Generate maars_lm, lm object with estimates of the variance

Generates an object of class "maars_lm", "lm" containing estimates of the variance of the coefficients in the regression model.

comp_var(
  mod_fit,
  boot_emp = NULL,
  boot_sub = NULL,
  boot_res = NULL,
  boot_mul = NULL
)

Arguments

mod_fit	An lm (OLS) object
boot_emp	(list) In the case of empirical bootstrap the expected input is of the form #' `list(B = 10, m = 100)`. Here the named element `m` is optional e.g. `list(B = 10)` is valid, or passed in as an explicit `NULL` e.g. `list(B = 10, m = NULL)`. Note that technically `B, m` should both be positive integers, but this assertion checking is handled explicitly in the `comp_boot_emp` function. So although passing in `list(B = -15, m = -20)` will pass this function without errors, these will be addressed explicitly in `comp_boot_emp` as invalid inputs.
boot_sub	(list) TODO: ADD
boot_res	(list) : In the case of residual bootstrap the expected input is of the form `list(B = 10)`. Note that technically `B` should be a positive integer, but this assertion checking is handled explicitly in the `comp_boot_res` function. So although passing in `list(B = -15)` will pass this function without errors, these will be addressed explicitly in `comp_boot_res` as invalid inputs.
boot_mul	(list) : In the case of multiplier bootstrap the expected input is of the form `list(B = 10, weights_type = "rademacher")`. Here the named element `weights_type` is optional e.g. `list(B = 10)` is valid, or passed in as an explicit `NULL` e.g. `list(B = 10, weights_type = NULL)`. Note that technically `B` should be a positive integer, and `weights_type` should be a character vector (see `comp_boot_mul`), but this assertion checking is handled explicitly in the `comp_boot_mul` function. So although passing in `list(B = -15, m = "test")` will pass this function without errors, these will be addressed explicitly in `comp_boot_mul` as invalid inputs.

Value

A "maars_lm" object containing the estimates of the variance of the regression coefficients, including the sandwich and the variance returned by stats::lm.

Details

The "maars_lm" object is basically an "lm" object with additional attributes (the additional estimates of the coefficients variance), which are stored within "var". For example, the estimates of the empirical bootstrap will be stored within "var$var_boot_emp". Each of the nested lists contains the following elements: the type of estimator of of the variance (var_type); An abbreviated string representing the type of the estimator of the variance (var_type_abb); the summary statistics of mod_fit based on this estimator of the variance (e.g., standard errors and p-values) (var_summary); the assumptions under which the estimator of the variance is consistent (var_assumptions); the covariance matrix for the coefficients estimates (cov_mat).

Examples

if (FALSE) {
# Simulate data from a linear model
set.seed(35542)
n <- 1e2
X <- stats::rnorm(n, 0, 1)
y <- 2 + X * 1 + stats::rnorm(n, 0, 1)

# Fit the linear model using OLS (ordinary least squares)
mod_fit <- stats::lm(y ~ X)

# Run the multiplier bootstrap on the fitted (OLS) linear model
set.seed(162632)
out <- comp_var(mod_fit, boot_mul = list(B = 100, weights_type = "rademacher"))

# print output
print(out)
print(out$var$var_boot_mul)
}