Matt Haberland

Please see scipy/scipy#14563 for context.

I'm not sure, but @tacaswell may have something in mind. I posted at his request.

I'd be happy to add other enhancements to `scipy.stats.bootstrap` if it would make it more attractive! As @FlorinAndrei mentioned, domain knowledge review of PRs would help - the main reason...

> I'd be happy to review a PR and/or tutorial Great! If you'd like to look at doc additions in scipy/scipy#16454, which returns the bootstrap distribution, I'd appreciate it! If...

@raphaelvallat Now that scipy/scipy#16651 addresses your idea 2 above, here are some thoughts about your idea 1. In retrospect, I would have made `confidence_interval` a method so that users could...

> Will it also require passing data and statistic as well as bootstrap_distribution Most CI methods need to know the observed value of the statistic to compute the confidence interval....

Just thought I'd mention that https://github.com/scipy/scipy/pull/16455 would add BCa bootstrap for multi-sample statistics. I think it's correct now, and I'm working on unit tests for it. Reviews appreciated!

And scipy/scipy#16714 adds the functionality suggested in https://github.com/raphaelvallat/pingouin/issues/189#issuecomment-1189750880 to address @raphaelvallat's idea 1. With that, one can change confidence level, CI type, and add resamples (or not) like: ```python3 res...

No problem. Thanks for taking a look at scipy/scipy#16714!

The multi-sample BCa PR merged. Hope it helps!