Open DominiqueMakowski opened 4 years ago
Not sure, but isn't the approach simply the same as here?
https://github.com/easystats/parameters/blob/master/R/bootstrap_model.R
Or simply generate bootstrap replicates of the data (see https://github.com/strengejacke/sjstats/blob/master/R/bootstrap.R), then compute correlation function on the data and take quantiles from the results.
which functions actually might need bootstrapped CIs?
cor_test_freq and the ones relying on cor_to_ci
, e.g., cor_test_biweight
See alse #181
since we have several different types of correlations for which p values are not straightforward, would be nice to include a method for bootstrapped CI. Any advice?