Categorical model with continuous dependent variable

[junpenglao]

@gBokiau is right you should set the shape=7 on p1.

However, since you said earlier:

@pdeboer

And I don’t want a separate Beta for every sample.

You should model p differently and had it multiple with b later on.