Lecture #10 Order Statistics

Objectives of this Lecture:

After reading this unit, you should be able to
• understand the basic concept of order statistics (OS)
• understand about the application of OS

Distribution of Range
Let X 1 , X 2 , . . . , X n be a random sample of size n from a continuous population with PDF f(x)
and CDF F(x). Let the sample values are arranged in order of magnitude as
X 1 : n≤ X 2 : n≤. . .≤ X n : n where X 1 : n and X n: n are the smallest and largest value in the ordered
sample, respectively.
The sample range is defined as the difference between the highest and lowest order statistics. So the
range is
W = X n :n − X 1 : n
Let X 1 : n= x and X n : n= y
The joint density of X 1 : n and X n: n is

f 1 , n: n ( xy)=n( n−1)[ F ( X n : n= y )−F ( x )]n−2 f (x ) f ( y ); −∞≤ x≤ y≤∞

After Jacobian transformation, the joint distribution of U and W is
f (uw)= f ( xy)| j|
= n(n−1)[ F ( w+ u)−F (u)] n−2 f (u) f (w+u) ; −∞ < u< ∞ , 0< w< ∞
The distribution of range is
∞
n−2
f W (w)=n( n−1) ∫ [F (w +u)− F (u)] f (u) f ( w+ u)du ; 0<w < ∞
−∞

Eves through the integration to be carried out in (i) does not assume a manageable form in many
cases, the CDF of W does take on a simpler form and may be derived as
w0

F W ( w0 )=P [W ≤w0 ]=∫ f W (w0 )dw

0
w0 ∞
n −2
= n( n−1)∫ ∫ [ F ( u+w)− f (u)] f (u) f (u +w)du dw
0 −∞
∞
= n ∫ [ F (u+ w0 )−F (u)] n−1 f (u)du ; 0< w< ∞
−∞

which is the required distribution function of range.

Order Statistics | Basic Concept | Lecture #10

 Lectured by
STT453: Order Statistics Md. Kaderi Kibria, STAT, HSTU

Distribution of Median
Consider the sample size n to be odd. Then, from the distribution of rth order statistics we have PDF
of the sample median X¯ n= X (n+1)/ 2 :n to be
n +1 n+1
n! 2
−1 n−
2
f x̄ ( x)= [ F ( x)] f ( x) [1− F ( x)]
n
n+1 n+1
(
2
−1 ! n− )(2
! )
n! (n −1 )/ 2 (n−1)/ 2
= 2
[F ( x )] [1−F ( x )] f ( x)
n−1
( {2
)! }
n! (n−1)/ 2
= 2
{ F ( x )[1−F (x)] } f ( x) −∞ < x < ∞
n−1
( {2
)! }
which is the distribution of median when n is odd. From the PDF of the sample median it is
observes that it is symmetric about 0.
Suppose the sample size n is even. The the sample median is given by
1
X¯ n= X n + X n
2 2 :n 2 [
+1 : n ]
Therefore the joint density function of X n/ 2 : n and X (n / 2)+1 : n to be
n n
n! −1 −1
f n n (x 1 x 2)= [ F ( x1 )]2 [1−F (x 2 )] 2 f ( x 1) f (x 2 )
, +1 : n n n n n
2 2 ( −1)!( +1− −1)!( n− −1)!
2 2 2 2
n n
n! −1 −1
= [ F ( x 1 )] [1−F (x 2 )] 2 f ( x1 ) f ( x 2)
2
n n
( −1)!( −1)!
2 2
After Jacobian transformation, the joint density function becomes
f X n/2 :n , X̄ n (x 1 x)= f n n ( x 1 x 2 )|J |
, +1: n
2 2
n n
2 n! 2
−1
2
−1
= 2
[F ( x 1 )] [1−F (2 x− x1 )] f ( x 1) f (2 x− x 1);−∞ < x 1 < x < ∞
n
{( ) }
2
−1 !

By integrating out x 1 we drive the PDF of the sample median X¯ n as

x n n
2 n! −1 −1
f X¯ ( x )=
n 2∫
[ F ( x 1)] 2 [1−F (2 x−x 1)] 2 f ( x 1 ) f ( 2 x−x 1 ) dx1 ;−∞ < x< ∞
n
{( ) }
2
−1 !
−∞

. . .. . . . .(i)
The integration to be performed in equation (I) does not assume a manageable form in most cases.
Yet the CDF of the sample median X¯ n can written in a simpler form as

Order Statistics | Basic Concept | Lecture #10

 Lectured by
STT453: Order Statistics Md. Kaderi Kibria, STAT, HSTU

F X̄ ( x 0)=P [ X¯ 0≤x 0 ]
n
x0 x n n
2 n! −1 −1
= 2 ∫∫
[ F (x 1)] 2 [1−F (2 x−x 1)] 2 f (x 1 ) f ( 2 x−x 1 )dx1 dx
n
{( ) }
−∞ − ∞
−1 !
2
By employing Fulbini’s theorem and changing the order of integration, we derive the CDF of X¯ n
as
x0 n x n
F X̄ ( x 0)=
n
n
2 n!
2 ∫ [F ( x 1 )] 2
−1
f ( x1 ) [∫ [1−F (2 x− x 1)] 2
−1
]
f (2 x− x 1)dx dx 1

{( ) }
−∞ −∞
−1 !
2
x0 n n x0 n n
=
n
2 n!
n
[∫ [ F ( x1 )] 2
−1
2
[1−F (x 1)] f (x 1 ) dx 1−∫ [ F ( x 1)] 2
−1
2
[1−F (2 x0 −x 1)] f ( x1 )dx 1 ]
( )( )
2
−1 !
2
!
−∞ −∞

1
Example. The CDF of the midpoint or mid-range M= [ X − X n: n ] in random samples of
2 1:n
size n from a continuous population with CDF F(x) is
x
n−1
F (x)=n ∫ [ F ( 2 xm)−F ( m) ] f (x )dx
inf

Solution: The joint density function of X 1 : n and X n: n is

n−2
f 1 , n: n ( x 1 x n )=n(n−1) [ F (x n )−F ( x 1 ) ] f ( x 1) f (x n ) ;−∞ < x 1< x n < ∞
Let us make a transformation
X 1+ X n
x= and m= X 1
2
After the Jacobian transformation the joint density function becomes
f (x m)= f ( x 1 x n)|J|
n− 2
= 2 n(n−1) [ F ( x n)−F ( x 1) ] f ( m) f ( 2 x−m) ; −∞ ≤m< x≤∞
The marginal density function of x is
x
n−2
f (x )=2 n(n−1) ∫ [ F (2 x−m)− F (m) ] f (m) f ( 2 x−m) dm ;
−∞

Therefore, the CDF of x is

x x
n−2
F (x)=2 n (n−1) ∫ ∫ [ F ( 2 x−m)−F (m) ] f (m) f (2 x−m)dm dx
−∞ −∞
x
n−1
= n ∫ [ F (2 x−m)−F (m) ] f ( m) dm
−∞

which is the CDF of sample mid-range.

Order Statistics | Basic Concept | Lecture #10

 Lectured by
STT453: Order Statistics Md. Kaderi Kibria, STAT, HSTU

Example. Suppose that a sample of size n (n is odd) from a standard uniform distribution. Find the
distribution of sample median.
Solution: Given that X ~U (0,1) then the density function becomes
f (x )=1 ; 0< x <1
and the distribution function is
F (x)=x
The distribution of sample median when n is odd. Therefore,
n+1 n+1
n! 2
−1 n−
2
f (n +1)/ 2 : n ( x )= [ F ( x)] [1−F ( x )] f (x)
n+1 n+1
( 2 )(
−1 ! n−
2
! )
n+1 n+1
n! 2
−1 n−
2
= [ x] [1−x ] .1
n+1 n+1
( 2 )(
−1 ! n−
2
! )
n −1 n−1
n! 2 2
= 2
x (1−x)
n−1
{( ) }
2
!
n−1
n! 2
= 2[
x (1−x) ]
n−1
{( ) }
2
!

This is the distribution of sample median of size n when it is drawn from standard uniform
distribution.
<><><><><><><><> End <><><><><><><><>

References
1. Balakrishnan, N. and A. C. Cohen (2004), Order Statistics and Inference: Estimation
Methods, Academy Press Inc., London.

2. Arnold, B.C., Balakrishnan, N. and Nagaraja, H.N. (2008), First Course in Order
Statistics, John Wiley and Sons, New York.

3. Anderson,T.W. (2003), An Introduction to Multivariate Statistical Analysis, 5th Ed.,

Wiley, N.Y.

4. Johnson, R. A. and Wichern, D. W (2002), Applied Multivariate Statistical Analysis, 5th

Ed., Prentice Hall, N.Y.

5. Izenman, A. J. (2008), Modern Multivariate Statistical Techniques, Springer, USA.

6. David, H.A. (2003), Order Statistics, 3rd Edition. Wiley Series, New Jersey, Canada.

Order Statistics | Basic Concept | Lecture #10

 Lectured by
STT453: Order Statistics Md. Kaderi Kibria, STAT, HSTU

Order Statistics | Basic Concept | Lecture #10