Figure 13

Maps of mean square error of prediction for the cervix cancer risk maps of Figure 9. The fill color in each county represents the mean square error of prediction of the risk of cervix cancer produced by the following predictors: population-weighted average, local and global empirical Bayes smoothers, and Poisson kriging. The units are age-adjusted mortality rate per 100,000 person-years, and the class boundaries correspond to the deciles of the histogram of mean square error of prediction.