310
G A R Y C . W H I T E
S
–
= ᎏᎏ
with theoretical variance (i.e., sum of the theoretical variances for each of the
estimates)
var(S
–
) = ᎏ
1
ᎏ
and empirical variance estimator
va r(S
–
) =
When the w i are the true (but unknown) weights, we have
ᎏ
1
ᎏ = ᎏ ᎏ
giving the following
1 = ᎏ
10 – 1
ᎏ
⌺
10
i=1
w i S
i
⌺
10
i=1
w i
⌺
10
i=1
w i
⌺
10
i=1
w i
( ⌺
10
i=1
w i )
(10 – 1)
( ⌺
10
i=1
w i )
(10 – 1)
⌺
10
i=1
w i ΂S i – S
–
΃
2
⌺
10
i=1
w i ΂S i – S
–
΃
2
⌺
10
i=1
w i ΂S i – S
–
΃
2