gen_a_start {OVtool}

gen_a_start

Description

This function is a wrapper to ov_simgrid. It generates the a. a is used to control the strength of the relationshp between the unobserved counfounder, U, and the treatment indicator

Usage

gen_a_start(y, tx, residuals, es, rho, my_estimand)

Arguments

y

A vector that represents the outcome.

tx

A vector for the treatment indicator (must be 0s and 1s).

residuals

A vector of residuals from regressing Y on X and controlling for treatment.

es

An effect size value to simulate over.

rho

A rho (correlation) value to simulate over.

my_estimand

"ATE" or "ATT"

Value

gen_a_start returns a list containing the following components:

n1

scalar representing sample size of treatment group (treat == 1)

ve1

1 - b1^2 multiplied by the variance of Ystar1

b1

bounded parameter for treatment group (treat == 1) that it with b0 are selected to set the correlation of the omitted variable and the outcome equal to rho

es

pi

proportion of population that is in the treatment group (treat == 1)

n0

scalar represnting sample size of control group (treat == 0)

ve0

1 - b0^2 multiplied by the variance of Ystar0

b0

bounded parameter for control group (treat == 0) that it with b1 are selected to set the correlation of the omitted variable and the outcome equal to rho

n

scalar representing the total sample size

ind

vector of positions in data that represent treatment group (treat == 1)

Rstar_1

Residuals in treatment group

Rstar_0

Residuals in control group

Examples

data(sud)
sud = data.frame(sud)
sud$treat = ifelse(sud$treat == "A", 1, 0)
sud$wts = sample(seq(1, 10, by=.01), size=nrow(sud), replace = TRUE)
outcome_mod = outcome_model(data = sud,
                            weights = "wts",
                            treatment = "treat",
                            outcome = "eps7p_3",
                            model_covariates = c("sfs8p_0"),
                            estimand = "ATE")

start = gen_a_start(y=sud$eps7p_3, tx=sud$treat,
                    residuals=residuals(outcome_mod$mod_results),
                    es = .01,
                    rho = .01,
                    my_estimand = "ATE")