Search Archives

Hi all, This is just to inform anyone who might be interested that Phillip Alday and I have developed a small R package that adds support for models with multiple membership random effects to lme4. Multiple membership models have been...

Hi Shira, I'm glad you're finding it useful! If I understand correctly, the model you've run already includes continuous covariates (e.g. (x | indicators)), so I think your question is mostly about the categorical groups? Your intuition...

Hi Shira, I'll just answer your questions in order: 1. For the generalized linear mixed model presented in the lmerMultiMember vignette, both anova() and lmtest::lrtest() perform a Chi-square test, so there's no meaningful difference other than...

Hi, Just to expand on Ben's last email: In principle, lmerMultiMember allows you to pass arbitrary indicator/weight matrices for the random effects to lme4 for model fitting as long as they have the correct shape. The package contains...

In case it's helpful to anyone following this email thread: I wrote a vignette explaining how to fit a Bradley-Terry model in lme4 using lmerMultiMember. You can find it at https://jvparidon.github.io/lmerMultiMember/articles/bradleyterry_models...

Hi Ben, It seems like Python's statsmodels might not be using EM either, which would mean that exactly none of the four stats packages listed in the original Wikipedia article actually use EM. My edits were a mix of...