Additional Exercise 1 for Session 9.

Based on the papers Barbosa & Goldstein (2000) and Yang, Heath & Goldstein (2000),
and Section 5 of the manual for MLwiN Macros for Advanced Multilevel Modelling (2001), 
explore the facilities in MLwiN for repeated measures, random effects models for binary
data.

i. Go through the example given to reproduce all values obtained. Run also a
purely hierarchical model, and explore if any underdispersion is present in 
that analysis.

ii. Pick on herd from the scc_40 dataset in VER, and carry out a similar
analysis for all cows from that herd.

iii. Explore whether the approach seems to extend to a 3-level model. If so,
carry out an analysis of the full scc_40 dataset.