@pauljohn32, this is out of my depth and beyond the scope of the post, but since it was brought up, don't the bias-corrected CI's account for lack of normality in the sampling distribution? I thought that was part of the reason it was preferred to the simple percentile method. I only have a conceptual understanding of most of these concepts so, when possible, conceptual explanations are most useful :).

Thanks for the feedback @pauljohn32. I'm not familiar with HPD, is that Highest Posterior Density in a Bayesian context? Just to be clear, I'm not actually working with r-squared but used it as a reproducible example...guess it wasn't the best choice of statistic.

Thank you @kbiolsi! I was close :). Also, not sure how I missed that website you added which shows examples of all kinds in R. I swear I searched high and low but kept getting directed to boot package examples. Thanks again for the prompt reply.