|
|
Hello,
SPSS20 user's guide says that pooled information about Linear Mixed Models procedure is available.
When I run MI using this procedure on a imputed dataset, splitting the file by imputation number, I get pooled results for a few tables (I consider pretty much irrelevant).
The relevant information I have to report in the paper, like tests for fixed effects, or the information criteria, are not pooled.
What would you recommend? Picking just one imputed dataset at random and then reporting these results kind of stands against the whole idea of MI, and averaging the results over all imputations "manually" also, I guess.
Thank you for the help
E
|
Locked
