Date:         Sat, 26 Jul 1997 14:48:13 -0400
Reply-To:     Josh Kim <u24c2@WVNVM.WVNET.EDU>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From:         Josh Kim <u24c2@WVNVM.WVNET.EDU>
Organization: Regional Research Institute
Subject:      logistic regression questions

What is the syntax to print the 95% confidence intervals for the
Exp(B) (odds ratios) in a logistic regression.  I can't find the answer in any
of my manuals..
 
Second question - relating to the R statistic
 
Lets say we change one value in one case in the dependent 0,1 variable (poor no
t-poor) and recode that as a new categorical independent variable RTEST.  In ot
her words, the new indepenent variable is exactly the same as the dependent var
iable except in one case.  Shouldn't this new independent variable be a near pe
rfect predictor of the dependent variable (as would be the case in an OLS
model) and therefore shouldn't the R stastic be very very very close to 1.  I
do this and get screwy results - what's going on.
 
Thanks.. Josh Kim at JKIM4@WVU.EDU