Compute doubly robust scores for a causal forest. — get_scores.causal

Compute doubly robust (AIPW) scores for average treatment effect estimation or average partial effect estimation with continuous treatment, using a causal forest. Under regularity conditions, the average of the DR.scores is an efficient estimate of the average treatment effect.

# S3 method for causal_forest
get_scores(
  forest,
  subset = NULL,
  debiasing.weights = NULL,
  num.trees.for.weights = 500,
  ...
)

Arguments

forest	A trained causal forest.
subset	Specifies subset of the training examples over which we estimate the ATE. WARNING: For valid statistical performance, the subset should be defined only using features Xi, not using the treatment Wi or the outcome Yi.
debiasing.weights	A vector of length n (or the subset length) of debiasing weights. If NULL (default) they are obtained via inverse-propensity weighting in the case of binary treatment or by estimating Var[W \| X = x] using a new forest in the case of a continuous treatment.
num.trees.for.weights	Number of trees used to estimate Var[W \| X = x]. Note: this argument is only used when debiasing.weights = NULL.
...	Additional arguments (currently ignored).

Value

A vector of scores.

References

Farrell, Max H. "Robust inference on average treatment effects with possibly more covariates than observations." Journal of Econometrics 189(1), 2015.

Graham, Bryan S., and Cristine Campos de Xavier Pinto. "Semiparametrically efficient estimation of the average linear regression function." arXiv preprint arXiv:1810.12511, 2018.

Hirshberg, David A., and Stefan Wager. "Augmented minimax linear estimation." arXiv preprint arXiv:1712.00038, 2017.

Robins, James M., and Andrea Rotnitzky. "Semiparametric efficiency in multivariate regression models with missing data." Journal of the American Statistical Association 90(429), 1995.