I need to fit a linear mixed model but my dependent variable is has some outliers that I can't discard. Then I used the rlmer() function (robustlmm package).

All this works fine. Nevertheless, I'm trying to compare several models and I need to get the p-values. AICc, BIC and the F-values. Is there any function that does this?


In robust linear mixed models, there is no AIC and BIC given that no likelihood exists for this kind of model. Regarding p-values, they seem bad if you try to calculate them using bootstrapping methods. It's better to use confidence intervals instead. These can be calculated using wald-type tests.

| cite | improve this answer | |
  • $\begingroup$ Hi @gung, this is not a proper answer to the question since the question is about a function but the answer counts the + and - and in the end it suggests sung CI. -1 $\endgroup$ – TPArrow Apr 26 '17 at 9:31
  • $\begingroup$ @TPArrow, this isn't my answer, I only edited it. It's true that the question asked for a function, but IMHO, sometimes the answer that will most benefit an OP is to unpack the assumptions made in the question & point out that they aren't really appropriate. $\endgroup$ – gung - Reinstate Monica Apr 26 '17 at 12:18
  • $\begingroup$ @gung oh right! However, the credit should be made for a proper answer, IMHO. $\endgroup$ – TPArrow Apr 26 '17 at 16:03
  • $\begingroup$ @Wagliss, Could you provide any reference about confidence intervals are better than p-values when performing robust linear mixed models? Thanks!! $\endgroup$ – Dekike May 17 at 8:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.