My comments were made in the context of trying to see what J. Maxwell Legg
might be meaning in refering to a probabalistic procedure of Patrick
Slater's for calculating distances between elements. In so far as I
understand it, I can't see where the probabilistic aspect comes in - unless
it's to do with the general problem of sampling error, that afflicts all
data-sampling, but especially small samples - as occur in rep grids.
That aspect of Rep. grid work does worry me. Where else would you see PCA
being done with N=10 (elements)? I do hope that Devi is right. His urging
that the person whose construing has produced the grid should be able to
recognise their construct system in the analysis does, perhaps, go a long
way towards keeping the problem under control, if not eliminating it.
Thinking of the use of grids outside a counselling context, e.g., (I pluck
the example out of the air!) in a research project aiming to assess changes
in construing as a result of a parent sensitivity training program, does
Devi's point imply that we should regard it as bad practise not to check
out the analysis with the participant? Fels like a good idea, but
immediately I'm thinking - Go back to the lab and do the analysis, then
another appointment, another home visit ... Good argument for doing it all
on a lap-top, I suppose, so as to be able to look at the the anlysis with
the research paticipant immediately.
...
What do people think?
Supplementary point; good agrument for doint he grid analuysis on
"palm-top", too! Anyone know of rep grid analysis programs for the Psion,
especially the Psion 3a?! Or any other palmtops ...
Tony Downing,
Dept. of Psychology, University of Newcastle upon Tyne, England.
A.C.Downing@ncl.ac.uk
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%