SMS scnews item created by Rafal Kulik at Wed 2 May 2007 0851
Type: Seminar
Distribution: World
Expiry: 8 Jun 2007
Calendar1: 8 Jun 2007 1405-1455
CalLoc1: Carslaw 373
Auth: rkuli(.ststaff;2434.3001)@p818.pc.maths.usyd.edu.au

# Statistics Seminar: Kabaila -- Valid inference after preliminary statistical model selection

               **********************************************
*                                            *
*          UNIVERSITY OF SYDNEY              *
*                                            *
*     SCHOOL OF MATHEMATICS & STATISTICS     *
*                                            *
*       STATISTICS SEMINAR SERIES - 2007     *
*                                            *
**********************************************

****************************
*        SEMINAR NOTICE    *
****************************

--------------------------------------------------------------------------

Valid inference after preliminary statistical model selection

Paul Kabaila (La Trobe University)

Friday, 8 June, 2.00pm

Carslaw 373

--------------------------------------------------------------------------

It is common in applied statistics to carry out a preliminary
model selection using e.g. AIC, BIC or hypothesis tests. This is usually
followed by the statistical inference of interest, using the same data,
based on the assumption that the selected model had been given to us a
priori. This assumption is false and it leads to invalid inferences.
Breiman (1992) calls this "a quiet scandal in the statistical
community". The purpose of the talk is to (a) describe the depth of this
scandal and (b) to describe the speaker’s proposed solution to this
problem.

---------------------------------------------------------------------------