ESTIMATION

Definition: Estimation is a procedure

by which we obtain estimate of the
true but unknown value of X , X X
1 2.. n
population parameter by using the
sample observation
EXAMPLE: We estimate the mean and
variance of a population by
computing the mean and variance of
a sample drawn from the population.
 ESTIMATE: An estimate is the numerical
value of the unknown parameter obtained
by applying a rule or a formula.
ESTIMATOR: Estimator stands for rule or
method that is used to estimate a
parameter.
EXAMPLE:If X 1 , X 2.. X n is a random
sample of size n drawn from a population
with mean  then  1
X
n
 X i

is an estimator of  and the value of


X
Is an estimate of 
1. Point Estimation
2. Interval Estimation

POINT ESTINATION:
DEFINATION: Point Estimation provides an
estimate a single value calculated from
the sample.

EXAMPLE: the value of X computes from a
sample of size is a point estimate of the
population parameter 
 EXAMPLE: a random sample of n=6
has the elements 56,10,13,14,18
and 20. compute a point estimate
(i) the population mean
(ii) the population S.D
(iii) the standard error of mean.
SOLUTION: the sample mean is
  Xi 6  10  13  14  18  20
X 
n 6
81
  13.5
6
 The point estimator of the population
mean  is 13.5 and X is an
estimator.
(ii) The S.D is
1  2
S
n
 (Xi  X )
2
 
 i   Xi
X 2
 1225  81 
2

     
n  n  6 6
 
 

 21.916  4.68
 The point estimation od the population S.D
is 4.68 and S is estimator
(iii) The Standard error


n

We use the
 sample S.D S to estimate the
S.E of X

Then
S 4.68
S x    1.91
n n
Where S xis the estimator of X
and 1.91 is the point estimate of the
standard error of mean.
 PROPERTIES OD GOOD POINT
ESTIMATERS
(i) Unbiased ness
(ii) Consistency
(iii) Efficiency
(iv) Sufficiency
 (i) UNBIASED NESS:
An estimator is said to be unbiased when
the statistic used as an estimator has its
expected value equal to the true value of
population parameter being estimated.

If  is an estimator of parameter  ,

then  will be called an inbiased
estimator if 
 E ( )  

If E ( )   then estimator  is called
a biased estimator.
 
If E ( )  0 +ve biased


E ( )  0 -ve biased

EXAMPLE: Let us consider the sample mean


X as an estimator of the population
mean 
. Then we have   

 1
X   Xi
n
Now  
E ( )  E ( X )
1
 E[ X 1  X 2  .... X n ]
n
1
 [ E ( X 1 )  E ( X 2 )  ..E ( X n )]
n
1
 .n.  
 n
X is an unbiased estimator of population
mean 
EXAMPLE: Let X 1 , X 2.. X n be a
random sample from the population
with mean 
and variance  2
we show that the sample variance
1  2
S 2
  (Xi  X )
n
is biased estimator of 
2

The sample variance is

1  2
S 2
  (Xi  X )
n
 1 
  [( X i   )  ( X   )]2

n
1  
   ( X i   )  2( X   ) ( X i   )  ( X   )
2 2

n
1 
  (Xi  )  (X  )
2 2

n
1  2
E (S )  E   ( X i   )  ( X   ) 
2 2

n 
1 
  E( X i   )  E( X  )
2 2

n
 .
1 
  var( X i )  var( X )
n
1
     X
2 2

n
 2
 i 2

n
 n 1  2
 
 n 
This shows that S 2 is biased estimators of
 2
(ii) CONSISTANCY: An estimator is said to
be consistent if the statistic to be used as
estimator becomes closer to the population
in parameter being estimated as the
sample size increases.

Symbolically  is a consistant estimator of
the parameter  if,


lim x  P 1   1  1
The sample mean
 is a unbiased
X
estimator and also consistent estimator
2 is a biased estimator but consistent
S
estimator of population variance  2 .
(iii) EFFICIENCY: An unbiased estimator
is defined to be efficient if the variance of
its sampling distribution is smaller than
that of the sampling distribution of any
other unbiased estimator of the same
parameter. In other words, suppose
there are two unbiased estimators T1
and T2 of the same parameter  , then
T1 will be said to be more efficient
estimator than T2 if var (T1)< var(T2).
(iv) SUFFICIENCY: An estimator is
defined to be sufficient, if the
statistic used as estimator uses all
the information that is contained in
the sample.
 ESTIMATION BY CONFIDENCE INTERVAL
(a) Normal population with  known.
LARGE SAMPLES:
let a random sample X1,X2..Xn of size n


is drawn from a normal population with
unknown  and known . Then

sampling distributions of mean X will be
normal with mean  and standard
deviation 
n

and the variable X 
Z
/ n
Will be exactly normal.
The normal distribution tells us that the
probability that a variable Z will fall in the
interval Z  /2
to Z
 /2
equal to 1  

where Z / 2 is equal to  /2
This we can make the following statement

 X  
P   Z / 2   Z / 2   1  
 / n 
 (i) We multiply all terms inside the
brackets by  / n and get
  
 Z / 2  X    Z / 2
n  n
(ii) We subtract X from each term
   
 X  Z / 2      X  Z / 2
n n
(iv) Multiply all terms by -1

   
X  Z / 2    X  Z / 2
n n
which is equivalent to
   
X  Z / 2    X  Z / 2
n n
 (v) we substitute the result in the
probability statement and get
   
P[ X  Z / 2    X  Z / 2 ]  1 
n n

Hence for a particular sample of size n, a

100(1   )%confidence interval for 
Is given by

   
( X  Z / 2 , X  Z / 2 )
n n
  
X   Z / 2
n
From the table Z 0.25 is 1.96 and the
confidence interval will be from X

 1.96
to   n
X  1.96
n
This means that the about 95% of the
intervals found in this way will contain the
parameter 
EXAMPLE: A random sample of 100 trees
from forest measure 30 feet on the
average height. If the S.D is 5 ft, obtain a
95% confidence interval estimate for the
mean height of trees in the forest.
 SOLUTION: We have the following
information

X =30 ft
 = 5 ft
n=100

 n  10
The 95% confidence interval may be as
   
X  1.96    X  1.96
n n
 .
 
X  1.96  30  1.96(5 /10)
n
=30-0.98
=29.02 approximately
Upper confidence interval
 
X  1.96  30  0.98
n
=30.98 =31 approximate
Confidence interval are 29 and 31
out of 100 trees selected out of 95 of
them have a height between 29 and
31.
SMALL SAMPLES:
95% confidence limits of
population are
 S
X  t / 2
n
EXAMPLE: A random sample of 16
students of B.com. made an
average score of 60 with a S.D score
of 15. Construct a 95% confidence
interval for the mean of the entire
B.com class.
SOLUTION: Since the sample size is
smaller than 30, we will use that t-
statistic to construct the required
confidence interval. We are provided
with
 .

X =60
S=15
n=16  n 4
The degrees of freedom =n-1=16-1=15

  0.05  0.025
2
Consulting the t-distribution table, for 15
degrees of freedom, we have,

t0.025  2.131
 Hence using the formula, the 95%
confidence limits would be,
LOWER LIMIT
 S  15 
X  t.025  60  2.131 
n  4
=60-7.99 =52 approximately
UPPER LIMIT
 S
X  t.025  60  7.99
n
=68 approximately
The true mean score  , therefore,
lies between 52 and 68 with a 95%
confindence.