In Praise of Statistical Theory
Persi Diaconis
Stanford University
Abstract
     Statistical theory is quite out of fashion.  It has been replaced
by an algorithmic, anything goes, "who needs theory when we have a
computer?" attitude. I will review the strengths and weaknesses of
this "new statistics" and point up places where new theory is badly
needed.
This is a report of joint work with Erich Lehmann.