Small Area Estimation - Sampling Spatial Units for Agricultural Surveys - page 288

Fig. 11.2 Comparison of
the variance of the direct
estimator with the MSE
estimate of the EBLUP
estimator

> data(cornsoybean)
> data(cornsoybeanmeans)
> attach(cornsoybean)
> attach(cornsoybeanmeans)
> Xmean < - data.frame(CountyIndex, MeanCornPixPerSeg, MeanSoyBeansPix
+ PerSeg)
> Popn < - data.frame(CountyIndex, PopnSegments)

The EBLUP estimates using the unit level model for corn and soybeans for the
12 SAs can be computed using the following code.

> saeCorn < - eblupBHF(CornHec ~ CornPix - 1, dom¼County,
+ meanxpop¼Xmean[,1:2], popnsize¼Popn, data¼cornsoybean)

> saeCorn

$eblup
domain eblup sampsize

1 1 122.1629 1

2 2 123.7423 1

3 3 112.3628 1

4 4 114.3091 2

5 5 137.3916 3

6 6 109.1529 3

7 7 116.0643 3

8 8 122.6724 3

9 9 111.3541 4