Simulate GLMM responses from conditional distribution — simulate_glmm

Internal helper that simulates responses from the conditional distribution of Y|b (given random effects) for GLMMs. This implements the correct theoretical approach where residual variance is inherent in the distribution.

Usage

simulate_glmm_response(
  yhat,
  family,
  link,
  weights = NULL,
  use.probability = FALSE
)

Arguments

yhat: Matrix of linear predictors (fixed + random effects).
family: Character string specifying the distribution family.
link: Character string specifying the link function.
weights: Optional weights (for binomial with n-trial parameter).
use.probability: For binomial, return proportions (0-1) instead of counts?

Value

Matrix of simulated responses with same dimensions as yhat.