[R-sig-ME] Equivalent of STATA absorb in lme4 (or HGLM)

[J.D. Haltigan]

Hi:

I am attempting to replicate some STATA analyses in R and have the
following question. In short:
https://github.com/lme4/lme4/issues/689#issuecomment-1186346282

Thanks, this is what I was trying to convey re: STATA's 'absorb' command
but failed to do.
https://stackoverflow.com/questions/49639238/r-equivalent-of-statas-absorb
<https://github.com/lme4/lme4/issues/url>

So, in STATA, absorb basically just aggregates estimates it seems across
the factor variable. I have now changed the pairID variable upstream to a
factor class which results in estimates for all 300 pairIDs. Is there
something similar to STATA's absorb in lme4? Or would it simply be | pairID
in the formula even if the original factor 'pairID' with all 300 estimates
were 'fixed effects.' Though I doubt | pairID would work with a variable
now defined as class factor which I just confirmed. So, is anything like
STATAs 'absorb' available?


-JDD

	[[alternative HTML version deleted]]