Date:         Thu, 1 Aug 2002 09:19:02 +0200
Reply-To:     Asesoría Bioestadística
              <bioestadistica@eresmas.net>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Asesoría Bioestadística
              <bioestadistica@eresmas.net>
Subject:      Re: question regarding PIN & POUT criteria in logistic regression
Content-Type: text/plain; charset=us-ascii

Hi Bob:

Try this:

/METHOD=ENTER
/CRITERIA PIN(.05) POUT(.10) bcon(0.000001) lcon(0.0000001) ITERATE(20)
CUT(.5) .

This will force the program to go on iterating until these very stricts
conditions are met. If the OR are tending to infinity (due to empty cell
patterns), you will get a result very different form your previous one. You
might even not reach convergence in 20 iterations, showing that the model is
crashing.

But, I am surprised, as:

/METHOD=ENTER
/CRITERIA PIN(.05) POUT(.10) ITERATE(20) CUT(.5) .

and this:

/METHOD=BSTEP(LR)
/CRITERIA PIN(1) POUT(1) ITERATE(20) CUT(.5)

Should give the same OR!. This last syntax is used to get all the LR p-values,
but the coefficient estimates should not change, unless there is some other
change in the syntax you have not included.

Best regards

Marta