misscalculation of predicted factor scores in SEM

nebojsa · Post by **nebojsa** » Tue Sep 19, 2023 6:34 am

I run the SEM and my model works on 1878 cases out of 2415 because of listwise deletion. However, when I saved predicted factor scores Jamovi just pasted it in data columns regardless of which case they belong to. It just attached it to the first 1878 cases. It looks to me as a bug, especially because there are only few missing data in manifest variables comprising factors for that part of the file?
How to solve it?
Please help!

mcfanda@gmail.com · Post by **mcfanda@gmail.com** » Tue Sep 19, 2023 11:19 am

HI
yes, it was a bug. It has been fixed in the new version 1.1.7, which is coming out. You'll see now that cases with some missing in the observed variables will have all predicted values missing if you select "missing listwise", whereas the cases will have missing only in the predicted values corresponding to the missing observed variables when other method of handling missing are chosen.

nebojsa · Post by **nebojsa** » Tue Sep 19, 2023 3:07 pm

Many thanks for your effort! I tried it again but it unfortunately still does not work. I got the same results no matter of missing data treatment.

mcfanda@gmail.com · Post by **mcfanda@gmail.com** » Tue Sep 19, 2023 8:20 pm

if you are in SEMLj 1.1.7 and it does not work, I'd ask to share an example so I can work it out, because in my system works fine. If you prefer, you can send a .omv file via email

jamovi