Missing Data/Imputation Discussion > Using AMOS to creat MI dataset
actually my biggest question is how to model the variables in amos to get accurate imputations. just allow all covariances and the impute? that seems logical.
July 18, 2012 |
Loren
Loren
To be honest, I haven't used AMOS much for creating imputed data sets. I usually use either R (most often) or sometimes SPSS. With that said, the way you went about it seems reasonable from what I read. I've also heard of people analyzing their imputed datasets in AMOS, by using the multi group analysis feature and specifying each group as one of their imputed datasets...
July 28, 2012 |
Jeremy Taylor
Jeremy Taylor
I don't have the SPSS missing values analysis add-on. Further, I have a very small dataset and it has been recommended to me that I use Bayesian imputation. To do so, I'm trying to use AMOS to create and save the MI dataset with 5 imputed datasets. I simply created 4 observed variables (those that I want to analyze) and then launched the imputation procedure in AMOS. Seems like I got the imputed datasets I was desiring, but I'm a little unsure. Any thoughts?