Model 11.5

[ASKurz]

I don’t believe brms allows for the beta-binomial likelihood, yet. So we don’t have a way to fit McElreath’s:

m11.5 <- map2stan(
    alist(
        admit ~ dbetabinom(applications,pbar,theta),
        logit(pbar) <- a,
        a ~ dnorm(0,2),
        theta ~ dexp(1)
    ),
    data=d,
    constraints=list(theta="lower=0"),
    start=list(theta=3),
    iter=4000 , warmup=1000 , chains=2 , cores=2 )

Correspondingly, we don’t have a good version of Figure 11.5.

[paul-buerkner]

See https://github.com/paul-buerkner/brms/blob/master/vignettes/brms_customfamilies.Rmd

[ASKurz]

Paul, this is tremendous and the vignette is coming along nicely, too!