Extract posterior variance for single random intercept (Poisson GLMM)

[jan-engelmann]

Hi,
thanks for putting this out!
my model is

frm = @formula(counts ~ 1 +C1 + C2 + ... + C37 + (1 | genotype_id))
modelp = fit(MixedModel,frm, df, Poisson(), fast = true) 

in lme4 I can get the random mode and conditional variance like:
lme4::ranef(model, condVar = TRUE)

In MixedModels I can get the means with only(raneftables(model))

For a LMM is it correct to extract the variance as below?

  ranefs = DataFrame(only(raneftables(model)))
  rename!(ranefs, Dict("(Intercept)" => "mean"))
  ranefs.var = vec(condVar(model)[1][1, 1, :])

For a Poisson GLMM, I get the following error:

MethodError: no method matching condVar(::GeneralizedLinearMixedModel{Float64, Poisson{Float64}})

Closest candidates are:
  condVar(::LinearMixedModel{T}, ::Any) where T
   @ MixedModels ~/.julia/packages/MixedModels/L0NHA/src/linearmixedmodel.jl:324
  condVar(::LinearMixedModel{T}) where T
   @ MixedModels ~/.julia/packages/MixedModels/L0NHA/src/linearmixedmodel.jl:320


Stacktrace:
 [1] top-level scope
   @ In[74]:1

What would be the right way to calculate the posterior variance of the random intercept per group?

Thanks a lot for your help!

[jan-engelmann]

Is the below a reasonable approximation? for a poisson GLMM?

vec(condVar(model.LMM)[1][1, 1, :])