Some optional exercises to go with Bock's Chapter 4 Section 1:

1) Bock's Exercise 4.1-1, part (a), on page 207 is a good one.

2) The presentation of the maximum likelihood development of least squares 
regression on the in-class "slides" began with a multivariate normal likelihood 
for the data considered as a single observation. An alternative likelihood-based 
development considers either (a) each group mean, or (b) each individual 
observation, as a draw from a univariate Gaussian distribution. In the latter 
case, the likelihood for the data (as a whole) is a product of those univariate 
Gaussian terms. This exercise, should you choose to accept it, is to re-develop 
the maximum likelihoo estimator for regression coefficients starting with the 
likelihood as a product of univariate Gaussian likelihoods.