JZS Prior Specification
Hello,
If I wanted to specify the same default priors used in the Bayesian Repeated Measures ANOVA module in JASP, what would it be?
I am trying to compare the analysis in JASP with brms, and this website says that it is Cauchy(0, 0.707) prior on the Intercept (see screenshot below).
I have two questions,
1) Is the Cauchy(0,0.707) the prior on population-level effects too (i.e., class = "b")?
2) How would I specify the Jeffreys prior on Sigma to completely mimic how the prior is specified in JASP?
Thank you
Comments
Hi Rohanp16,
Not sure about the workings of brms. We once coded this in WinBUGS (http://www.ejwagenmakers.com/2009/WetzelsEtAl2009Ttest.pdf). The Lee & Wagenmakers (2013) textbook (https://bayesmodels.com/) also provides JAGS/Stan code. Ioannis Ntzoufras has written a few tutorial papers on coding the JZS prior using BUGS.
Cheers,
E.J.
Thanks for that EJ, will give them a read.
Cheers