Skip to contents

AIPW (Doubly Robust) Estimator

Source: R/observational.R
est_via_aipw.Rd

Augmented IPW estimator. Consistent if either the outcome model or the propensity model is correctly specified.

Usage

est_via_aipw(obj, ...)

# S3 method for class 'observational'
est_via_aipw(
  obj,
  outcome_model = NULL,
  propensity_model = NULL,
  treated_pred = NULL,
  control_pred = NULL,
  propensity = NULL,
  ...
)

Arguments

obj

An observational or clustered object.

...

Additional arguments passed to methods.

outcome_model

A model wrapper with fit and predict. Default uses OLS.

propensity_model

A model wrapper with fit and predict_proba. Default uses logistic regression.

treated_pred

Optional pre-computed treated outcome predictions.

control_pred

Optional pre-computed control outcome predictions.

propensity

Optional pre-computed propensity scores.

Value

A cm_result object.