Skip to contents

Sensitivity Summary

summarize_sensitivity.Rd

Returns a data.frame or Kable table with summary measures of sensitivity

Usage

summarize_sensitivity(
  weights = NULL,
  Y = NULL,
  Z = NULL,
  b_star = 0,
  estimate = NULL,
  SE = NULL,
  unweighted = NULL,
  sigma2 = NULL,
  estimand = "ATT",
  pretty = FALSE,
  svy_srs = NULL,
  svy_wt = NULL,
  sig.fig = 2
)

Arguments

weights

Vector of estimated weights

Y

Outcome of interest

Z

Treatment assignment (Not needed for settings when users are analyzing surveys)

b_star

Killer confounder threshold. If not specified, will be automatically set to 0.

estimate

(Optional) Weighted point estimate. If not specified, function will automatically generate the weighted estimator, given the inputs in Y and weights

SE

(Optional) Standard error associated with the weighted point estimate

unweighted

(Optional) Unweighted point estimate.

sigma2

(Optional) Variance of outcomes or individual-level treatment effect in PATE case. In the case of a PATE estimator, if not specified, function will automatically estimate an upper bound for the variance of the individual-level treatment effect.

estimand

Specifies estimand; possible parameters include "ATT", "PATE", or "Survey"

pretty

If set to TRUE, will return a Kable table. If set to FALSE, will return a data.frame.

svy_srs

Unweighted svymean object

svy_wt

Weighted svymean object

sig.fig

Significant figures to round the output to (default set to 2)

Value

Sensitivity summary

Examples

data(jtpa_women)
site_name <- "NE"
df_site <- jtpa_women[which(jtpa_women$site == site_name), ]
df_else <- jtpa_women[which(jtpa_women$site != site_name), ]

# Estimate unweighted estimator:
model_dim <- estimatr::lm_robust(Y ~ T, data = df_site)
PATE <- coef(lm(Y ~ T, data = df_else))[2]
DiM <- coef(model_dim)[2]
# Generate weights using observed covariates:
df_all <- jtpa_women
df_all$S <- ifelse(jtpa_women$site == "NE", 1, 0)
model_ps <- WeightIt::weightit(
  (1 - S) ~ . - site - T - Y, 
  data = df_all, method = "ebal", estimand = "ATT"
)
weights <- model_ps$weights[df_all$S == 1]
# Estimate IPW model:
model_ipw <- estimatr::lm_robust(Y ~ T, data = df_site, weights = weights)
ipw <- coef(model_ipw)[2]
# Estimate bound for var(tau):
vartau <- var(df_site$Y[df_site$T == 1]) - var(df_site$Y[df_site$T == 0])
summarize_sensitivity(weights = weights, 
Y = df_site$Y, 
Z = df_site$T, 
sigma2 = vartau,
 estimand = "PATE")
#>   Unweighted Unweighted_SE Estimate     SE   RV sigma_tau_bound cor_w
#> Z    1107.35        982.65  1356.66 1417.3 0.36          2897.9  0.07