different parameter estimates in SPSS, R and SAS (GENLOG/GENLIN)

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

different parameter estimates in SPSS, R and SAS (GENLOG/GENLIN)

Filip
This post has NOT been accepted by the mailing list yet.
Hello,

I am learning loglinear regression using Agresti's textbook. I use one of the datasets that he discusses (three dummy variables, no missing data, count).
When I run the 'glm' function in R I get the same parameter estimates as he got (in SAS). However, in SPSS when I run GENLOG on that (weighted) dataset, I get different parameter estimates than I got in R and he got in SAS. (Specifically, the intercept is the same but other estimates are very different). The same happens when I run GENLIN (using count as the response variable, unweighted dataset); the estimates from GENLIN are identical to those from GENLOG but different to the estimates from R and SAS). I have tried to reverse the order of values in GENLIN (order=descending) but that didn't change anything. Two questions here:

1) SPSS only gives me parameter estimates for some parameters but not for others ('This parameter is set to zero because it is redundant.'). R and SAS give estimates for all parameters. Is there a way of forcing SPSS to produce estimates for all parameters?

2) Have I done anything wrong?

Thanks in advance!

PS. I use SPSS 22.
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: different parameter estimates in SPSS, R and SAS (GENLOG/GENLIN)

David Marso
Administrator
Here is a ghost from circa 1990 when I used to use LOGLINEAR and HILOGLINEAR very frequently.
Perhaps see if those give you similar migrainal symptomology.

Filip wrote
Hello,

I am learning loglinear regression using Agresti's textbook. I use one of the datasets that he discusses (three dummy variables, no missing data, count).
When I run the 'glm' function in R I get the same parameter estimates as he got (in SAS). However, in SPSS when I run GENLOG on that (weighted) dataset, I get different parameter estimates than I got in R and he got in SAS. (Specifically, the intercept is the same but other estimates are very different). The same happens when I run GENLIN (using count as the response variable, unweighted dataset); the estimates from GENLIN are identical to those from GENLOG but different to the estimates from R and SAS). I have tried to reverse the order of values in GENLIN (order=descending) but that didn't change anything. Two questions here:

1) SPSS only gives me parameter estimates for some parameters but not for others ('This parameter is set to zero because it is redundant.'). R and SAS give estimates for all parameters. Is there a way of forcing SPSS to produce estimates for all parameters?

2) Have I done anything wrong?

Thanks in advance!

PS. I use SPSS 22.
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me.
---
"Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis."
Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?"
Loading...