bcalibrate {CopSens} | R Documentation |
Calibration for Binary Outcomes
Description
Calibrates the naive estimates to account for unobserved confounding when outcome variables are binary. The calibration can be done with user-specific sensitivity parameter or with our pre-provided calibration methods, the worst-case calibration for a single contrast or multivariate calibration for multiple contrasts.
Usage
bcalibrate(
y,
tr,
t,
gamma,
R2 = NULL,
mu_y_t = NULL,
mu_u_tr = NULL,
mu_u_t = NULL,
cov_u_t = NULL,
nU = NULL,
nsim = 4000,
...
)
Arguments
y |
|
tr |
|
t |
|
gamma |
a vector specifying the direction of sensitivity parameters. |
R2 |
an optional scalar or vector specifying the proportion of residual variance in outcome given the treatment that can be explained by confounders, which determines the magnitude of sensitivity parameters. |
mu_y_t |
an optional scalar or vector that contains naive estimates of treatment effects ignoring confounding. |
mu_u_tr |
an optional matrix of conditional confounder means for all observed treatments with latent variables in columns. |
mu_u_t |
an optional matrix of conditional confounder means for treatments of interest with latent variables in columns. |
cov_u_t |
an optional covariance matrix of confounders conditional on treatments. |
nU |
Number of latent confounders to consider. |
nsim |
an optional scalar specifying the number of sample draws. |
... |
Value
A data.frame
with naive and calibrated estimates of population average outcome receiving
treatment t
.
Examples
# load the example data #
y <- GaussianT_BinaryY$y
tr <- subset(GaussianT_BinaryY, select = -c(y))
t1 <- tr[1:5,]
t2 <- rep(0, times = ncol(tr))
# calibration #
est_b <- bcalibrate(y = y, tr = tr, t = rbind(t1, t2),
nU = 3, gamma = c(1.27, -0.28, 0),
R2 = c(0.2, 0.7))
est_b_rr <- list(est_df = est_b$est_df[1:5,] / as.numeric(est_b$est_df[6,]),
R2 = c(0.2, 0.7))
plot_estimates(est_b_rr)