Separate Bayesian analyses?

Comments

• EJ

  April 2016 edited 10:57AM

  Hi JR,

  This is largely a matter of taste I think. The first idea we had was to incorporate only the Bayesian analyses, but then we figured people might like to have the classical analyses as well. Anyway, one problem with combining the two is that it might become a little messy in terms of the user interface. In addition, the current way we are doing it now makes it very apparent that for most classical analyses there exists a Bayesian alternative. A different team could have made a different decision. So far, what we've done seems to work well for most users.

  Cheers,
  E.J.
• coledavis

  May 2016 edited 10:57AM

  Looking from another point of view, there is a trend of people moving to JASP from SPSS, some of whom will want classical analyses only (although hopefully they'll start to see the benefit of Bayesian analyses as they get accustomed to JASP). I do see the combination point of view in one area: If you want to input variables into a factorial repeated measures ANOVA, you've got to replicate the laborious process to get both classical and Bayesian results. From a results point of view, however, I think it makes sense to have separate displays, because there are just so many things you might want to set up, with different options for the different approaches; for a beginner it could make the whole thing rather intimidating.