"All models are wrong; some are useful".
He would also agree that model fitting is an iterative process of forming
hypeotheses, gathering data, analyzing data, and revising the hypotheses.
I think that's a healthy attitude toward modeling. Models are approximations.
You can never perfectly model reality; you may be able to model it well
enough to use if for something, but you should never forget that the model
does not equal reality.
My $.02 worth.
Karen Pitts
kpitts@sarnoff.com
Hopewell Presbyterian Church, Hopewell, NJ, teacher of NT Greek
Sarnoff Corporation, Princeton, NJ, statistician