Effect sizes for Bayesian Mixed ANOVA and Mann Whitney U test

Comments

• MSB

  July 2021

  You might be interested in the effectsize package for both these issues:

  https://easystats.github.io/effectsize/articles/bayesian_models.html#eta2-1

  https://easystats.github.io/effectsize/articles/simple_htests.html#difference-in-ranks

  
  
• max_gold

  July 2021

  Thanks for the reference to the package!

  Does anyone have any experiences with using JASP for the analysis but then stan for the effect size estimation? Is there a way to get the "underlying" R code that is used for an analysis that was carried out in JASP?
• EJ

  July 2021

  1. You can use Stan, but note that we also have a JAGS module inside JASP.
  2. JASP should offer effect size estimates in addition to the testing results. If there is anything you think is missing please let us know on our GitHub page.
  3. The underlying R code will be made easily available very soon. This has been our top priority for a while but after the upcoming 0.15 version this will be our main focus.

  Cheers,

  E.J.
• max_gold

  July 2021

  Hey E.J.,

  Thanks, sounds good! But concerning 2: Where can I find the effect sizes for a repeated-measures ANOVA in JASP? Are they computed for each predictor and interaction or for each model?

  I'm not sure if it is missing because I'm not entirely sure what to look for.

  Cheers,

  Max
• vandenman

  July 2021

  @max_gold , The only available effect size measure for the Bayesian ANOVAs is R^2. This is computed for each model and you can obtain a model-averaged R^2. The frequentist ANOVAs provide η^2 and several other effect size measures for which we do not (yet) have Bayesian counterparts. That certainly warrants a feature request.
• max_gold

  July 2021

  Okay, thank you for the info!