mi impute question

[Mia Cajita]

Hello, I used mi estimate: ologit and tried to check whether the imputed data meets the parallel regression assumption (using oparallel); however, it returned an error message. If the 'original' data (i.e., the one with missing values) meets the assumption, does it follow that the imputed data will also meet the assumption? And does that hold true for other types of imputed data? I really appreciate any help you can provide.

[Rich Goldstein]

in my opinion, this is not a very important assumption; see

https://www.fharrell.com/post/impactpo/
or
https://www.fharrell.com/post/po/

to specifically answer your question, that the assumption holds in incomplete data does not tell you whether it will hold in imputed data; if you really feel that this is important here, you can probably check it within each imputed data set; see

help mi xeq

to organize a way to get at that

[Mia Cajita]

Thank you for the resources and helpful suggestion!