Date:         Fri, 24 Jun 2005 10:42:35 -0400
Reply-To:     Peter Flom <flom@NDRI.ORG>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Peter Flom <flom@NDRI.ORG>
Subject:      Re: proc GLM output
Comments: To: zawalazingo@YAHOO.COM
Content-Type: text/plain; charset=US-ASCII

>>> Baris Sagiroglu  wrote
<<<
I need to know which variables make sense.
>>>

Good.  Statistics can't help with that.  That's a substantive
question.

<<<
 I'd love to have stepwise glm but as far as I know there's no such
thing.
>>>

Augh!  OK, David and I and others have ranted on this enough.  A search
of the archives will reveal our rants.  In short, stepwise is almost
never a good idea.  It CERTAINLY isn't a good idea when you don't know
which variables make sense.

If you want to explore the relationships of the IVs to the DV, I would
start with lots of plots.  With a continuous DV and categorial IVs, I
like parallet box and whisker plots.

<<<
I also thought about using proc reg to get the collinearity diagnostics
but I have to turn those 40 variables into dummies
>>>

You could probably write a macro to do this; probably someone already
has......

<<<
I ran proc glm with few variables too (like 4 or 5 variables) and I
still have the 'B' next to the estimates so I don't think I can get rid
of it. Thanks.
>>>

Hmmm......

This indicates some problem with the data.  Hard to say what. I've seen
this happen when I mistakenly included all the categories, rather than
leaving one as a reference, but I don't think that's what's going on
here.

HTH

Peter

Peter L. Flom, PhD
Assistant Director, Statistics and Data Analysis Core
Center for Drug Use and HIV Research
National Development and Research Institutes
71 W. 23rd St
www.peterflom.com
New York, NY 10010
(212) 845-4485 (voice)
(917) 438-0894 (fax)