Date:         Thu, 4 Nov 1999 19:10:19 -0500
Reply-To:     HERMANS1 <HERMANS1@WESTAT.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         HERMANS1 <HERMANS1@WESTAT.COM>
Subject:      Re: Estimating effects for interactions in PROC LOGISTIC
Comments: To: "Cande V. Ananth" <ananthcv@EPI.UMDNJ.EDU>
Content-Type: text/plain; charset=US-ASCII

I must be missing something subtle here.  I don't see why you would not user
PROC LOGISTIC or PROC GENMOD
to estimate the parameters of the model, including the interactions.  In
Versions 7 and 8 the CLASS statement creates
the classes automatically and you can specify the interactions directly.  Sig

____________________Reply Separator____________________
Subject:    Estimating effects for interactions in PROC LOGISTIC
Author: "Cande V. Ananth" <ananthcv@EPI.UMDNJ.EDU>
Date:       11/3/99 11:29 AM

folks:

    I need to estimate odds ratios (with 95% confidence intervals)
    for 3 effects in a logistic regression model. There are several
    interaction terms among these 3 effects:

    x1 + x1*x1 + x2 + x1*x2 + x3*x3 + x3*x2

    x1 and x3 are continuous while x2 is categorical (5 levels,
    and so, there are 4 indicators for x2). I need to estimate
    odds ratios and 95% CI for each of the combinations for
    x1, x2 and x3, as dictated by the model. I can accomplish this
    task by hand (whew, I hope not !), but was wondering if anyone
    has written any macro/IML statements to do this. Thanks.

    I'm running SAS ver 8.0 on the Solaris O/S.

--
Cande V. Ananth                           ananthcv@epi.umdnj.edu
Div of Epidemiology and Biostatistics     rwj-obgyn.umdnj.edu/
Dept of OB/GYN, UMDNJ-RWJMS               Tel:  (732) 235-7940
125 Paterson St, New Brunswick NJ 08901   Fax:  (732) 235-6627