3
$\begingroup$

I have yet to find an answer to this problem, so here goes.

I am fitting a robust mixed regression model using the robustlmm package on R. Unlike the lme4 package, from which we can obtain p values using the lmerTest package, the robustlmm package provides no method of obtaining p values from my model.

Does anyone have any experience with this issue? How can we calculate p values from a robust mixed model?

$\endgroup$
1
  • 5
    $\begingroup$ Don't. P values are the devil's work and are not outputted by the software for good reasons (the same reasons lme4 does not output them). Resist any request from reviewers or editors for their inclusion and interpret the effect sizes instead. $\endgroup$ Aug 31 '20 at 13:04
3
$\begingroup$

If you really, really want to obtain p values, then one approach you can take is to fit the model with robustlmm::rlmer and an identical model with lmerTest::lmer and use the degrees of freedom reported by the latter to compute the p value for the t statistic reported by the former:

 pt(t.from.robustlmm, df.from.lmerTest, lower.tail = FALSE) * 2

Note that the degrees of freedom are an approximation, and as such the p value obtained should also be considered an approximation and therefore great care should be taken in it's use. Note also that the degrees of freedom can, somewhat absurdly, be fractional.

$\endgroup$
1
  • $\begingroup$ @bagel does this answer your question ? If so, please consider marking it as the accepted answer. If not then please let us know why $\endgroup$ Sep 26 '20 at 18:22

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.