Outcome-blind exposure overlap diagnostics — margot_exposure

`margot_exposure_overlap()` estimates or accepts propensity scores using only the exposure and baseline covariates. It is intended for registration-stage support checks before any outcome model, causal forest, LMTP estimator, or causal contrast is fitted.

Usage

margot_exposure_overlap(
  data,
  exposure,
  covariates,
  weights = NULL,
  method = c("supplied", "logistic", "probability_forest"),
  bounds = c(0.05, 0.95),
  plot = TRUE,
  propensity = NULL,
  grf_defaults = list(),
  seed = 12345,
  verbose = TRUE,
  ...
)

Arguments

data: A data frame containing the exposure and, when `covariates` is a character vector, the baseline covariates.
exposure: Character string naming a binary exposure coded 0/1.
covariates: Character vector of baseline covariate names, or a numeric matrix/data frame with one row per observation in `data`.
weights: Optional character string naming design weights in `data`, or a numeric vector with one value per observation.
method: Character. One of `"supplied"`, `"logistic"`, or `"probability_forest"`.
bounds: Numeric length-2 vector giving the lower and upper propensity score bounds used for overlap and trimming summaries.
plot: Logical; if `TRUE`, return a propensity-score histogram.
propensity: Optional numeric vector of propensity scores. Required for `method = "supplied"`.
grf_defaults: Optional list passed to `grf::probability_forest()` when `method = "probability_forest"`.
seed: Integer seed used by the probability forest path.
verbose: Logical; emit progress messages.
...: Reserved. Outcome-like arguments supplied here trigger an error.

Value

A list with support summaries, balance summaries, optional plot, and enough metadata to document the outcome-blind exposure diagnostic.