The papers below claims that BCa bootstrap improves bootstrap estimate accuracy over standard quantile bootstrapping, I think by adjusting for the underlying distribution's skew and bias. I am looking to get an intuitive understanding of how this works.

What is the issue with using standard quantile-based bootstrapping for a very skewed distribution, and how does this method result in improvement? It seems to be somehow adjusting for bias in quantile-based bootstrapping, but I'm not completely sure how this works?

I've tried reading these papers, as well as some previous crossvalidated posts on BCa bootstrap but I'm still confused by the topic.


DiCiccio, Thomas J.; Efron, Bradley. Bootstrap confidence intervals. Statist. Sci. 11 (1996), no. 3, 189--228. doi:10.1214/ss/1032280214.


Kelley, Ken. "The effects of nonnormal distributions on confidence intervals around the standardized mean difference: Bootstrap and parametric confidence intervals." Educational and Psychological Measurement 65.1 (2005): 51-69.


crossvalidated questions

Why are BCa bootstrap confidence intervals second order exact?

Comparison between bootstrap percentiles confidence interval and BCa confidence interval?

  • 1
    $\begingroup$ Please include full citations for the papers, as links may break in the future; we want posts to remain valid even if this happens $\endgroup$ – user20160 Nov 4 '19 at 17:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.