At 06:19 14/05/99 +1000, you wrote:
>Is the idea of a phantom construct or element as in options, 2 and b, to
>cancel out the effect of the value in the original construct or element?
Not so much a cancelling out as permitting an element to have a dual
relationship with a construct.
>
>If you had a lot of such values (e.g 6), could you still use such a
>strategy, or would the results be too bodgey?
The more you use the strategy in a grid the more you will impose
artefactual constraints on the solution. For the both pole situation it
will make the configuration more perfectly circular - see your ref from
another mailing
Russell J and Carroll J (1999). On the Bipolarity of Postive and Negative
affect. Psychological Bulletin. 125(1), 3-30.
[by the way thanks for that - it's one of my many talks in berlin and I
hadn't caught up with that reference]
For the neither pole situation I'm not sure what would happen - chaos is a
good bet.
Richard
PS I'll get to your gridstat/gpack question soon. Have just fixed up a
couple of GRIDSTAT bugs that Sharon Micken & you have picked up.
Richard C. Bell
Associate Professor
Department of Psychology
University of Melbourne
Parkville Vic 3052 Australia
ph: +61 3 9344 6364
fax: +61 3 9347 6618
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%