caliplot: Calibrate sensitivity parameters based on pre-treatment...

In JiajingZ/TukeySens: Sensitivity Analysis Based on Tukey's Factorization

Description

Calibrate the magnitude of sensitivity parameters to the amount of variation in the treatment assignment T that is explained by outcome Y(t), above and beyond what is accounted for by observed pre-treatment variables X.

Usage

caliplot(x, trt, y, gamma_seq)

Arguments

x	a tibble or data frame with observed pre-treatment variables
trt	a vector with binary treatment indicators
y	a vector with outcomes
gamma_seq	a vector with chosen values for sensitivity parameter γ_t

Details

caliplot returns a plot of sensitivity parameter γ_t vs partial coefficient of variation from outcome Y, ρ^2_{Y|X}. For comparison, the largest four partial coefficients of variation from covariates are also plotted if the number of observed covariates is larger or equal to four; otherwise, all the partial coefficients of variation from covariates are plotted.

Examples

# Observed Pre-treatment Variables 
x = NHANES %>% select(-one_of("trt_dbp", "ave_dbp"))

# Treatment 
trt = NHANES %>% select(trt_dbp)

# Outcomes 
y = NHANES %>% select(ave_dbp)

# Sensitivity Parameter Sequence 
gamma = seq(0.01, 0.1, by = 0.001)

# plot 
caliplot(x, trt, y, gamma)

JiajingZ/TukeySens documentation built on Jan. 23, 2020, 3:44 a.m.