MASTER OF COMMERCE AND

BUSINESS ADMINISTRATION

LEARNING

MODULE 5
√
1+ −3
Sampling Techniques ρ :=
2

The Open University of Kenya

Developed by: Dr. Jacob Ong’ala

Reviewed by:
Programme Title Masters of Commerce & Business Administration

Course Title MCA 807 - Statistical Methods

Learning Module number Module 5

Learning module title Sampling Techniques

Module Developer Dr. Jacob Ong’ala

Reviewed by
LEARNING MODULE 5 MCA 807 - Statistical Methods

Module 5 - Sampling Techniques

Instructional Hours: 4

Ð Module Overview

This module focuses on sampling methods and their critical role in business research. You will
explore various sampling techniques, learn how to design surveys or experiments tailored to
specific business problems, and calculate the appropriate sample size for different research
scenarios. Additionally, you will develop strategies to minimize sampling errors and biases,
ensuring the accuracy and reliability of your research outcomes

Module Learning Outcomes

By the end of this module, you should be able to:
1. Describe different types of sampling methods and their applications in
business research.
2. Use appropriate sampling technique for business reserch problem.
3. Calculate the appropriate sample size required for different research sce-
narios, using statistical formulas.
4. Develop strategies for minimizing sampling errors and biases in business
research.

Learning Activities

1. Read the lecture notes (PDF)

2. Read the assigned reference materials
3. Watch lecture videos
4. Complete the module assessment/Quiz

START OFF REFLECTION

Imagine you’re tasked with conducting a nationwide survey to understand the spending habits
of university students. You’ve mastered how to analyze and present data, and you’re familiar
with patterns and probabilities. But now, the challenge is: How do you choose a small group
of students to survey that truly represents the entire student population, while avoiding er-
rors and biases? To find the solution, you’ll need to explore methods for selecting samples
effectively, which is the focus of this next module

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 1 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

READING TASKS

Read the following Lecture notes and attempt the tasks that follows

1. Introduction to sampling

1.1. What is sampling

Sampling is the process of selecting a subset from a population to draw conclusions about the
entire population. Below are some key sampling techniques:

ACTIVITIES

You will read the notes provided in the link below and watch the video also after-which you
are required to participate in the discussion forum

• Lecture Notes (page 1 upto page 4 only): Introduction to Sampling Theory

Content Curated Video

ç • Sampling Vs Population
• What is sampling

ONLINE DISCUSSION FORUM AND REFLECTION

Do this question and post your thought into a discussion forum.

• A business analyst needs to address a key business problem through research. Even if
resources are available to gather data from the entire population, why might the analyst
choose to take a sample instead?

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 2 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

1.1.1. Probability and Nonprobability Sampling

Probability sampling is a sampling process that utilizes some form of random selection. In
probability sampling, each unit is drawn with known probability, or has a nonzero chance of
being selected in the sample. With probability sampling, a measure of sampling variation can
be obtained objectively from the sample itself.
Nonprobability sampling or judgment sampling depends on subjective judgment. The non-
probability method of sampling is a process where probabilities cannot be assigned to the
units objectively, and hence it becomes difficult to determine the reliability of the sample re-
sults in terms of probability. In nonpraobability sampling, often, the surveyor selects a sample
according to his convenience, or generality in nature. Nonprobability sampling is well suited
for exploratory research intended to generate new ideas that will be systematically tested
later.

1.1.2. Sampling Errors

Sampling errors arise when estimates (such as the mean, total, or proportion) are derived from
a sample rather than the entire population. These errors occur because the sample estimate
is unlikely to match the true population value exactly. For instance, if you use a sample of
neighborhoods to estimate the total number of residents in a city, and the neighborhoods in
your sample are larger than average, your estimate will likely overestimate the actual popula-
tion. This discrepancy happens because the sample may not accurately represent the overall
population, leading to an erroneous conclusion.

ő Example 1:
Scenario: A company wants to estimate the average satisfaction score of its cus-
tomers based on a survey. Suppose the company uses a sample of 100 customers
from a total customer base of 1,000.
Example:
• Population Mean Satisfaction Score: 7.5 (on a scale of 1 to 10)
• Sample Mean Satisfaction Score: 8.0
If the sample is not representative (e.g., if it includes more highly satisfied customers
due to a biased survey invitation), the sample mean might overestimate the true
population mean.
Illustration:
Sum of all satisfaction scores
True Population Mean = = 7.5
Total number of customers
Sum of satisfaction scores in the sample
Sample Mean = = 8.0
Number of sampled customers
The difference between the sample mean and the population mean is a sampling
error.

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 3 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

1.1.3. Nonsampling Errors

The accuracy of an estimate can be affected by errors that arise from factors other than sam-
pling, known as nonsampling errors. These errors may include incomplete coverage, faulty
estimation procedures, and observational errors. The goal of a survey is to obtain informa-
tion as close as possible to the true population value, within the available resources. The
difference between the survey value and the true value is termed as observational error or
response error.

ő Example 2:
• Improper Records: Errors due to incorrect or incomplete data entries.
• Careless Reporting: Mistakes made when reporting data.
• Deliberate Modification: Data falsification by collectors to align with their inter-
ests.
• Nonresponse Error: Occurs when a significant number of individuals do not
respond to the survey or when the respondents differ significantly from non-
respondents in ways important to the study.
Illustration: If a survey aims to estimate the average income of a population but
only collects data from people who are employed, the resulting estimate will likely
be biased, as it does not account for those who are unemployed or otherwise not
represented in the sample.

1.1.4. Bias

Bias refers to systematic errors that can affect the results of a study, especially when judgment
sampling is used instead of probability sampling. For instance, if a surveyor is tasked with
selecting 20 books from a total of 200 to estimate the average number of pages per book, they
might choose books that appear to be of average size. This method is problematic because
the selection process may introduce systematic errors.

ő Example 3:
• If the surveyor consciously or unconsciously selects books that seem to be of av-
erage size, the resulting sample may not accurately represent the true diversity of
book sizes in the population. This can lead to a biased estimate of the average
number of pages.
Illustration: Suppose the true average number of pages in all 200 books is 250, but
due to biased sampling, the average number of pages in the selected 20 books is
220. The discrepancy reflects the bias introduced by the selection process.

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 4 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

ONLINE DISCUSSION FORUM AND REFLECTION

Discuss the following questions in your groups and come up with possible control over sam-
pling errors, nonsampling errors and bias in research.

Discussion Question 1

A company is conducting a survey to assess customer satisfaction with its new product. They
select a sample of customers who recently made a purchase from their online store and send
them a satisfaction survey. After analyzing the responses, they find that the average satisfac-
tion score is significantly higher than expected.
1. Possible Sampling Errors: What are some potential sampling errors that could have
occurred in this scenario? How might the sample not represent the entire customer base
accurately?
2. Nonsampling Errors: What nonsampling errors might have influenced the survey re-
sults? Consider factors such as data recording or response accuracy.
3. Bias: What types of bias could be introduced by selecting only customers who made
recent online purchases? How might this affect the overall satisfaction estimate?

Discussion Question 2

A market research firm is tasked with estimating the average annual spending of consumers
on electronics. They use a judgment sampling method, where researchers select individuals
who they believe are typical electronics buyers based on their appearance and behavior at a
shopping mall.
1. Possible Sampling Errors: What are the potential sampling errors associated with this
method of selecting participants? How might this sampling method impact the accuracy
of the estimate?
2. Nonsampling Errors: What nonsampling errors could arise from this judgment-based
sampling approach? Think about errors related to data collection and reporting.
3. Bias: What types of bias are likely in this scenario? How might the selection process
based on appearance and behavior at the mall affect the representativeness of the sam-
ple and the resulting estimate of annual spending?

2. Probabilistic/Random Sampling techniques

2.1. Simple Random Sampling (SRS)

Each member of the population has an equal chance of being selected. This method is unbi-
ased but may not be efficient for heterogeneous populations.

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 5 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

Sample surveys are typically designed to draw samples from a population consisting of a finite
number of N units. If the units in the population can be distinguished from one another, the
number of possible distinct samples of size n that can be selected from N units is given by the
combinatorial formula. The goal in simple random sampling is to ensure that each of these
combinations has an equal probability of being chosen, meaning every unit in the population
has an equal chance of being included in the sample .
The most common way to implement simple random sampling is to use:
• A table of random numbers (Click here to see the table )
• A computer-based random number generator (in this case we will use R)).
• Mechanical devices for random selection

ő Example 4:
Consider a school with N = 850 students, and you want to select a simple random
sample of n = 10 students. Each student is assigned a unique number from 1 to 850.
Since our population has three-digit numbers, we use a random number generator
or table to select random numbers with three digits. Any number exceeding 850 is
ignored, and if a number repeats, it is disregarded.
For instance, using columns 4,5, and 6 of the random number table provided in in
the extract below, the selected sample is as follows:

627, 054, 044, 745, 532, 035, 077, 105, 158, 212

Simple random sampling provides an unbiased method of selecting units from a population,
ensuring that each possible sample has an equal chance of being chosen. This method is
foundational in many sampling designs.

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 6 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

ACTIVITIES

2.1.1. Simple random sampling using R

Run the following R script to get the sample.

Tasks
• Try running the script several times several times.
• What do you notice?
• Explain what your results in the group forum
1 # Set the population size
2 N <- 850
3

4 # Set the sample size

5 n <- 10
6

7 # Generate a simple random sample without replacement

8 [Link] (123) # Setting a seed for reproducibility
9 sample _ students <- sample (1:N, n, replace = FALSE )
10

11 # Display the selected sample

12 sample _ students

2.2. Stratified Sampling

The population is divided into subgroups called strata, and a random sample is taken from
each stratum. This ensures representation from each group and reduces variance within strata.

ACTIVITIES

Go through this article by By Adam Hayes How Stratified Random Sampling Works, With
Examples He has explained in a simple way how stratified sampling works after which you
will be expected to take a quiz at the end of this section.

Content Curated Video

ç Here is also another video that will help you understand the con-
cept of stratified sampling.
This video is going to demonstrate how stratified sampling is done

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 7 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

ç stratified Sampling

2.3. Systematic Sampling

Systematic sampling is an effective method for selecting samples from a large, well-organized
population. Unlike simple random sampling, where each unit has an equal chance of being
selected, systematic sampling involves choosing units at regular intervals from a list of the
population. This approach begins with a random starting point and then selects every k-th
unit, where k is determined by the total number of units N divided by the desired sample size
N
n, i.e., k = .
n

Procedure

To perform systematic sampling, follow these steps:

1. Number the units: Assign a unique identifier to each of the N units in the population.
2. Determine the sample size: Decide the sample size n that you want to take from the
population.
N
3. Calculate the interval size: Find the interval k, which is determined by k = . This
n
means you will select every k-th unit.
4. Random start: Randomly select a starting unit from the first k units in the population.
5. Select every k-th unit: From the randomly selected starting point, choose every k-th
unit to be part of the sample until you have n units.

ő Example 5:
Let’s consider a practical example in the business world. A company wants to con-
duct a survey to understand customer satisfaction with a new product. The popula-
tion consists of N = 1000 customers, and the company wants to sample n = 100
customers for the survey.
1. First, all 1000 customers are listed and assigned a number from 1 to 1000.
1000
2. The company calculates the interval size k = = 10. This means every
100
10th customer will be selected.
3. A random number between 1 and 10 is chosen. Assume the random number
is 4.
4. Starting from the 4th customer, the company selects every 10th customer (4th,
14th, 24th, …) until they have 100 customers in the sample.

Advantages:
• Easy to implement, especially for large populations.
• Ensures a spread-out sample across the entire population.
• Less time-consuming and more efficient than simple random sampling.

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 8 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

Disadvantages:
• May introduce bias if the population has periodic patterns that align with the sampling in-
terval.
• Requires a complete list of the population in advance.

ACTIVITIES

Here is an R script Run it in R environment to learn how to sample systematically from a

population using R.

Practical Example in R

Below is an R code snippet demonstrating how to perform systematic sampling on a dataset

of 1,000 customers.
1 # Load necessary libraries
2 library (dplyr )
3

4 # Set seed for reproducibility

5 [Link] (123)
6

7 # Create a sample dataset of 1 ,000 customers

8 customer _ids <- 1:1000
9 customer _data <- data. frame (
10 CustomerID = customer _ids ,
11 SatisfactionScore = sample (1:10 , 1000 , replace = TRUE) # Random
satisfaction scores from 1 to 10
12 )
13

14 # Parameters for systematic sampling

15 N <- nrow( customer _data) # Total number of customers
16 n <- 100 # Desired sample size
17 k <- floor (N / n) # Interval size
18 start <- sample (1:k, 1) # Random starting point
19

20 # Perform systematic sampling

21 systematic _ sample _ indices <- seq(start , N, by = k)
22 systematic _ sample <- customer _data[ systematic _ sample _indices , ]
23

24 # Display the sample

25 print ( systematic _ sample )
26

27 # Optional : Summary of the sampled data

28 summary ( systematic _ sample )

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 9 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

Explanation of the R Code

The R code performs the following steps:

1. Loads the ‘dplyr‘ library for data manipulation.
2. Creates a dataset of 1,000 customers with a ‘CustomerID‘ and a ‘SatisfactionScore‘.
3. Defines parameters for systematic sampling: total number of customers (‘N‘), sample
size (‘n‘), and interval size (‘k‘), and selects a random starting point.
4. Performs systematic sampling by selecting every k-th customer starting from the ran-
dom point.
5. Displays the sampled data and provides a summary.

2.4. Cluster Sampling

Cluster sampling is a method used when a population is divided into clusters or groups, and
a random sample of these clusters is selected. All members of the selected clusters are then
included in the sample. This method is particularly useful when dealing with large populations
that are spread across different locations or when a complete list of the population is not
available. This method is cost-effective but increases variability within clusters.

Procedure

1. Divide the population into clusters: The population is segmented into groups or clus-
ters, which are often naturally occurring.
2. Randomly select clusters: Choose a random sample of clusters.
3. Include all units within selected clusters: Every unit within the selected clusters is
included in the sample.
In a business environment, cluster sampling is useful when conducting surveys or quality
checks across multiple branches or locations. For example, if a company has 50 stores and
wants to evaluate customer satisfaction, they could randomly select a few stores (clusters)
and survey all customers within those stores.

ACTIVITIES

in this activity, read through the article which explains a step by step to cluster sampling. Also
run throgh the R snipset to learn how to use Software in cluster sampling. Then you are ex-
pected to undertake the quiz at the end of this section

• A Simple Step-by-Step Guide to Cluster sampling with Examples

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 10 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

Practical Example in R

Below is an R code snippet demonstrating how to perform cluster sampling on a dataset

representing different stores and customers.
1 # Load necessary libraries
2 library (dplyr )
3

4 # Set seed for reproducibility

5 [Link] (123)
6

7 # Create a sample dataset of 50 stores , each with 20 customers

8 store _ids <- 1:50
9 customer _per_ store <- 20
10 data_list <- lapply (store _ids , function (store ) {
11 data. frame (
12 StoreID = store ,
13 CustomerID = paste0 (store , "_", 1: customer _per_ store ),
14 SatisfactionScore = sample (1:10 , customer _per_store , replace = TRUE)
15 )
16 })
17 store _data <- [Link](rbind , data_list)
18

19 # Parameters for cluster sampling

20 clusters <- unique ( store _data$ StoreID ) # List of clusters ( stores )
21 n_ clusters <- 10 # Number of clusters to sample
22

23 # Randomly select clusters

24 selected _ clusters <- sample (clusters , n_ clusters )
25

26 # Perform cluster sampling

27 cluster _ sample <- store _data %>%
28 filter ( StoreID %in% selected _ clusters )
29

30 # Display the sample

31 print ( cluster _ sample )
32

33 # Optional : Summary of the sampled data

34 summary ( cluster _ sample )

Explanation of the R Code

The R code performs the following steps:

1. Loads the ‘dplyr‘ library for data manipulation.
2. Creates a dataset with 50 stores, each having 20 customers with a ‘StoreID‘ and ‘Satis-
factionScore‘.
3. Defines the parameters for cluster sampling: the list of clusters (stores) and the number
of clusters to sample.
4. Randomly selects a specified number of clusters and extracts all customers from these
selected clusters.

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 11 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

5. Displays the sampled data and provides a summary.

2.5. Non Probabilistic Sampling Techniques

To lean how samples can be selected with non-probabilistic approach, undertake the follow-
ing activity

ACTIVITIES

The notes on probabilistic sampling can be accessed in the article Non-probability sampling
go through the article and the video below then attempt the questions below.

Content Curated Video

This video explains the non-probabilistic sampling techniques. go

ç through it. It will give you more insight to non-probability sam-

pling. You will now be able to undertake the quiz with ease.
Link to video

QUIZ/QUESTIONS

1. What is a key characteristic of non-probability sampling?

A. Every unit has an equal chance of being selected.
B. Units are selected using random methods.
C. The probability of selecting any unit is known.
D. Selection of units is subjective and non-random.
Answer: D. Selection of units is subjective and non-random.
2. Which of the following is a major disadvantage of non-probability sampling?
A. It is slow and expensive.
B. It assumes the sample is always representative of the population.
C. It guarantees unbiased results.
D. It always ensures an accurate estimate of sampling variability.
Answer: B. It assumes the sample is always representative of the population.
3. What is a common use of volunteer sampling?
A. For selecting individuals who happen to walk by randomly.
B. For selecting individuals for qualitative testing such as focus groups.
C. For selecting a random sample from the entire population.
D. For selecting individuals based on expert judgment.
Answer: B. For selecting individuals for qualitative testing such as focus groups.

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 12 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

4. Which non-probability sampling method grows its sample by asking respondents to

refer others like themselves?
A. Quota sampling
B. Convenience sampling
C. Snowball sampling
D. Judgment sampling
Answer: C. Snowball sampling.
5. Which of the following is an advantage of non-probability sampling?
A. It reduces the risk of bias in the sample.
B. It allows precise calculation of sampling error.
C. It is quick, inexpensive, and easy to implement.
D. It ensures that every unit in the population has a chance of being included.
Answer: C. It is quick, inexpensive, and easy to implement.

3. Sample size determination

Sample size determination is a crucial step in the design of research studies. It ensures that
the study is adequately powered to detect meaningful effects while maintaining cost efficiency
and feasibility. An improperly determined sample size can lead to inaccurate results, wasted
resources, or ethical concerns. In this lecture, we explore the key considerations in determining
sample size and the formulas used in various types of studies.

3.1. Factors Affecting Sample Size

The sample size required for a study depends on several factors, including:
• Study Objectives: The purpose of the study and the type of research question being ad-
dressed.
• Variability in the Population: Higher variability in the population requires a larger sample
size to achieve the same level of precision.
• Desired Precision: The degree of accuracy required in the estimates. Smaller margins of
error require larger sample sizes.
• Confidence Level: The probability that the true population parameter lies within the confi-
dence interval (typically 95% or 99%).
• Power of the Test: The ability of the study to detect a true effect if it exists. Studies typically
aim for a power of 80% or 90%.
• Effect Size: The minimum detectable difference or effect in the population. A smaller effect
size requires a larger sample to detect.
• Population Size: The total number of individuals in the population. For large populations,
the sample size is often determined independently of the population size.

3.2. Sample Size Determination for Estimating Proportions

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 13 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

3.2.1. Formula for Proportions

To estimate the proportion (p) of a population with a certain characteristic, the sample size
can be calculated using the following formula:

Z2 p(1 − p)
n=
E2
where:
• n is the required sample size,
• Z is the Z-score corresponding to the desired confidence level (e.g., 1.96 for 95% confi-
dence),
• p is the estimated proportion in the population (if unknown, use p = 0.5 to maximize the
sample size),
• E is the margin of error (the desired precision).

ő Example 6:
Suppose we want to estimate the proportion of people in a city who support a new
policy with a 95% confidence level and a margin of error of 5%. If we assume the
proportion p = 0.5, then:

(1.96)2 × 0.5 × (1 − 0.5)

n= = 384.16
0.052
Thus, the required sample size is 385 individuals.

3.2.2. Application: Customer Satisfaction Survey

Scenario

A retail company wants to estimate the proportion of its customers who are satisfied with a
newly launched product. The company aims to make business decisions based on customer
satisfaction levels and plans to conduct a survey among its customers.

Research Objective

To estimate the proportion of satisfied customers with a 95% confidence level and a margin
of error of ±5%.

Steps for Sample Size Determination

1. Define the Population: The population consists of all customers who purchased the
new product in the past 6 months.
2. Select Confidence Level and Margin of Error:
• Confidence Level: 95% (corresponding to Z = 1.96).
• Margin of Error: E = 0.05.

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 14 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

3. Estimate Proportion p: Since no prior data is available, the company assumes p = 0.5,
the worst-case scenario.
4. Apply the Formula: The formula for sample size calculation is:

Z2 p(1 − p)
n=
E2
Plugging in the values:

(1.96)2 × 0.5 × (1 − 0.5) 3.8416 × 0.25

n= = = 384.16
0.052 0.0025
Hence, the required sample size is approximately n = 385.

Conducting the Survey

The company will randomly select 385 customers and ask them whether they are satisfied
with the product (yes/no).

3.3. Sample Size Determination for Estimating Means

3.3.1. Formula for Means

To estimate the population mean (µ), the sample size can be determined using the following
formula:

Z2 σ 2
n=
E2
where:
• n is the required sample size,
• Z is the Z-score corresponding to the desired confidence level,
• σ is the population standard deviation (or an estimate from a pilot study),
• E is the margin of error.

ő Example 7:
Suppose we want to estimate the average height of students in a university with a
standard deviation of 10 cm, a 95% confidence level, and a margin of error of 2 cm.
Then:

(1.96)2 × (10)2
n= = 96.04
22
Thus, the required sample size is 97 students.

3.3.2. Application: Estimating Average Monthly Customer Spending

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 15 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

Research Scenario

A retail business wants to estimate the average amount of money customers spend in their
stores each month. They aim to estimate this mean with 95% confidence and a margin of
error of $10. Based on previous data, the standard deviation of monthly customer spending
is estimated to be $50.

Calculation of Sample Size

We can use the sample size determination formula for estimating the mean:

Z2 σ 2
n=
E2
Substituting the known values:
• Z = 1.96 (for 95% confidence level),
• σ = 50 (estimated standard deviation),
• E = 10 (desired margin of error),
The sample size is:

(1.96)2 × (50)2
n= = 96.04
(10)2

Thus, the company needs to survey at least 97 customers to estimate the average monthly
spending with a 95% confidence level and a margin of error of $10.

3.4. Sample Size for Survey Designs

In surveys, sample size calculation depends on the design of the study. The sample size for
simple random sampling (SRS) can be adjusted based on the survey design by applying a
design effect.

3.4.1. Design Effect

The design effect (D) accounts for the fact that complex sampling designs, such as stratified
or clustered sampling, often increase the required sample size compared to simple random
sampling. The formula becomes:

Z2 p(1 − p)
n=D×
E2
where D is the design effect. A typical value for the design effect is between 1.5 and 2, but it
depends on the specifics of the survey design.

3.4.2. Application: Estimating Customer Satisfaction in a Retail Business

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 16 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

Scenario

A retail business wants to conduct a survey to estimate the proportion of customers who
are satisfied with their in-store experience. They plan to use a stratified sampling design to
account for different customer segments (e.g., by age group or shopping frequency).
They wish to estimate this proportion with a margin of error of 5%, using a 95% confidence
level. Based on previous surveys, the proportion of satisfied customers is estimated to be
p = 0.75. The design effect for the stratified design is estimated at 1.5.

Calculation of Sample Size for SRS

First, we calculate the sample size for a simple random sample (SRS) using the formula for
estimating a proportion:

(1.96)2 × 0.75(1 − 0.75) 3.8416 × 0.1875

n= = = 288.12
(0.05)2 0.0025
So, the sample size required for a simple random sample would be approximately 289 cus-
tomers.

Adjusting for the Design Effect

Given that the design effect for the stratified sampling design is 1.5, the adjusted sample size
is:

nadjusted = 289 × 1.5 = 433.5

Thus, the retail business needs to survey approximately 434 customers to account for the
stratified design and achieve the desired precision.

Other Applications in Business

Sample size determination for surveys can be applied in various other business research sce-
narios, such as:
• Estimating the proportion of customers satisfied with a new product launch,
• Surveying employee satisfaction across different departments in a large corporation,
• Assessing brand awareness among customers segmented by region or demographic.

ONLINE DISCUSSION FORUM AND REFLECTION

After reading the lecture notes participate the following forum discussion. Post your answers
tio the forum then critique your peers work.

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 17 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

Discussion Question

Your company is preparing to launch a new line of eco-friendly household cleaning products.
As part of the product launch strategy, the marketing team wants to conduct a survey to es-
timate the proportion of potential customers who would be willing to purchase the product
within the first six months of release. Based on previous market research, it is estimated that
approximately 40% of the target population would be interested in purchasing eco-friendly
products.
1. Simple Random Sampling (SRS): Calculate the required sample size for this survey
using simple random sampling (SRS), assuming a 95% confidence level and a 5% margin
of error.
2. Adjustment for Design Effect: Suppose the company decides to stratify the sample
based on customer demographics (e.g., age, income, etc.). If the design effect is esti-
mated to be 1.2, how would you adjust your sample size calculation?
3. Discussion: Discuss the potential consequences of underestimating or overestimating
the required sample size in this scenario. How might this affect the business decision-
making process?
4. Application of Concepts: Reflect on how the concepts of sample size determination
and design effect can be applied to other business research areas, such as customer
satisfaction surveys or employee engagement studies. Provide specific examples.

4. Central Limit Theorem and Its Applications

The Central Limit Theorem (CLT) is a key concept in statistics. It states that the sampling
distribution of the sample mean approaches a normal distribution as the sample size increases,
regardless of the population’s distribution, provided the samples are independent.

4.1. Applications of CLT

• Sampling Distributions: The CLT allows us to make inferences about population parame-
ters from sample data.
• Confidence Intervals: The CLT provides the foundation for constructing confidence intervals
around the sample mean.
• Hypothesis Testing: The normal approximation of the CLT enables us to conduct hypothesis
tests and assess statistical significance.
The Central Limit Theorem (CLT) is a fundamental theorem in statistics that describes the dis-
tribution of sample means. It states that the distribution of the sample mean approaches a
normal distribution as the sample size becomes large, regardless of the shape of the popula-
tion distribution.

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 18 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

5. Central Limit Theorem (CLT)

The CLT states that, regardless of the population distribution, the sampling distribution of the
sample mean will tend to be normally distributed if the sample size is sufficiently large. This
is true even if the original population distribution is not normal.

5.1. Theorem Statement

Let X1 , X2 , . . . , Xn be a random sample of size n from a population with mean µ and variance
σ 2 . The Central Limit Theorem states that as n approaches infinity, the distribution of the
σ2
sample mean X̄ approaches a normal distribution with mean µ and variance .
n
Formally,
X̄ − µ d
σ → N (0, 1)
−
√
n
d
where −→ denotes convergence in distribution to a standard normal distribution N (0, 1). A
concept that will help you understand the CLT is sampling distribution which is discussed
below.

5.2. Sampling Distribution of the Sample Mean

When you repeatedly take random samples from a population and calculate the sample mean
for each sample, the distribution of these sample means is called the sampling distribution of
the sample mean. This distribution will vary depending on the sample size and the original
population distribution.

ő Example 8:
We demonstrate the concept of sampling distribution by taking samples from a pop-
ulation and analyzing their distribution. We will plot the population distribution and
the sampling distribution of the sample mean for a given sample size and number of
samples.

Concept of Sampling Distribution

The sampling distribution is the probability distribution of a given statistic (e.g., sam-
ple mean) obtained from a large number of samples drawn from a specific popula-
tion. The Central Limit Theorem states that, as the sample size increases, the sam-
pling distribution of the sample mean will approach a normal distribution, regardless
of the population’s distribution.

Methodology
• Population Size: 50
• Sample Size: 10

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 19 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

• Number of Samples: 5

Samples and Their Means

The samples taken and their means are as follows:

Sample Number Sample Values Mean

Sample 1 5.53, 6.11, 3.62, 6.46, 2.87, 7.23, 4.31, 9, 0.83, 6.62 5.26
Sample 2 3.62, 6.11, 0.65, 4.89, 7.79, 9.73, 4.56, 1.46, 7.32, 7.52 5.36
Sample 3 6.94, 4.86, 0.73, 6.07, 3.12, 3.33, 0.65, 3.5, 6.46, 9.73 4.54
Sample 4 5.73, 0.65, 3.62, 6.94, 6.11, 0.73, 4.86, 2.87, 3.12, 6.07 4.07
Sample 5 0.68, 2.13, 5.73, 7.79, 4.89, 4.31, 6.62, 1.46, 4.56, 4.86 4.30

Samples and Their Means

Population Distribution

The population is generated from a uniform distribution between 0 and 10. The histogram of
the population distribution is shown below.

Population Distribution

Sampling Distribution

We take 5 samples, each of size 10, from the population. The histogram of the sample means
is shown below. It is called sampling Distributrion.
The sampling distribution here does not look like normal distribution. Now lets apply Central
Limit Theorem. i.e. Increasing the sample size.

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 20 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

Sampling Distribution of the Sample Mean

ACTIVITIES

Run the following R script

1 # Load necessary libraries
2 library ( ggplot2 )
3 library ( gridExtra )
4

5 # Set parameters
6 population _size <- 5000
7 sample _size_small <- 5
8 sample _size_large <- 1000
9 num_ samples <- 1000
10

11 # Generate a population from a uniform distribution

12 [Link] (123) # For reproducibility
13 population <- runif ( population _size , min = 0, max = 10)
14

15 # Function to take samples and compute means

16 get_ sample _ means <- function (population , sample _size , num_ samples ) {
17 sample _ means <- numeric (num_ samples )
18 for (i in 1: num_ samples ) {
19 sample <- sample (population , sample _size , replace = FALSE )
20 sample _ means [i] <- mean( sample )
21 }
22 return ( sample _ means )
23 }

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 21 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

24

25 # Generate sample means for small sample size

26 sample _ means _ small <- get_ sample _ means ( population , sample _size_small , num_
samples )
27

28 # Generate sample means for larger sample size

29 sample _ means _ large <- get_ sample _ means ( population , sample _size_large , num_
samples )
30

31 # Plot population distribution

32 pop_hist <- ggplot (data = data. frame (x = population ), aes(x)) +
33 geom_ histogram ( binwidth = 1, fill = " lightblue ", color = "black ") +
34 ggtitle (" Population Distribution ") +
35 xlab(" Value ") + ylab(" Frequency ")
36

37 # Plot sampling distribution for small sample sizes

38 sampling _dist_ small <- ggplot (data = data. frame (mean = sample _ means _ small ),
aes(mean)) +
39 geom_ histogram ( binwidth = 0.2 , fill = " lightcoral ", color = "black ") +
40 ggtitle (" Sampling Distribution of the Sample Mean ( Small Sample Size)") +
41 xlab(" Sample Mean") + ylab(" Frequency ")
42

43 # Plot sampling distribution for larger sample sizes

44 sampling _dist_ large <- ggplot (data = data. frame (mean = get_ sample _means (
population , sample _size_large , num_ samples )), aes(mean)) +
45 geom_ histogram ( binwidth = 0.02 , fill = " lightgreen ", color = " black ") +
46 ggtitle (" Sampling Distribution of the Sample Mean ( Large Sample Size)") +
47 xlab(" Sample Mean") + ylab(" Frequency ")
48

49 # Save plots
50 ggsave (" population _ distribution .png", plot = pop_hist , width = 8, height =
6)
51 ggsave (" sampling _ distribution _ small .png", plot = sampling _dist_small , width
= 8, height = 6)
52 ggsave (" sampling _ distribution _ large .png", plot = sampling _dist_large , width
= 8, height = 6)
53

54 # Print plots to the console

55 grid. arrange (pop_hist , sampling _dist_small , sampling _dist_large , ncol = 1)

below are the results. You can see that we sample size is increase then you end up with tha
normally distributed Sampling distribution.

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 22 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

Population Distribution

Sampling Distribution for small sample

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 23 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

Sampling Distribution large Sample

The Central Limit Theorem simplifies the analysis of sample data by allowing researchers and
analysts to apply normal distribution-based methods and tools, even when the underlying
population distribution is not normal by taking a larger sample. This facilitates better decision-
making, more reliable estimates, and robust statistical testing in various aspects of business
research.

END OF MODULE ASSESSMENT

Attempt all the questions

1. You are tasked with conducting a market research survey to assess customer satisfaction
with a new product. Describe which sampling method you would use to ensure that all
customer segments are adequately represented. Justify your choice of sampling method
and explain how it would help you achieve a representative sample.
2. A company wants to evaluate employee satisfaction across different departments. The
departments vary significantly in size. How would you design a sampling plan to ensure
that the sample is representative of the entire organization? What sampling technique
would you use and why?
3. A business researcher needs to estimate the average annual expenditure on office sup-
plies for a company with a large number of employees. The researcher aims to achieve a
95% confidence level with a margin of error of $500 and knows the standard deviation
of the population is $2000. Calculate the required sample size.
4. In a survey aimed at understanding consumer preferences for a new product, what strate-
gies would you implement to minimize sampling errors and biases? Consider aspects
such as survey design, sampling techniques, and data collection methods.
5. Imagine you are analyzing the distribution of annual sales figures for a company. You

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 24 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

take several random samples of size 15 from the population and calculate the mean for
each sample. Plot the sampling distribution of the sample means and discuss how the
Central Limit Theorem (CLT) applies in this context.
6. You are conducting a business research project to estimate the average time taken by
employees to complete a task. Describe how you would use the Central Limit Theorem to
ensure that your sample estimates are reliable. Include an explanation of how increasing
the sample size impacts the sampling distribution of the sample mean.

CORE READING AND REFERENCES

1. Sampling Techniques : W.G. Cochran, Wiley (Low price edition available)

2. Theory and Methods of Survey Sampling : Parimal Mukhopadhyay, Prentice Hall of India
3. Theory of Sample surveys with applications : P.V. Sukhatme, B.V Sukhatme, S. Sukhatme
and C. Asok, IASRI, Delhi
4. Sampling Methodologies and Applications : P.S.R.S. Rao, Chapman and Hall/ CRC
5. Sampling Theory and Methods : M.N. Murthy, Statistical Publishing Society, Calcutta
(Out of print)
6. Elements of sampling theory and methods : Z. Govindrajalu, Prentice Hall

MCA 807: Statistical Methods

Module 5 End of Module Assessment-Rubric
1. For the market research survey, stratified sampling would be an appropriate method.
This technique involves dividing the population into distinct subgroups (strata) such as
age, income, or geographic location, and then sampling from each stratum. This ensures
that all segments are represented in the final sample, providing a more accurate and
representative assessment of customer satisfaction.
2. For evaluating employee satisfaction across departments of varying sizes, cluster sam-
pling can be effective. First, departments (clusters) are randomly selected, and then em-
ployees within those departments are sampled. This method accounts for the variation
in department sizes and simplifies the sampling process while ensuring each department
is represented.
3. Using the formula for sample size determination:

Z2 · σ 2
n=
E2
where Z = 1.96, σ = 2000, and E = 500:

1.962 · 20002 3.8416 · 4000000 15366400

n= = = ≈ 61.47
5002 250000 250000
Thus, a sample size of approximately 62 is required.
4. To minimize sampling errors and biases:
• Use a random sampling method to ensure every participant has an equal chance of
being selected.

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 25 of 26

LEARNING MODULE 5 MCA 807 - Statistical Methods

• Pretest the survey instrument to ensure clarity and validity.

• Ensure proper training for interviewers to avoid interviewer bias.
• Use stratification to account for key subgroups and ensure their representation.
• Apply data cleaning techniques to identify and address potential errors in responses.
5. To demonstrate the sampling distribution of sample means:
• Generate multiple samples of size 15 from a population (e.g., uniform distribution).
• Calculate the mean for each sample.
• Plot the histogram of sample means.
• The distribution of sample means should approximate a normal distribution as per the
Central Limit Theorem.
6. To use the Central Limit Theorem (CLT):
• Collect multiple random samples of sufficient size from the population.
• Calculate the sample means.
• The distribution of these sample means will approximate a normal distribution, re-
gardless of the original population distribution, given a large enough sample size.
• This allows for making inferences about the population mean with known confidence
intervals and margins of error.

MASTER OF COMMERCE AND BUSINESS ADMINISTRATION Page 26 of 26