Package index

Overview

A tour of the package, from weights to effect estimates.

propensity propensity-package

propensity: A Toolkit for Calculating and Working with Propensity Scores

Weights

Propensity score weights for each estimand, across binary, categorical, and continuous exposures.

wt_ate() wt_att() wt_atu() wt_atm() wt_ato() wt_entropy() wt_atc() wt_cens()

Calculate propensity score weights

Weight objects

Construct propensity score weights and read what they record.

new_psw() psw() is_psw() is_stabilized() stabilization_score() as_psw()

Propensity Score Weight Vectors

exposure_type() density_meta() format(<propensity_density_meta>) print(<propensity_density_meta>)

What a set of weights records about the exposure

numerator_model()

The model a set of weights was stabilized on

Joint treatments

Weights for a sequence of two treatments, from a factorized pair of treatment models.

joint_wt_models() is_joint_wt_models()

Record the two treatment models of a joint exposure

wt_joint() is_joint_wt() joint_wt_meta()

Product weights for a joint intervention on two treatments

Continuous exposure densities

The density families a continuous exposure’s weights are built from.

dens_normal() dens_laplace() dens_t() dens_kernel() dens_fn()

Density specifications for continuous exposures

Trimming

Restrict a set of propensity scores to a region of overlap, and refit the model on what remains.

ps_trim()

Trim Propensity Scores

ps_trim_meta()

Extract trimming metadata from a ps_trim object

is_ps_trimmed()

Test whether propensity scores have been trimmed

is_unit_trimmed()

Identify which units were trimmed

ps_refit()

Refit a Propensity Score Model on Retained Observations

is_refit()

Check if propensity scores have been refit

Truncation

Bound extreme propensity scores in place rather than dropping them.

ps_trunc()

Truncate (Winsorize) Propensity Scores

ps_trunc_meta()

Extract truncation metadata from a ps_trunc object

is_ps_truncated()

Test whether propensity scores have been truncated

is_unit_truncated()

Identify which units were truncated

Calibration and tilting

Recalibrate fitted propensity scores, and evaluate the tilting function an estimand implies.

ps_calibrate()

Calibrate propensity scores

is_ps_calibrated()

Check if propensity scores are calibrated

ps_tilt()

Propensity score tilting functions

Effect estimation

Inverse probability weighted estimation and the surfaces its results are read through.

ipw(<joint_wt_models>) ipw(<multinom>) ipw(<lm>) ipw(<glm>)

Inverse Probability Weighted Estimation

print(<ipw_diagnostic_se>) as.data.frame(<ipw_diagnostic_se>) tidy(<ipw_diagnostic_se>)

Print a result whose standard errors are a diagnostic

tidy(<ipw>)

Tidy an inverse probability weighted result

glance(<ipw>)

Glance at an inverse probability weighted result

augment(<ipw>)

Augment an inverse probability weighted result with per-observation columns

tidy(<ipw_pooled>)

Tidy a pooled inverse probability weighted result

glance(<ipw_pooled>)

Glance at a pooled inverse probability weighted result

Re-exports

reexports ipw as_marginal as_conditional is_causal_wt estimand estimand<- tidy glance augment pool_ipw

Objects exported from other packages