A B for every p

Discuss statistics related things

by RHainez » Thu Jun 20, 2019 12:27 pm


A colleague gave me this article https://royalsocietypublishing.org/doi/ ... .2019.0174 about the limitations of the p-value. It's a review of 4 methods that anyone can use to overcome those limitations and get a better understanding of the data.

The authors talk, among other things, about a simplified Baye Factor called the Bayes factor upper bound, a way of getting a way more simple interpretation of the p-value.
Jamovi has a module about Bayesian stats, Jasp-style, but I was wondering if it could be interesting to give the user the possibility to tick a case allowing the computing of this Bayes factor upper bound. Like a first taste of the Bayesian world available for the p-value maniacs all over the world (and for humans too). Or compute it by default, to get à B for every p, like the author of the article said.

What do the stats wizards think about this approach?

Have a nice day.
Posts: 81
Joined: Wed Feb 08, 2017 3:14 pm

by Vit » Wed May 06, 2020 12:15 pm

I would love that!
Posts: 27
Joined: Fri Apr 19, 2019 11:51 am

by jonathon » Wed May 06, 2020 10:19 pm

yeah, we could provide it in a module like the `moretests` module, where additional stats get added to existing analyses.

we just need to find someone with the time to do it (the perennial problem!) :)

User avatar
Posts: 1382
Joined: Fri Jan 27, 2017 10:04 am

Return to Statistics