CHAPTER 3

STATISTICA
L CHAPTER 3

ESTIMATIO
N
 Chapter Objectives

Recognize the concepts of statistical estimation.

Estimate the population mean and solve real-world

business problems.

Estimate the population proportion and solve real-

world business problems.

Understand the concept of sample size determination.

• Statistical inference : is a process of making
inference/drawing conclusion about a population
from a sample drawn from that population.

• It is about making sensible decisions under

uncertainty
• Make general conclusion about population from
limited amounts of data.
• It mainly includes: statistical estimation and
hypothesis testing.
Estimator and Estimate
• Estimator: Statistical measure that is used to
make inference about unknown population
parameter
• Any quantity that is calculated from sample data
and it is used to give information about unknown
values of a population
Examples: Measures of Central Tendency,
Measures of Dispersion, Frequencies.
Estimate: a single value of an estimator
• It is an indicator of value of a parameter
Example: Consider the data: 1,2,3,3,6
The mean is 3 …. The mean is estimator and 3 is
an estimate.
Types of Estimation:
• Point Estimation
• Interval Estimation

I. Point Estimation
• In point estimation a parameter is estimated
by a single value ( i.e we give a number as
an estimate for the parameter)
Examples: sample mean, sample proportion,
sample variance, sample standard deviation.
 Properties of best point estimators

1. Unbiasedness
 The expected value of the estimator equals the value
of the parameter to be estimated (i.e an estimator
neither overestimates nor underestimate the
parameter on average).
Example: sample mean, mode and median are
estimators of population mean, but it is only sample
mean that is unbiased estimator of population mean
2. Consistency
Effect of sample size on the accuracy of estimator
A statistic is consistent estimator of the parameter if it
gets closer and closer to the parameter as n increases
3. Efficiency
Refers to sampling variability of an estimator.
Given two competing estimators; the one that has
the minimum variance is more efficient than the
other.
Smaller variance >>> the more concentrated is the
distribution of estimator around the parameter >>>
better estimator.
4. Sufficiency
Ensures that all information a sample can furnish
with respect to estimation of parameter being
utilized.
Value of any sample from the distribution gives no
more information about the parameter than does
the value of the estimator on that sample
The following are some examples of parameters
and their best estimators:

 Sample mean is best point estimator of population

mean.
 Sample proportion is best point estimator of
population Proportion
 Sample variance is best point estimator of
population Variance.
 Sample standard deviation is best point estimator
of population.
II. Interval Estimation
 Describes a range of values within which a parameter is
expected to lie.
Confidence Interval estimates of population mean
 It is an interval of values a to b within which an unknown
population mean is expected to lie Or
It is an interval estimate together with a statement of how
confident we are that the interval is correct.
The interval is an inference based upon:

 the value of sample mean

 Known facts about the sampling distribution of the mean

NB: The choice of method used in constructing a confidence

interval for population mean depends on whether or not the
population is normal and whether the population standard
deviation is known or unknown.
Case 1:

• Normal population
•Population standard deviation () is known/given
•Any sample size(n); small or large
A 100(1-)% confidence interval estimate of the
population mean () is given by:
Example 1: A manufacturer claims that a particular automobile
model will get 50 miles per gallon on the highway. The
researchers at a consumer-oriented magazine believe that
this claim is high and plan a test with a simple random sample
of 30 cars and the sample mean is 49 miles per gallon. What
should the researchers conclude if confidence coefficient is
0.95? Assume that the population speed of cars is normal, and
the standard deviation between individual cars is 2.3 miles
per gallon for the population data.
E=0.82 48.18 to 49.82

Example 2: In a random sample of 49 banks in a country, the

average profit per month is $20000 with S.D of profit of all
banks in the country be $1000. Assuming that profit is
normally distributed,
 Find the point estimate of the profit of all banks in the country
 Construct a 95% C.I estimate of the profit of all banks in the
country
19720 to 20280……e= 280
Case 2

•Non-Normal population
•sample size(n) is large (n>=30)
•Population standard deviation() is unknown

A 100(1-)% confidence interval estimate of

the population mean () is given by:
Example: A producer of a given furniture claims that his
products serve for 10 years on average. A consumer
protection agency takes a random sample of 400
furniture produced by the manufacturer and finds an
average of 9.2 years with standard deviation of 3 years.

i. Find a 99% confidence interval estimate for the mean

service age of all furniture of the manufacturer. Ans:
e=.387, 8.8 to 9.6
ii. Does the result support the claim?
Exercise: In order to ensure efficient usage of a server, it
is necessary to estimate the mean number of concurrent
users. The sample mean and sample standard deviation
of number of concurrent users of 100 randomly selected
items is 37.7 and 9.2, respectively. Construct a 99%
confidence interval for the mean number of concurrent
users.
Case 3

•Normal population
•Standard deviation() is unknown
•Small sample size; n<30

If sample size is small the sampling distribution of

sample mean will not be Normal; it will rather follow t-
distribution with n-1 degrees of freedom.

x 
T  ~ t ( n  1)
S
n
A 100(1-)% confidence interval estimate of  is given
by:

Example 1: A logistics company wants to estimate the average

delivery time for a certain route. They randomly select 25
deliveries and record the time it takes for each delivery. They
calculate a mean delivery time of 3 days with a standard
deviation of 0.5 days. What is the 95% confidence interval for
the true delivery time? (Assume that delivery time is normal).
 Example 2: The environmental protection officer of a large
industry plant sought to determine the mean daily amount
of sulfur oxides ( a pollutant emitted by the plant). Because
measurement costs were high, only a random sample of 10
days’ measurements (tons per day) were obtained with
mean 9.5 and standard deviation 3.24. Suppose emissions
per day are normally distributed. Estimate the mean amount
of all sulfur oxides emitted per day, using a confidence
interval with a confidence coefficient of 0.95.
Ans: (7.2, 11.8)
 Exercises
1. There are 500 employees in a company, and their annual salaries
follow a normal distribution with an unknown mean μ and standard
deviation σ= $4,000. A researcher has randomly selected 80
employees from the population and the mean annual salary is
calculated to be $50000. Find the 95% confidence interval estimate for
the average salary of the population under study.

2. A company wants to understand customer satisfaction levels (1=

very dissatisfied, 2= dissatisfied, 3= neutral, 4= satisfied, 5= very
satisfied) across its 1,500 active customers who made a purchase in
the last month. The researcher surveyed a sample of 50 customers and
calculated the average satisfaction score as 4.2. What is the 99%
confidence interval for the mean satisfaction score of all 1,500
customers?

3. An IT department manages 80 servers in a data center. They are

interested in estimating the mean downtime (in minutes) per week.
After measuring downtime on a random sample of 20 servers, they
calculate the average downtime and standard deviation as 15 minutes
and 3 minutes, respectively. What is the 99% confidence interval
Confidence interval estimate of population
proportion (P)

•For large n; i.e nP>5 and nQ>5

Example 1: A company wants to estimate the proportion of all
customers who are likely to repurchase their product. They randomly
sample 200 customers and find that 140 of them have repurchased the
product. Find the 95% confidence interval estimate for the proportion of
the population under study.

Conclusion: The 95% confidence interval is (0.64, 0.76), which would

suggest that the true proportion of customers who are likely to
repurchase the product is likely to fall between these two values with
95% confidence. This information can help the company make informed
decisions regarding their marketing and business strategies.

Exercise: A random sample of 400 members of labor force in a five

stage region showed that 32 were unemployed. Construct a 99%
confidence interval estimate for the population of unemployed in
the region.
e = 0.03; (0.05, 0.11)
B: For finite population, we take finite population correction
tor into account, and hence a 100(1-)% Confidence interva
timate for:
Population Mean (µ ) is:

Case 1

Case 2

Case 3

i. Population Proportion (P) is:

 Sample size determination when estimating
population Mean
Research is done to find a solution to a particular problem (formulated
as a research question) based on statistics. In an ideal situation, the
entire population should be studied but this is almost impossible.

Other than census, which is conducted on each and every observation

of the population, all other studies are performed on limited
observations drawn from the concerned population known as target
population.

The data obtained is analyzed and conclusions are drawn which are
extrapolated to the population under study.
•Sample size determination is an important aspect of Marketing
Management that helps researchers to conduct research and make
informed decisions based on the collected data.
1. Infinite population:

The sample size required when estimating population

mean for infinite population is given by:

where,
 n is the sample size desired
 α is the level of significance
Z is the critical value that cuts off an area of α/2 in the normal
α/2
curve (the standard normal deviant)
 is the standard deviation
 E is an error of estimation
2. Finite population:

For finite population, the sample size required to estimate

population proportion is given by:

Take FPC into account

where
 Example 1: A manufacturer wants to conduct a survey to determine the average delivery time for a
particular supplier. The manufacturer expects the delivery time to be 5 days with a standard deviation
of 2 days. How large of a sample is needed to estimate the true delivery time with a 95% confidence
level and a margin of error of ±0.5 days?
Ans: 30.8 ~ 31.
Hence, we need a sample size of at least 31 to estimate the mean delivery time
with a 95% confidence level and a margin of error of ±0.5 days.
Example 2: A farm has 1000 young pigs with an initial weight of about 50 lbs. They put them on a
new diet for 3 weeks and want to know how many pigs to sample so that they can estimate the
average weight gain. They want the answer to be within 2 lbs., with 95 percent confidence. (assume
that σ=10bs). no=96.04, n=87.62~88
Exercises:
1.Suppose Mr. Abebe would like to start an Internet service provider (ISP) and need to estimate the
average Internet usage of households in one week for his business plan and model. How many
households must he randomly select to be 95 percent sure that the marginal error is within 1 minute
of the population mean? Assume that a previous survey of household usage has shown = 6.95
minutes. Ans: 185.558~186
2. A company wants to conduct market research to identify the average income of their target
audience. They want a confidence level of 95% and a margin of error of no more than $5,000. The
company knows that the standard deviation for the population is $20,000. What is the minimum
sample size required for this study?
 Sample size determination when estimating population
proportion

The sample size required when estimating population proportion

for infinite population is given by:

For finite population, the sample size required to estimate

population proportion is given by

where
Examples
1. The researcher in company ABC wants to estimate the proportion of customers that will
purchase its new product. The researcher wants to be 95% confident that the estimate is to be
within .03 of the true proportion. How many customers should he sample? n=1067.1~1068

2. A researcher is studying the population of a small town in Ethiopia of N = 2000 people. She is
interested in estimating P for several yes/no questions on a survey. How many
people (n) does she have to randomly sample to ensure that her estimates (sample proportion)
are within 0.04 of the true proportion P? (assume that level of significance is 0.05 ).

Exercises
1. A logistics company wants to determine the proportion of late deliveries for a particular route.
The company expects the proportion of late deliveries to be 20%. How large of a sample is
needed to estimate the true proportion with a 99% confidence level and a margin of error of
±5%?
Ans: n = 333.6~334
2: A marketing research firm wants to conduct a survey to estimate the proportion of consumers
who are likely to buy a new product that a client is introducing. The firm has determined that
they want to be 95% confident in their results and have a margin of error of no more than 5%.
The population size is 10,000. What is the minimum sample size required for this survey?
(Assume p-hat = 0.5) )
THE END