How to conduct Bayesian Stats for non-normal distributions in JASP?
I am new to JASP and I am wondering if the current MAC version of software has an option to conduct Bayesian stats when normality and equality of variance assumptions are violated? What should I do in this case? Do I need to somehow transform my data to gaussian distribution or is there any Bayesian non-parametric analysis I can go for? Thanks so much in advance for any help on this matter!
Comments
Yes, you can always transform to normality. We also offer several rank-based Bayesian methods (depending on the analysis you want to do)
E.J.
Hi EJ! Thank you very much for your reply. I want to compare two groups (a patient group and a control group in a single variable) so I think I need to do the Bayesian equivalent to Mann-Whitney test (the non-parametric option to an independent sample T-test). How do I access in JASP to the rank-based Bayesian methods? I cannot find it in the menu options unfortunately. Also is there any tutorial I can learn from ? Thanks!
Hi Paula,
Just execute a Bayesian independent-samples t-test and then pick "Mann-Whitney" under "Tests". So it is part of the GUI for the regular t-test. You can check out this paper, for details: https://www.tandfonline.com/doi/full/10.1080/02664763.2019.1709053
Cheers,
E.J.