Graduate Lectures and Problems in Quality

Control and Engineering Statistics:

Theory and Methods

To Accompany
Statistical Quality Assurance Methods for Engineers
by
Vardeman and Jobe

Stephen B. Vardeman
V2.0: January 2001

c Stephen Vardeman 2001. Permission to copy for educational

purposes granted by the author, subject to the requirement that

this title page be axed to each copy (full or partial) produced.

A Useful Probabilistic
Approximation
Here we present the general delta method or propagation of error approximation that stands behind several variance approximations in these notes as
well as much of 5.4 of V&J. Suppose that a p 1 random vector
1
0
X1
B X2 C
C
B
X=B . C
@ .. A
Xp

has a mean vector

B
B
=B
@

EX1
EX2
..
.
EXp

C B
C B
C=B
A @

1
2
..
.
p

1
C
C
C
A

and p p variance-covariance matrix

0
VarX1
Cov (X1 ; X2 ) Cov (X1 ; Xp1 )
Cov (X1 ; Xp )
B Cov (X1 ; X2 )
VarX

Cov
(X
;
X
)
Cov
(X2 ; Xp )
2
2
p1
B
B
..
.
.
.
.
.
.
.
.
= B
.
.
.
.
.
B
@ Cov (X1 ; Xp1 ) Cov (X2 ; Xp1 )
VarXp1
Cov (Xp1 ; Xp )
Cov (X1 ; Xp )
Cov (X2 ; Xp ) Cov (Xp1 ; Xp )
VarXp
0
1
2
1
12 1 2
1;p1 1 p1
1p 1 p
2
B 12 1 2
C

2p 2 p
2
2;p1 2 p1
B
C
B
C
.
.
.
.
.
..
..
..
..
..
= B
C
B
C
2
@ 2p 2 p 2;p1 2 p1
p1
p1;p p1 p A
1p 1 p
2p 2 p
p1;p p1 p
2p

= ij i j

(Recall that if X1 and Xj are independent, ij = 0.)

127

1
C
C
C
C
C
A

128

A USEFUL PROBABILISTIC APPROXIMATION

Then for a k p matrix of constants

A = (aij )
consider the random vector
Y = A X

k1

kp p1

It is a standard piece of probability that Y has mean vector

0
1
EY1
B EY2 C
B
C
B .. C = A
@ . A
EYk

and variance-covariance matrix

Cov Y = A A0
(The k = 1 version of this for uncorrelated Xi is essentially quoted in (5.23)
and (5.24) of V&J.)
The propagation of error method says that if instead of the relationship
Y = A X, I concern myself with k functions g1 ; g2 ; :::; gk (each mapping Rp to
R) and dene
1
0
g1 (X)
B g2 (X) C
C
B
Y =B
C
..
A
@
.
gk (X)

a multivariate Taylors Theorem argument and the facts above provide an approximate mean vector and an approximate covariance matrix for Y . That is,
if the functions gi are dierentiable, let

@gi
D =
kp
@xj
1 ;2 ;:::;p

A multivariate Taylor approximation says that for each xi near i

0
1 0
1
g1 (x)
g1 ()
B g2 (x) C B g2 () C
B
C B
C
y=B
CB
C + D (x )
..
..
@
A @
A
.
.
gk (x)

gk ()

So if the variances of the Xi are small (so that with high probability Y is near
, that is that the linear approximation above is usually valid) it is plausible

129
that Y has mean vector
0
B
B
B
@

EY1
EY2
..
.
EYk

and variance-covariance matrix

C B
C B
CB
A @

g1 ()
g2 ()
..
.
gk ()

Cov Y D D0

1
C
C
C
A