Pool regression estimates using Rubin's Rules

Usage

pool_estimates_by_term(
  combined_results,
  exposure_terms,
  reference_model,
  model_type = "mixed"
)

Arguments

combined_results: Data frame with term, estimate, and std.error columns
exposure_terms: Character vector of exposure terms to pool
reference_model: A fitted model to extract degrees of freedom
model_type: One of "fixed", "random", or "mixed". Used to select how residual degrees of freedom are extracted. Defaults to "mixed".

Value

Data frame with pooled estimates and confidence intervals. When combined_results contains a y.level column (multinomial models), estimates are pooled per outcome level and the column is retained.