13. In Example 11.11 (scram rates) consider a model with varying over time, and taking {logit(t ), bt} to followa bivariate normal randomwalk. Omit

13. In Example 11.11 (scram rates) consider a model with ω varying over time, and taking

{logit(ωt ), bt} to followa bivariate normal randomwalk. Omit the 10th year’s observations

(namely replace yi,10 by NA though keeping the offsets Hi,10 as they are). The actual data for the last year will then be a separate vector. Compare the predictions (e.g. posterior mean of absolute deviations between predictions and actual divided by 66) of the constant

ω model (and RW1 prior in bt only) with the extended model.