http://www.andrewheiss.com/blog/2019/01/29/diff-means-half-dozen-ways/

[andrewheiss]

Half a dozen frequentist and Bayesian ways to measure the difference in means in two groups | Andrew Heiss
http://www.andrewheiss.com/blog/2019/01/29/diff-means-half-dozen-ways/

[andrewheiss]

Originally posted by Rasmus Nikolai Nes on 2019-02-13 08:46:05

Great review and comparison, thank you very much for the great post! A
bayesian interpretation of the bootstrapping "posterior" distribution,
like e.g. Rasmus Bååth writes about
[here](http://www.sumsar.net/blog/2015/04/the-non-parametric-bootstrap-as-a-bayesian-model/ "http://www.sumsar.net/blog/2015/04/the-non-parametric-bootstrap-as-a-bayesian-model/"),
is a method that I am enthusiastic about these days. It seems we can use
the simple approach of making bootstrap distributions from the samples
-- like what you've done above -- and at the same time you can get
the convenient bayesian probabilistic interpretation of the results.

What do you think about this approach, and how would you compare it to
the more "traditional" bayesian hypothesis test methods that you've
introduced here?

[andrewheiss]

Originally posted by blogcollector on 2021-01-29 12:07:39

Great!!!
But in your stan model in the transformed parameters block shouldn't
the transformed parameters be mean_y = y - mean(y)?

[andrewheiss]

Originally posted by Rainer Düsing on 2021-07-08 16:30:11

Thanks for the great overview, but may it be that your calculation of
cohens_d is not correct? Shouldn't it be the variances in the
denominator instead of the standard deviations? Same for the CL ES?