Statistic Random Samples

Chapter 5
The rv’s X1,…,Xn are said to form a simple
5.3 A statistic is any quantity whose value can be
calculated from sample data. Prior to obtaining
Joint Probability data, there is uncertainty as to what value of any
random sample of size n if

Distributions and Statistics particular statistic will result. A statistic is a 1. The Xi’s are independent rv’s.
random variable denoted by an uppercase letter; a
Random Samples and their lowercase letter is used to represent the calculated 2. Every Xi has the same probability
or observed value of the statistic. distribution.
Distributions

Simulation Experiments Using the Sample Mean

The following characteristics must be specified: Normal Population Distribution
5.4 Let X1,…, Xn be a random sample from a
distribution with mean value µ and standard
1. The statistic of interest. deviation σ . Then Let X1,…, Xn be a random sample from a
2. The population distribution. The Distribution ( )
1. E X = µ X = µ
normal distribution with mean value µ and
standard deviation σ . Then for any n, X
3. The sample size n. of the 2. V ( X ) = σ X2 = σ
2 is normally distributed.
n
4. The number of replications k. Sample Mean µ
In addition, with To = X1 +…+ Xn,
E (To ) = nµ , V (To ) = nσ 2 , and σ To = nσ .

The Central Limit Theorem

The Central Limit Theorem
X small to
Rule of Thumb 5.5
Let X1,…, Xn be a random sample from a moderate n X large n
distribution with mean value µ and variance σ 2 .
Then if n sufficiently large, X has Population
distribution
If n > 30, the Central Limit Theorem can
be used. The Distribution
approximately a normal distribution with
2
µ X = µ and σ X2 = σ n , and To also has of a
approximately a normal distribution with
µTo = nµ , σ To = nσ 2 . The larger the value of
µ Linear Combination
n, the better the approximation.

Linear Combination Expected Value of a Linear Variance of a Linear Combination Variance of a Linear Combination
Combination
Given a collection of n random variables Let X1,…, Xn have mean values µ1, µ2 ,..., µn If X1,…, Xn are independent, For any X1,…, Xn,
X1,…, Xn and n numerical constants a1,…,an, and variances ofσ12 , σ 22 ,..., σ n2 , respectively V ( a1 X1 + ... + an X n ) = a12V ( X1 ) + ... + an2V ( X n ) n n
the rv
n
V ( a1 X1 + ... + an X n ) = (
ai a j Cov X i , X j )
Whether or not the Xi’s are independent, = a12σ12 + ... + an2σ n2 i =1 j =1
Y = a1 X1 + ... + an X n = a iX i
i =1 E ( a1 X1 + ... + an X n ) = a1E ( X1 ) + ... + an E ( X n ) and
is called a linear combination of the Xi’s.
= a1µ1 + ... + an µn σ a1 X1 +...+ an X n = a12σ12 + ... + an2σ n2
Difference Between Two Random
Variables
E ( X1 − X 2 ) = E ( X1 ) − E ( X 2 )
and, if X1 and X2 are independent,
V ( X1 − X 2 ) = V ( X1 ) + V ( X 2 )
Difference Between Normal Random
Variables
If X1, X2,…Xn are independent, normally
distributed rv’s, then any linear combination
of the Xi’s also has a normal distribution. The
difference X1 – X2 between two independent,
normally distributed variables is itself
normally distributed.