Phoenix
10-14-2008, 05:55 PM
Hey there, I’m really glad that this forum exists, cause I’m having an analytical issue with SPSS (Graduate Pack 14.0 for Windows) that’s got me completely stumped.
I'm currently analyzing a very large dataset with about 1.5 million cases. For one of the linear regression models I am trying to run, I include about 30 variables as predictors. The problem is that when I include all of the predictors in the model, one of them keeps getting excluded because that variable's tolerance value is too low. I soon learned that this is an automatic feature of the program, but I was wondering if there was any way I can get SPSS to keep all of the variables in the analysis (despite low tolerance values).
The analyses I need to run simply requires that all the variables be in there, because I'm trying to simply replicate a finding here. Also, I cannot collect any more data (this is a pre-existing dataset). I'm thinking that there must be some way, perhaps some option I can choose, to allow all of the predictor variables to stay in, but I haven't been able to figure out how. Is there a way, or maybe some method others have used to get around this problem?
Any help or suggestions about where to find an answer would be greatly appreciated!
Thanks,
Phx
I'm currently analyzing a very large dataset with about 1.5 million cases. For one of the linear regression models I am trying to run, I include about 30 variables as predictors. The problem is that when I include all of the predictors in the model, one of them keeps getting excluded because that variable's tolerance value is too low. I soon learned that this is an automatic feature of the program, but I was wondering if there was any way I can get SPSS to keep all of the variables in the analysis (despite low tolerance values).
The analyses I need to run simply requires that all the variables be in there, because I'm trying to simply replicate a finding here. Also, I cannot collect any more data (this is a pre-existing dataset). I'm thinking that there must be some way, perhaps some option I can choose, to allow all of the predictor variables to stay in, but I haven't been able to figure out how. Is there a way, or maybe some method others have used to get around this problem?
Any help or suggestions about where to find an answer would be greatly appreciated!
Thanks,
Phx