|Date: ||Thu, 1 Jul 1999 00:08:44 -0500|
|Reply-To: ||"Nichols, David" <nichols@SPSS.COM>|
|Sender: ||"SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>|
|From: ||"Nichols, David" <nichols@SPSS.COM>|
|Subject: ||Re: GLM|
How did you specify that IV? If you put it in the factor box, it's being
treated as categorical, just as if you specified it as a class variable in
SAS. If you specify it as a covariate, it won't be treated that way. If
you're using command syntax, you put it after a WITH on the GLM command, not
after a BY, e.g., GLM V1 TO VK WITH IV, etc.
Principal Support Statistician and
Manager of Statistical Support
> -----Original Message-----
> From: Annette Towler [SMTP:towleraj@RUF.RICE.EDU]
> Sent: Wednesday, June 30, 1999 7:38 AM
> To: SPSSX-L@LISTSERV.UGA.EDU
> Subject: GLM
> Hi there
> SAS users tell me that PROC GLM allows analysis of a factorial design that
> includes a continuous variable as an IV. Previously, I was running
> Regression so as not to perform a dichotomous split on my continuous
> variable. I have an old version of SPSS, which does not have GLM.
> Having been given this advice, I ran my data using SPSS's GLM function
> (repeated measures) and discovered that it is treating the continuous
> variable as categorical. I think that is what is happening because this
> variable is eating up 21 degrees of freedom. As a result I have lost some
> findings I obtained using the regression.
> Is there any way around this or will I need to revert back to using
> regression? Sorry this seems like a stupid question.
> Thanks in advance.
> Annette Towler
> Psychology Department
> Rice University
> 466 Sewall Hall, MS 25
> 6100 Main
> Houston, Texas 77005-1892