Retrieve the assumptions for the variance estimators to be consistent for a a fitted OLS maars_lm, lm class object

Retrieve the assumptions for the variance estimators to be consistent for a a fitted OLS maars_lm, lm class object.

get_assumptions(
  mod_fit,
  sand = NULL,
  boot_emp = NULL,
  boot_sub = NULL,
  boot_mul = NULL,
  boot_res = NULL,
  well_specified = NULL
)

Arguments

mod_fit	(maars_lm, lm) A fitted OLS maars_lm, lm class object
sand	(logical) : TRUE if sandwich estimator output is required, FALSE to exclude this output from the request
boot_emp	(logical) : TRUE if empirical bootstrap standard error output is required, FALSE to exclude this output from the request
boot_sub	(logical) : TRUE if subsampling standard error output is required, FALSE to exclude this output from the request
boot_mul	(logical) : TRUE if multiplier bootstrap standard error output is required, FALSE to exclude this output from the request
boot_res	(logical) : TRUE if residual bootstrap standard error output is required, FALSE to exclude this output from the request
well_specified	(logical) : TRUE if lm standard errors (well specified) output is required, FALSE to exclude this output from the request

Value

(vector) : Vectors containing the assumptions under which each estimator of the variance is consistent.

Examples

if (FALSE) {
set.seed(1243434)

# generate data
n <- 1e3
X_1 <- stats::rnorm(n, 0, 1)
X_2 <- stats::rnorm(n, 10, 20)
eps <- stats::rnorm(n, 0, 1)

# OLS data and model
y <- 2 + X_1 * 1 + X_2 * 5 + eps
lm_fit <- stats::lm(y ~ X_1 + X_2)

# DEFINE common column names - these stay the same across all
# reported error types
common_vars <- c("term", "estimate")

# Empirical Bootstrap check
set.seed(454354534)
comp_var1 <- comp_var(
  mod_fit = lm_fit,
  boot_emp = list(B = 20, m = 200),
  boot_res = list(B = 30)
)

# This returns everything but boot_mul, since we didn't run it in the original
# original maars_lm model
get_assumptions(
  mod_fit = comp_var1)
}