4. MSE estimation
Jae-kwang Kim, Seunghwan Park and Seo-young Kim
Previous | Next
We
now discuss mean squared error (MSE) estimation of the GLS estimator
which is given
by (2.9). Note that the GLS estimator is a function of
and
If the model
parameters are known, then the MSE of
is equal to
as discussed in
Remark 1. That is, writing
and
the actual
prediction for
is computed by
To account for
the effect of estimating the model parameters, we first note the following
decomposition of
which was originally proved by Kackar and Harville (1984) under normality
assumptions. The first term,
is of order
where
is the size of
and the second
term,
is of order
with
The second term
is often much smaller than the first term.
We
consider a jackknife approach to estimate the MSE. Use of the jackknife for
bias-corrected estimation was originally proposed by Quenouille (1956). Jiang,
Lahiri and Wan (2002) provided a rigorous justification of the jackknife method
for the MSE estimation in small area estimation. The following steps can be
used for the jackknife computation.
- Step 1 Calculate the
replicate
of
by deleting the
area data set
from the full
data set
This calculation
is done for each
to get
replicates of
which, in turn,
provide
replicates of
where
- Step 2 Calculate the estimator of
as
- Step 3 Calculate the estimator of
as
- where
is a
bias-corrected estimator of
given by
- and
Remark 4 For the transformation in (3.13),
we use the bias-corrected estimator in (3.14) and its MSE estimation method
needs to be changed. Using
to denote the bias-corrected estimator in (3.14)
evaluated at
we can have the
where the first equality
follows that
is of order
The MSE of
the EGLS estimator of
after transformation, is computed by (4.1) and
(4.2). Once
is estimated, we should multiply it by
to obtain the MSE estimator of the
back-transformed EGLS estimator
Previous | Next