\magnification=1200
\baselineskip=20pt
\nopagenumbers
\font\big=cmr12 scaled \magstep2

\centerline{\bf STANFORD UNIVERSITY}
\centerline{\bf DEPARTMENT OF STATISTICS}
\centerline{\big DEPARTMENTAL SEMINAR}
\bigskip

\baselineskip=12pt
\centerline{4:15 p.m., Tuesday, October 8, 2002}
\centerline{Sequoia Hall Room 200}
\centerline{(Cookies at 3:45 in 1st Floor Lounge)}

\bigskip
\baselineskip=15pt
\centerline{\sl Tim Hesterberg}
\centerline{\sl Insightful Corporation}
\bigskip
\centerline{\bf Bootstrap Tilting and non-sampling inferences and diagnostics}
\bigskip

Bootstrap tilting condence intervals could be the method of choice in many applications
for reasons of both speed and accuracy. With the right implementation, tilting
intervals are 37 times as fast as bootstrap BC-a limits, in terms of the number of bootstrap
samples needed for comparable simulation accuracy. Thus 100 bootstrap samples
might suffice instead of 3700.

Tilting limits have other desirable properties -- second-order accuracy, transformation
invariance, and better finite-sample coverage and/or shorter intervals on average than
competing procedures.

We also discuss condence interval procedures that require no sampling, similar to
ABC intervals, but with better finite-sample properties.

Bootstrap tilting also is useful for diagnostic purposes. They provide an immediate
warning against the most common error in using the bootstrap.

S-PLUS software for these procedures is available.


\bye