Compare naive vs. design-based corrected estimates — dsl_cmp • craft

Runs the same regression with and without DSL correction and returns a tidy comparison data frame.

Usage

dsl_cmp(
  data,
  formula,
  predicted_var,
  prediction,
  sample_prob,
  fixed_effect = NULL,
  index = NULL,
  cluster = NULL,
  ...
)

Arguments

data: Data frame.
formula: Outcome formula. The LHS predicted_var is replaced by prediction for the naive fit.
predicted_var, prediction, sample_prob, fixed_effect, index, cluster: See dsl_fit.
...: Forwarded to dsl::dsl().

Value

A long data frame with columns term, model (one of "Original" or "DSL"), estimate, std.error, conf.low, conf.high, p.value.