Confidence Intervals for IPW Estimates of the Causal Effects of Stochatic Shift Interventions

# S3 method for ipw_haldensify
confint(object, parm = seq_len(object$psi), level = 0.95, ...)

Arguments

object: An object of class ipw_haldensify, produced by invoking the function ipw_shift, for which a confidence interval is to be computed.
parm: A numeric vector indicating indices of object$est for which to return confidence intervals.
level: A numeric indicating the nominal level of the confidence interval to be computed.
...: Other arguments. Not currently used.

Value

A named numeric vector containing the parameter estimate from a ipw_haldensify object, alongside lower/upper Wald-style confidence intervals at a specified coverage level.

Details

Compute confidence intervals for estimates produced by ipw_shift.

Examples

# simulate data
n_obs <- 50
W1 <- rbinom(n_obs, 1, 0.6)
W2 <- rbinom(n_obs, 1, 0.2)
W3 <- rpois(n_obs, 3)
A <- rpois(n_obs, 3 * W1 - W2 + 2 * W1 * W2 + 4)
Y <- rbinom(n_obs, 1, plogis(A + W1 + W2 - W3 - W1 * W3))

# fit the IPW estimator
est_ipw <- ipw_shift(
  W = cbind(W1, W2, W3), A = A, Y = Y,
  delta = 0.5, cv_folds = 2L,
  n_bins = 5L, bin_type = "equal_range",
  lambda_seq = exp(seq(-1, -10, length = 100L)),
  # arguments passed to hal9001::fit_hal()
  max_degree = 3,
  smoothness_orders = 0,
  num_knots = NULL,
  reduce_basis = 1 / sqrt(n_obs)
)
#> 2% of observations outside training support...predictions trimmed.
confint(est_ipw)
#> # A tibble: 2 × 8
#>   lwr_ci   psi upr_ci se_est type     l1_norm lambda_idx gn_nbins
#>    <dbl> <dbl>  <dbl>  <dbl> <chr>      <dbl>      <dbl>    <int>
#> 1  0.576 0.72   0.830 0.0657 dcar_tol    4.53         35        5
#> 2  0.578 0.719  0.828 0.0647 dcar_min   20.0          64        5