Although OLS can be used to estimate and
test the
unstandardized effects, it is inappropriate
for forming the confidence intervals for the
standardized effects. See
@yuan_biases_2011 on the issue of standardized
regression coefficients.

To form nonparametric bootstrap confidence intervals for
effects to be computed, add `boot_ci = TRUE`,
`R` to the number of bootstrap samples
(should be 5000 or even 10000, for
multiple regression), and `seed` (set
it to an integer to ensure the results are
reproducible).
