Binary logistic regression with multiple imputed data: how to get pooled estimates?

[tamtam1]

Dear all,

I have performed multiple imputation on my dataset using SPSS 20, giving me 10 different datasets. I have understood that you could easily perform logistic regression on these and that you would get the pooled estimates (saw some examples of this as well, but only with the enter method).
Now I want to use the stepwise forward method, but unfurtunately I get the results for the 10 different datasets, but no pooled estimates. Does enyone know how this could be?

Thanks in advance!

 

[therp]

Re: Binary logistic regression with multiple imputed data: how to get pooled estimate

hey!
i am no expert in MI but i had the same problem with correlations on imputed data with SPSS18 (Portable version). I then found out that apparently you need a specific license for getting pooled results. I run the analysis on a university computer and got the pooled results there.
hope this helps,
therp