1

I am using the R packages lme4, merTools and Amelia to run a multi-level model with multiply imputed data. The merTools function "print.summary.merModList" gives the model results, including for the fixed effects part of the model (it gives the estimates, standard errors and t values for the fixed effects). However, I need to report p values for the fixed effects. I understand the problem with p values for multi-level models and will focus on confidence intervals in my interpretation, but I need to report the p values out of convention.

I see that there are several options to display p values in the lme4 package (derived using Satterthwaite approximations). However, I cannot seem to find a a way to do this with multiply imputed data when the model results are in the form of a merModList object (from merTools). It would also be helpful if the confidence intervals displayed automatically, rather than having to calculate them. Does anyone know if either of these things is possible?

Thoraya
  • 11
  • 1
  • 1
    There is a mailing list dedicated to mixed models in R which may be the place to ask. –  Jul 08 '20 at 12:23
  • Hi there, lme has let to quite a few questions on Stack Overflow, the programming site, because it can be quite esoteric to get the model working as planned in R. I suggest also asking in that community, so voting to shift this to Stack Overflow. – Michelle Jul 11 '20 at 12:13
  • Don't follow the convention. P values are not reported for good reasons. – Robert Long Aug 05 '20 at 17:36

0 Answers0