Point and Interval Estimation

Estimation

The objective of estimation is to determine
the approximate value of a population
parameter on the basis of a sample statistic.

For example, sample mean is employed to
estimate the population mean. We refer to
the sample mean as the estimator of the
population mean.

Once the sample mean has been computed,
its value is called the estimate.

Point Estimation

We can use sample data to estimate a population
parameter in two ways.

Point Estimation and Interval Estimation

In point estimation procedure, we make an attempt
to compute a numerical value, from sample
observation, which could be taken as an
approximate to the parameter.

Sample mean is the best possible estimator
of the population mean , because it is unbiased,
consistent, and relatively efficient.
n
X
X

=
Evaluating the Goodness of a
Point Estimator
Unbiasedness
An unbiased estimator of a population parameter
is an estimator whose expected value is equal to
that parameter.
Sample mean is an unbiased estimator of the
population mean because
If we define sample variance using n in the
denominator, the resulting statistic would be a
biased estimator of the population variance.
But if, then
So now is an unbiased estimator of the
population variance
= ) ( X E
1
) (
2
2

=

n
X X
S
2 2
) ( o = S E
2
S
2
o
Evaluating the Goodness of a
Point Estimator

Consistency
An unbiased estimator is said to be consistent, if
the difference between the estimator and the
population parameter becomes smaller as the
sample size grows larger. is a consistent
estimator of , because the variance of is

This implies that as n grows larger, the variance of
grows smaller. As a consequence, larger
sample size will results better estimate.
X
X
. /
2
n o
X
Evaluating the Goodness of a
Point Estimator
Relative Efficiency
If there are two unbiased estimators of a
parameter, the one whose variance is smaller is
said to be relatively efficient.
Sample mean is an unbiased estimator of the
population mean and its variance is

Statisticians have established that (sampling from
a normal population) the sample median is also an
unbiased estimator of the population mean, but its
variance is 1.57 Consequently, we say that
the sample mean is relatively more efficient than
the sample median.
. /
2
n o
. /
2
n o
Interval Estimation
We have to be extremely lucky to have a sample which has
a mean exactly equal to the population mean, otherwise it
will be a little higher or a little lower.

We, therefore, try to determine two values, instead of one
point estimate, within which the true value of the parameter
is expected to fall.

We can also attached a certain degree of confidence to our
estimated value. The two values which are expected to
contain the true value of a population parameter are called
the confidence limits (the lower confidence limit (LCL) and
the upper confidence limit (UCL) and the two together are
called confidence intervals.
Confidence Interval Estimate of
Mean (Large Sample)

o
o

o
o
o

o
o
o

o o
o o
o o
=
(

+ ( (
=
(

( (
=
(
(
(

(
1
1
1
2 2
2 2
2
2 /
n
Z X
n
Z X P
n
Z X
n
Z P
Z
n
X
Z P
o
o

o
o
o

o
o o
o o
=
(

+ ( (
=
(

) ) +
1
1
2 2
2 2
n
Z X
n
Z X P
n
Z X
n
Z X P
n
Z X
o
o
2

n
Z X
o
o
2
+


Upper Confidence Limit (UCL) =



Lower Confidence Limit (UCL) =
o
o o
=
(

( ( 1
2 2
Z Z Z P
Confidence Interval Estimate of
Mean (Large Sample)
90% confidence limits are
n
X and
n
X
o o
64 . 1 64 . 1 +
95% confidence limits are
n
X
n
X
o o
96 . 1 , 96 . 1 +
99% confidence limits are
n
X
n
X
o o
575 . 2 , 575 . 2 +
is usually not known, therefore, for large samples it can be
replaced by s (the sample standard deviation) which may be
calculated from the sample by using the following the formula
o
1
) (
2
2

=

n
X X
S
Example

A random sample of 64 students in a class
made an average score of 60, with a
standard deviation of 15. Construct 99%
confidence interval estimate for the mean
score of entire class.

=> Also construct 90% and 95 % confidence
interval estimate for the mean score of entire
class.
Example- Solution
ely approximat
n
X it conf idence Upper
the Also
ely approximat
n
X it conf idence Lower
n
X erval Conf idence
65
83 . 64
8
15
58 . 2 60
58 . 2 lim
55
16 . 55
8
15
58 . 2 60
58 . 2 lim
58 . 2 int % 99
=
=
|
.
|

\
|
+ =
+ =
=
=
|
.
|

\
|
=
=
+ =
o
o
o
Example- Solution
675 . 63 08 . 63
325 . 56 925 . 56 8 15
\ / 64
% 95 # % 90 60
= =
= = = = =
=
=
UCL UCL
LCL LCL n S
n
X
o
Confidence Interval Estimate of
Mean (Small Sample)

In practice is usually not
known, also the application of central limit
theorem is feasible only when sample size is
large. We thus need to know some such
distribution, based on random samples,
where we could overcome these difficulties.
So if population S.D is known and sample
size is large, (n = 30 is considered as large)
use Z distribution. If population S.D is not
known, but sample size is large, replace
by s (sample S.D) and use Z distribution
) . . ( population of D S o
o
Confidence Interval Estimate of
Mean (Small Sample)

But if sample size is small (n<30) then use t
distribution


With n-1 degrees of freedom.
So confidence limits for the mean of
the population are
n
s
X
t

=
% 100 ) 1 ( o
n
s
df t X it conf idence Upper
n
s
df t X it conf idence Lower
2
2
lim
lim
o
o
+

Definition
Degrees of Freedom (df ) = n - 1

Any
#
Specific
#
so that x = 80
Any
#
n = 10 df = 10 - 1 = 9
Any
#
Any
#
Any
#
Any
#
Any
#
Any
#
Any
#
Important Properties of the Student t
Distribution
1. The Student t distribution is different for different sample sizes (see
Figure 6-5 for the cases n = 3 and n = 12).
2. The Student t distribution has the same general symmetric bell
shape as the normal distribution but it reflects the greater
variability (with wider distributions) that is expected with small
samples.
3. The Student t distribution has a mean of t = 0 (just as the standard
normal distribution has a mean of z = 0).
4. The standard deviation of the Student t distribution varies with the
sample size and is greater than 1 (unlike the standard normal
distribution, which has a o = 1).
5. As the sample size n gets larger, the Student t distribution gets
closer to the normal distribution. For values of n = 30, the
differences are so small that we can use the critical z values
instead of developing a much larger table of critical t values.
Student t
distribution
with n = 3
Student t Distributions for
n = 3 and n = 12
0
Student t
distribution
with n = 12
Standard
normal
distribution
Figure 6-5
Confidence Interval Estimate of
Mean (Small Sample)

Example
A random sample of 10 packets was
taken and is found to have a mean
weight of 60 grams and a standard
deviation of 12 grams. What is the mean
weight of the population

(a) with 95% confidence?
(b) with 99% confidence?
Confidence Interval Estimate of
Mean (Small Sample)



95% confidence limits are
confidence
df S X
% 95
9 1 10 12 60 = = = = = u
025 . 0
2
05 . 0 = = = +
o
o
n
s
t X
58 . 68 ~ 42 . 51 262 . 2
025 . 0 , 9
= t
33 . 72 67 . 47
33 . 12 60 250 . 3
10
12
250 . 3 60
005 . 0 01 . 0 % 99
005 . 0 , 9
2
and
t t
=
+ = = =
|
.
|

\
|
+
= =
o
o o
Inference about a Population
Variance



2
o
.
1
) (
2
2
|
|
.
|

\
|

=

n
X X
S
2
S
2
o
Point Estimator
The point estimator for is the sample variance
is an unbiased, consistent estimator of
Sampling distribution of
To create the sampling distribution of the sample variance, we
repeatedly take samples of size n from a normal population
whose variance is , calculate for each sample
2
S
2
o
2
S

=
2 2
2
2
) ( ) 1 (
1
) (
X X S n
n
X X
S
Mathematician have shown that the sum of squared difference

2
) ( X X | |
2
) 1 ( S n to equal is which
divided by the population variance is distributed according
to what is called the chi-squared distribution provided that the
population is normal.
where
n = sample size
s
2
= sample variance
o
2

= population variance

Chi-Square Distribution
X
2
=
o
2
(n - 1) s
2
Properties of the Distribution of
the Chi-Square Statistic
1. The chi-square distribution is not symmetric, unlike
the normal and Student t distributions.
As the number of degrees of freedom increases, the
distribution becomes more symmetric. (continued)
0 5 10 15 20 25 30 35 40 45
Figure 6-8 Chi-Square Distribution for df = 10
and df = 20
df = 10
df = 20
Figure 6-7 Chi-Square Distribution
All values are nonnegative
Not symmetric
x
2
0
Properties of the Distribution of
the Chi-Square Statistic
(continued)
2. The values of chi-square can be zero or positive, but
they cannot be negative.
3. The chi-square distribution is different for each
number of degrees of freedom, which is df = n - 1
in this section. As the number increases, the chi-
square distribution approaches a normal
distribution.

A
A
_
2
A
_
2
1-A

The _
2
table
Degrees of
freedom
1 0.0000393 0.0001571 0.0009821 . . 6.6349 7.87944
.
.
10 2.15585 2.55821 3.24697 . . 23.2093 25.1882
. . . . . .
. . . . . . . .
_
2
.995
_
2
.990
_
2
.975
_
2
.010
_
2
.005

.990 .010
=.01
=.01
1 - A =.99
A
_
2
.01,10
= 23.2093
Inference about a Population
Variance
| |
o
o
o
o
o o
o o
=
(

(
= ( (

1
) 1 (
1
lim % 100 ) 1 (
2
2
2
2
2
1
2
2
2 2
2
1
2
X
S n
X P
X X X P
are its conf idence The
o o
o o
o
o
o o
o o
o o
=
(
(

( (

=
(
(

) )

=
(
(

( (

1
) 1 ( ) 1 (
1
) 1 ( ) 1 (
1
) 1 (
1
) 1 (
2
1
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2 2
2
1
2
X
S n
X
S n
P
X
S n
X
S n
P
S n
X
S n
X
P
Inference about a Population
Variance
The
% 100 ) 1 ( o
lower and upper confidence limits of the
population variance are
2
1
2
2
2
2
2
) 1 (
lim
) 1 (
lim
o
o

=
X
S n
it conf idence Upper
X
S n
it conf idence Lower
2
2
o
X
2
1
2
o
X
Where n-1 are the degrees of freedom and the value of
and
are available from the chi-square table against
(n-1) degrees of freedom and the appropriate level of significance.
95 . 0
2
05 . 0
2
05 . 0
2
1 . 0 X X If = =
o
o
area the of contain will X X % 95 025 . 0
2
05 . 0 975 . 0
2
025 . 0
2
= =
o
o
Inference about a Population
Variance
Example
A random sample of 10 bottles of a cough syrup
found to have an average alcohol content of 3.5
m.l. with a variance of 0.64 m.l. Construct a 99
percent confidence interval for the true standard
deviation alcohol contents of the cough syrup.