runehaubo
commented
2 years ago Hi Charity,
With clmm you can only fit variance structures like those of lmer - not all those of lme, ie. variance structures on the random effects, not those on residuals. And usually the concern in repeated measures analyses is the variance structure on the residuals (unstructured, compound symmetry etc). But In ordinal regression there are no residuals so the usual thinking from linear world does not apply here and there is no equivalent to lme. Some of the time-to-time variation can potentially be handled by careful specification of random-effect structures but how to do this, and the extent to which it can be done, is not an easy task and sufficiently unclear to me that I cannot give further concrete advice.
Cheers Rune
On Thu, 11 Nov 2021 at 22:53, Charity Goeckeritz @.***> wrote:
Hi Rune,
I've got a somewhat complicated dataset with repeated measures within 8 collection timepoints. I think, for the most part, I am beginning to understand how to form my model -- but I'm very confused as to how I show clmm2 that I have repeated measures. I'm thinking the correlation of repeated measures within each time point may differ from each other a bit, so I need to try an unstructured variance-covariance structure; I was thinking compound symmetry. Can clmm2 handle this? And if so, how??
Any help is much appreciated -- thank you so much for maintaining this package!
Kindly, Charity
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/runehaubo/ordinal/issues/47, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABHZZLLI6NX77ES2FCZN7ADULQ3L5ANCNFSM5H3NVE6Q . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.
Hi Rune,
I've got a somewhat complicated dataset with repeated measures within 8 collections. I think, for the most part, I am beginning to understand how to form my model -- but I'm very confused as to how I show clmm2 that I have repeated measures. I'm thinking the correlation of repeated measures within each collection may differ from each other a bit, so I need to try an unstructured variance-covariance structure; I was thinking compound symmetry. Can clmm2 handle this? And if so, how?? I.e., what does the syntax look like?
Any help is much appreciated -- thank you so much for maintaining this package!
Kindly, Charity