to a rating of 6 at a certain level of fuzziness.

I am working on a method for establishing the consensual nature of a grid in
terms of the constructs expressed as fuzzy subsets. These fuzzy subsets can
be compared for similarity using matrix methods. Eventually one can
consruct a consensus matrix which shows the pairwise similarity of
constructs and provides a measure of grid consensus. This latter measure
can be used as a measure of cognitive complexity. Low consensus 'may' mean
complex structure.

The consensus matrix can be used to derive a dissimilarity matrix and then
plotted using MDS methods.

What do you think about such approaches.


Alastair Anderson

Melbourne Australia