Bootstrapping confidence intervals around ICCs & variance components (in a mixed-effects model)
Date created: | Last Updated:
: DOI | ARK
Creating DOI. Please wait...
Description: Researchers may wish to compare ICCs to one another. For instance, to test whether between-target variance in appearance (i.e., target-ICC) contributes to evaluations of women more so than men. This code modifies the lmeresample() package in R (which modifies the boot() function) to bootstrap 95% confidence intervals around ICC estimates and variance components. The 95% CIs of each ICC can then be examined for overlap to determine if they are different. [Updated Nov2021]