# Advanced options in Bayesian ANOVA

I have just downloaded the new version of JASP (0.8.0.0.) and wish to gain a better understanding of how the new functions work. In particular, for Bayesian repeated-measures ANOVA, there is now the option to modify the r values for covariates, fixed effects, and random effects. Is there any reason to change the assigned default values? For example, if I expect an effect size of 0.3 (partial eta square) based on prior research, how do I change the r values to reflect this prior belief?

Additionally, how does the 'Samples' function work (with the options 'Auto' and 'Manual')? What is the 'auto' sample value and is there any reason to manually change this?

Any clarification on this matter would be greatly appreciated - thanks!

## Comments

When you set the "samples" function to manual and you increase the default number (which is what "auto" does -- I guess we could have called it "default") then you can decrease the %error in the estimation of your BFs. The price you pay is that it takes more time to get the answer.

My advice is not to tinker with the default r values unless you collaborate with a Bayes expert. Also, note that the current priors are centered on the null. We are working to relax that restriction.

Cheers,

E.J.

Thanks EJ for the clarifications! I have just one more question regarding the r values - would it be necessary to report these default r values in a research paper?

Either that, or refer to the default and cite the BayesFactor package documentation for details.

E.J.