Jerry,
You are correct that the same data set can support a number of SEM models.
This is not the case with corresponding correlations/regressions, however,
CR discloses THE causal model for a given set of data, To debate whether or
not this disclosure is provable or not when we model phenomena is pointless,
But it is possible that when we use numbers to simulate causal patterns
that proofs can be developed for corresponding regressions restricting the
proof to numbers, Thus if we constrain our endeavor to numbers then we can
develop proofs of the validity of corresponding regressions, We can prove
that CR reveals that y=x1+x2 when x1 and x2 are uniform random numbers, This
keeps the endeavor within the domain of numbers and proof that x1 and x2
cause y is possible within that domain, This is the best a statistician can
do,as far as proof goes,
By the way, I finally got my McAfee update and my computer and files are
virus free, Anybody, including Frank Harrell and Rich Ulrich can now have
my most recent paper. Maybe Rich and Frank will demonstrate that their
"expertise" is more than knee jerk reactions of canned statisticians,
Bill Chambers
===========================================================================
This list is open to everyone. Occasionally, people lacking respect
for other members of the list send messages that are inappropriate
or unrelated to the list's discussion topics. Please just delete the
offensive email.
For information concerning the list, please see the following web page:
http://jse.stat.ncsu.edu/
===========================================================================
