Grid analyses
Rob Franz (RFRANZ@UKANVM.EARN)
Fri, 03 Jun 94 12:39:15 CDT
Hello network--
I've recently been conducting some analyses of grids that I collected using
OMNIGRID Version 1.5. I have found some things that puzzled me and hoped that
someone out there might have some good advice. Anyway, here goes:
First, I have done some principal components analyses using SPSS on 21 grids
(20 X 20 each). The variance accounted for by the first factor (PVAFF) has
correlated .99 (Spearman's rho) with Intensity as calculated by OMNIGRID.
OMNIGRID calculates intensity by squaring each possible construct
intercorrelation, multiplying it by 100, and then summing over them. However,
the percentage of variance accounted for by the first three factors correlates
only .70 with Intensity. If the PVAFF - Intensity correlation were artifactual
I would expect the other correlation to be much greater, wouldn't I? I am also
puzzled because Feixas et al. (IJPCP, 1992) reported a correlation of only .25
between PVAFF and Intensity. However, Bannister & Fransella (1977) did state
that the correlation between the two indices should be quite high -- but .99?
Any thoughts regarding 1) the magnitude of the correlations or 2) the relative
merits of using PVAFF as a second measure of complexity versus using the
variance accounted for by the first three factors??
Second, I have also obtained a rather high correlation between Intensity and
Variability of Intensity (again, in the .9s). Radley did state that in a
relatively "normal" population (e.g. college students, as in my sample?) that
the two measures should be closely related -- but again, the magnitude of the
correlation makes me edgy. Theoretically, if I understand the two constructs,
a highly "intense" person would probably not have a very high variability of
intensity, would they? After all, it seems that if a person is for all intents
and purposes rating others along a global, single dimension (very high
Intensity) then that person should not demonstrate a network of interrelated
"islands" of constructs (which is how I understand Variability of Intensity).
Or am I missing something here? Again, any thoughts would be greatly
appreciated!!
Thanks in advance for your consideration and thoughts |)
Rob Franz
Department of Psychology
University of Kansas
Lawrence, KS, USA 66045
RFRANZ@UKANVM.EDU.CC.UKANS