Bayes Factor of regression model with specific prior using JAGS module

narcilili

hello jaspers,

i used JAGS module in estimating the posterior slope of a simple linear regression model with a specific prior from the literature. 

i wonder if there is a way to determine the Bayes Factor of the slope? thanks.

cheers

EJ

Hi Narcilili,

The easiest way is the Savage-Dickey density ratio (see papers on http://www.ejwagenmakers.com/papers.html).

Cheers,

E.J.

narcilili

hello E.J.,

thanks for the reply. i will read your paper and try to work it out. thanks a lot.

cheers

narcilili

hello jaspers,

i used JAGS module in JASP to estimate the posterior slope of a simple linear regression model with a specific prior from the literature of 3.04. the analysis showed a posterior slope of 2.99. I wonder if someone could share R code to determine if the Bayes Factor for the null hypothesis that the posterior slope = zero? thanks a lot.

cheers

EJ

You could check out some of the R code that goes with the 2013 book by Michael Lee and myself: https://bayesmodels.com/ (the later chapters are about Savage-Dickey Bayes factors)

narcilili

hello E.J.,

thanks a lot for your help. i will check it out.

cheers