Causation and Construction
Mon, 19 Feb 1996 17:34:25 -0500

In as much as PCP concerns Man-the-Scientist, it is appropriate that we
consider the assessment of causal relations among constructs. As a
scientist, the person's constructions are products of his complance to
causal theories.These constructions are determined by theoretical assumptions
and empirical data. This does not mean his theory can not be revized. So long
as construction is predicated on scientific assumptions, however, there will
be determination of constructions. This is how hypotheses are derived. Breaks
in the determination occur when theory needs revision.

In PCP the therapist probably should be testing his theory of a client's
constructions. It is conceivable that LISREL could be used in something like
a grid in order to do this. The multiple group factor analysis procedure in
Circumgrids could be used to do this as well. However, more useful would be
an inductive strategy that simply spells out the logical sequences in a
client's construction.

If a person rated self, mother and wife on -say- 50 constructs, could a PCP
therapist analyze the numbers and determine whether or not the person's views
of wife are determined by his views of mother and/or self? You can not do
this with LISREL nor with correlations. You can not infer causation from
correlations. You apparently can do this, however, with the method of
Corresponding Regressions. Think about the problem and its potential. Try to
bracket your very appropriate doubts and any cowardly diagnoses that might
obstruct your circumspection. I'll post the solution soon. Thoughts and
questions would be appreciated.

William V. Chambers