You are here

Question about error codes in general

2 posts / 0 new
Last post
mirusem's picture
Joined: 11/19/2018 - 19:32
Question about error codes in general

So it looks like for umxConfint() it doesn't report error codes by default; I am wondering, if there were no NaNs, does that necessarily mean there are no errors? I am wondering if I should specifically look at the error codes even if estimates/parameters were returned without an issue. If so, are there specific error codes in which are certainly unacceptable [or are all of them aside from 0?] (and should it be checked for every type of model, even if they ran successfully without any explicit warning in umx)? I have read the list at, but it is not clear which ones are technically okay (other than when it explicitly gives an error).

Thanks again!

AdminNeale's picture
Joined: 03/01/2013 - 14:09
0 and 1 usually ok

Codes 0 and 1 are usually ok. There are circumstances with strong local minima (mixture distribution models in particular) where even these are false positives.

There are usually codes generated for the likelihood-based CIs, which appear as ! to the right of the CI estimates.. More info can be obtained from summary() if the argument verbose=T is included.