Date:         Tue, 11 Nov 2008 12:40:33 -0800
Reply-To:     clyde013@YAHOO.COM
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         clyde013@YAHOO.COM
Organization: http://groups.google.com
Subject:      Re: SAS wald test
Comments: To: sas-l@uga.edu
Content-Type: text/plain; charset=ISO-8859-1

On Nov 10, 9:39 pm, clyde...@yahoo.com wrote:
> On Nov 10, 8:03 pm, stringplaye...@yahoo.com (Dale McLerran) wrote:
>
>
>
>
>
> > --- On Mon, 11/10/08, clyde...@YAHOO.COM <clyde...@YAHOO.COM> wrote:
>
> > > From: clyde...@YAHOO.COM <clyde...@YAHOO.COM>
> > > Subject: Re: SAS wald test
> > > To: SA...@LISTSERV.UGA.EDU
> > > Date: Monday, November 10, 2008, 4:53 PM
> > > On Nov 10, 1:20 pm, stringplaye...@yahoo.com (Dale McLerran)
> > > wrote:
> > > > --- On Mon, 11/10/08, clyde...@YAHOO.COM
> > > <clyde...@YAHOO.COM> wrote:
>
> > > > > From: clyde...@YAHOO.COM
> > > <clyde...@YAHOO.COM>
> > > > > Subject: SAS wald test
> > > > > To: SA...@LISTSERV.UGA.EDU
> > > > > Date: Monday, November 10, 2008, 9:35 AM
> > > > > I have one data set consisting of 2 groups.
> > > Group 1 f=1 and
> > > > > Group 2
> > > > > f=0.  I am looking to compare parameter estimates
> > > for these
> > > > > two groups
> > > > > with a Wald Chi-square test statistic.  Can
> > > someone specify
> > > > > how I'd go
> > > > > about doing that?
>
> > > > > For example, right now my code for the regression
> > > looks
> > > > > like this:
>
> > > > > proc reg data=gg11;
> > > > > model retpi = dummy; by f; run;
>
> > > > > what I need to do is compare the coefficient on
> > > dummy for
> > > > > when f=1 and
> > > > > when f=0 using a wald test. Thanks in advance for
> > > any help!
>
> > > > Clyde,
>
> > > > This appears to be some sort of homework problem, so I
> > > won't
> > > > give you all the details of how to get a Wald
> > > chi-square test.
> > > > However, I feel it is safe to provide some direction
> > > for where
> > > > to look.  Take a look at the documentation of the
> > > GENMOD
> > > > procedure, and especially at the sections
>
> > > >     SYNTAX --> MODEL Statement
> > > >     DETAILS --> Type 3 Analysis
>
> > > > Good luck,
>
> > > > Dale
>
> > > > ---------------------------------------
> > > > Dale McLerran
> > > > Fred Hutchinson Cancer Research Center
> > > > mailto: dmclerra@NO_SPAMfhcrc.org
> > > > Ph:  (206) 667-2926
> > > > Fax: (206) 667-5977
> > > > ---------------------------------------- Hide quoted
> > > text -
>
> > > > - Show quoted text -
>
> > > Thanks.  I tried the following code and am still getting
> > > the wrong
> > > statistic.  What am i forgetting??
>
> > > proc genmod data=gg11;
> > > class f;
> > > model retpi=dummy /type3 wald ;run;
>
> > > Is this doing what I need it to be? Is it comparing the
> > > parameter
> > > estimate on the dummy variable for class f=1 to f=0??
>
> > Clyde,
>
> > You are close.  I think I can assist further.  You have named
> > f on the CLASS statement.  But where does f appear in your
> > model?
>
> > Additional hint: you need to compare both intercept and slope
> > parameters, right?  So f needs to modify two parameters of
> > your model.  How can you get f to modify both slope and
> > intercept?
>
> > HTH,
>
> > Dale
>
> > ---------------------------------------
> > Dale McLerran
> > Fred Hutchinson Cancer Research Center
> > mailto: dmclerra@NO_SPAMfhcrc.org
> > Ph:  (206) 667-2926
> > Fax: (206) 667-5977
> > ---------------------------------------- Hide quoted text -
>
> > - Show quoted text -
>
> I'm not sure if this is what you mean, but I have tried:
>
> proc genmod data=gg11;
> class f;
> model retpi=dummy dummy*f /type3 wald ;run;
>
> I do not get the desired statistic doing that either.  I am not
> comparing the intercept, just the coefficient on the dummy variables
> for 2 classes.  Thanks so much for you help!- Hide quoted text -
>
> - Show quoted text -

Anyone?