st: any virtues to stepwise regression?

["Christopher W. Ryan" <cryan@binghamton.edu>]

I'm familiar with the "evils" of stepwise reegression techniques that 
occasionally are discussed on Statalist.  I don't mean to stir them up 
again.  In journal club for my family practice residents, I had occasion 
to present a great example of stepwise multiple logistic regression 
producing 6 significant predictor variables for perioperative morbidity, 
while "all-at-once" regression confirmed only 4 of them.  [Lee TH, et 
al.  Derivation and prospective validation of a simple index for 
prediction of cardiac risk of major non-cardiac surgery.  Circulation 
1999; 100:1043-1049.]  It was instructive.

Are there *any* circumstances in which stepwise multiple regression 
would be the preferred approach?

Thanks.

--Chris
--
Christopher W. Ryan, MD
SUNY Upstate Medical University Clinical Campus at Binghamton
and Wilson Family Practice Residency, Johnson City, NY
cryanatbinghamtondotedu
GnuPG and PGP public keys available at http://pgp.mit.edu

"If you want to build a ship, don't drum up the men to gather wood, 
divide the work and give orders. Instead, teach them to yearn for the 
vast and endless sea."  [Antoine de St. Exupery]

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/