ABSTRACT
This chapter changes the discussion from the basic workings of Bayes’ Law in a probabil-
ity context to a focus on the use of Bayes’ Law for realistic statistical models. Consequently,
the first order of business is to go from our previous vague definition of data, D, to a rect-
angular n×k matrix of data, X. In this chapter we also make the move from the unspecific p() for posterior distributions to the more clear π() notation in order to distinguish them
from priors, likelihoods, and other functions. Also, from now on we use the vector form of
theta, θ, since nearly all interesting social science models are multidimensional.