Unit V: Estimation

Objectives:
The main focus of this unit is to explain the concept of
point and interval estimation, their properties, CLT,
confidence interval for mean and proportion, and
sample size estimation
Estimation:
Concept:
Estimation:
 The techniques by which sample is selected from
the population and calculate the sample statistic
then make an estimate about the population
parameter.

 It is the process of finding the true value of

unknown population parameter by using the
known value of sample statistic.
Purpose of Estimating Parameter:
 The objective or goal of any researcher is to estimate
the value of population parameter which is generally
unknown.
 The value of population parameter is estimated from
the sample taken from that population which has
limited information about the population.
 One of the important tasks of the inferential statistics
is related to such kind of estimation of population
from the sample.
 It is the art of drawing conclusion about the
population from sample drawn from it.
Steps for the estimation
 The population related to the study is defined.
 Identify the parameter of the interest for
estimation.
 Sample is selected from that population.
 Information is collected from each unit of the
sample.
 The value of statistic relevant to the parameter is
estimated from sample.
 The value of population parameter is estimated
from the sample statistic.
Basic Terms
Estimate: The particular value of the sample
statistic calculated from the sample is said to be
estimate of the parameter.
Estimator: The function or formula which is used
for estimating the sample statistic (i.e. estimate) is
estimator.
Types of estimation:
1. Point estimation: Summarize the sample by a
single number that is an estimate of the population
parameter.

2. Interval estimation: A range of values within

which, we believe, the true parameter lies with
specified probability.
Point Estimation:
 The estimation of a single value of unknown
population parameter from the sample is the point
estimation.
 In this estimation parameter value is exactly
estimated.
 It is single value which summarizes the population.
 The estimated value is either accepted or rejected.
Example:
• If a researcher wanted to estimate the mean
income of a class of business students then a mean
income for the sample of 250 students is found to
be Rs.12000 per week, it is point estimate.
• In a random sample of 100 television viewers, if 55
persons like a particular TV program then 0.55 is
the point estimate of population proportion.
Methods of Point Estimate:
1. Methods of moments (MOM).
2. Methods of least square (MLS).
3. Methods of maximum likelihood (MLE).
4. Numerical method
Point Estimate of Mean:
 The sample mean 𝒙 is point estimate of population mean.
 If the mean profit of 25 firms is Rs.2.1 lakhs drawn from the
population having 1000 firms then it is point estimate of the
whole firms.

Point estimate of proportion:

 The sample proportion 𝒑 is point estimate of population
proportion P.
 In a sample of 500 persons 175 are supported for the social
security fund proposed by the government then the sample
proportion 0.35 is the point estimate.
Properties of good estimator
An estimator is said to be good if it is closed to the true
population parameter.
Four different criteria of good estimator:
If t be the statistic and  be the parameter then
a) Unbiasness : The estimate is said to be Unbiasness if the
expected or the average values of the sample statistic is equal
to the population parameter. If these two values are not equal
then the difference between them is biased value.
E(t) = . Then, statistic ‘t’ is unbiased estimate of parameter .
 The sample mean is unbiased estimate of the
population mean.
 The sample proportion is unbiased estimate of the
population proportion.
 The sample variance is unbiased estimate of population
variance.
 The sample standard deviation is biased estimate of
population standard deviation.
 The sample median (or mode) is biased estimate of
population mean.
(b) Consistency:
The statistic which is estimator of the population
parameter is said to be consistent estimate as the
sample size n is increases the value of statistic is
approached to the value of population parameter.
𝒍𝒊𝒎 𝑬 𝒕𝒏 = 𝜽
𝒏→∞
That is the statistic ‘t’ tends to the parameter ‘’ as n
becomes large.
(c) Efficiency:
 If t1 and t2 are two unbiased estimator of a population
parameter, then t1 is said to be more efficient
estimator than t2 if the variance of statistic t1 is less
than that of statistic t2.
 That is, the statistic which have less variance among
many statistics is efficient estimator of parameter .
 The estimator having minimum standard error is the
efficient estimator.
 Var(t1) < Var(t2), then t1 is efficient estimator than the
t2.
(d) Sufficiency: A statistic ‘t’ is said to be sufficient
estimator of the population parameter  if it
contains all information about the population
parameter contained in the sample.
For example: the samples mean is sufficient estimate
of population mean but not median.
Confidence Coefficient
 The confidence level stated in the proportion
rather than in the percentage.
 It is the area under the curve and between
calculated lower and upper limits of the
parameter.
 Higher the value of confidence coefficient, more is
the certainty of the results of the estimate.
 For example, if the confidence level is 95% then
the confidence coefficient is 0.95.
 And for 99% confidence level the confidence
coefficient is 0.99.
Confidence Interval Example
 Population has µ = 368 and σ = 15.
 If you take a sample of size n = 25 you know
 368 ± 1.96 * 15 /25= (362.12, 373.88) contains 95%
of the sample means
 When you don’t know µ, you use 𝑿 to estimate µ
 If 𝑿 = 362.3 the interval is 362.3 ± 1.96 * 15 /25 =
(356.42, 368.18)
 Since 356.42 ≤ µ ≤ 368.18 the interval based on this
sample makes a correct statement about µ.
But what about the intervals from other possible samples
of size 25?
 Confidence Interval Example:
Sample Sample Lower Upper Contain
No. Mean Limt Limit ?
(𝑋)
1 362.30 356.42 368.18 Yes
2 369.50 363.62 375.38 Yes
3 360.00 354.12 365.88 No
4 362.12 356.24 368.00 Yes
5 373.88 368.00 379.76 Yes
• In practice you only take one sample of size n
• Probably, you do not know µ, so you do not know if the
interval actually contains µ
• However you do know that 95% of the intervals formed in
this manner will contain .
• Based on the one sample, you actually selected that you can
be 95% confident your interval will contain µ (this is a 95%
confidence interval)
Note: 95% confidence is based on the fact that we used Z =
1.96.
Steps for Confidence interval (Fudicial Limits) for
Estimating Population Parameter.
* Obtain the appropriate sample statistics `T`
* Compute the standard error of the sampling
distribution of the statistics `T` i.e. S.E.(T)
* Choose the appropriate confidence level (1-α)
depending on the exactness of the estimate.
* Use the expression as
CI for Parameter = Point Estimate ± (Critical
Value)(Standard Error)
OR
CI of θ  T  Critical Value * SE(T)
Where
T = Point estimate is sample statistic
Critical value= Table value based on the sampling
distribution of the statistic and desired confidence
level.
Standard Error = The standard deviation of the
sample statistic.
Confidence Level :
 It is the value in percentage that how much
confidence the interval will contain the unknown
population parameter.
 It is less than 100%.
 It is also written as 1-  i.e. if confidence level is 95%
then (1 -) = 0.95 (then  = 0.05)
 Interpretation: It is interpreted as 95% of all the
confidence intervals will contain the true population
parameter.
Confidence Interval Estimation:

Confidence
Intervals

Population Population
Mean Proportion

 Known  Unknown
Confidence Interval for  when  is known
Assumption:
 The value of population standard deviation is
known.
 The population is normally distributed.
 If population is not normally distributed, large
sample is used.
Formula:

X  Z / 2 .SE ( X )  X  Z / 2 .
n
Where,
𝑋 = Sample mean
 = population standard deviation.
 = level of significance
Z/2 = Critical value from Z- table.
 Common Level of Confidence
Confidence Confidence  /2 Critical
Level Coefficient Value
(1-) Z/2
50% 0.50 0.50 0.25 0.6745
80% 0.80 0.20 0.10 1.28
90% 0.90 0.10 0.05 1.645
95% 0.95 0.05 0.025 1.96
98% 0.98 0.02 0.01 2.33
99% 0.99 0.01 0.005 2.58
99.9% 0.999 0.001 0.0005 3.27
Example 1: A sample of 150 light bulbs from the lot
having mean life of 2010 hours. From the past data
the standard deviation of the life of light bulb is
found to be 20 hours.
Calculate the standard error of mean and set up
95% and 99% confidence limits for the true
population mean.
Solution : Here, n= 150, 𝑋= 2010 hours,  = 20
hours
(i) To find the standard error of mean:
We have,
 20
SE ( X )    1.63
n 150
(ii) To find the confidence interval for true
population mean
(a) For 95% confidence level

X  Z / 2  SE ( X )  X  Z / 2 
n
 2010  1.96  1.63
 2010  3.20
 (2006.80,2013.20)
i.e. 2006.80 ≤  ≤ 2013.20
Interpretation: We are 95% confident that the
interval (2006.80, 2013.20) contains the true
population mean life of light bulbs.
(b) For 99% confidence level

X  Z / 2  SE ( X )  X  Z / 2 
n
 2010  2.58  1.63
 2010  4.21
 (2005.79,2014.21)
i.e. 2005.79 ≤  ≤ 2014.21
Interpretation: We are 99% confident that the
interval (2005.79, 2014.21) contains the true
population mean life of light bulbs.
Thank You