Statistics Essentials for Data Science

Sampling and Sampling Techniques

 Learning Objectives

By the end of this lesson, you will be able to:

Understand the concept of sampling along with its advantages and

disadvantages
Identify various sampling techniques

Discover the process employed in non-probability sampling

methods
Differentiate between the situations when probability and non-
probability sampling are applied

Clarify the role of probability distribution in sampling

 Business Scenario

ABC is a government organization that stores and maintains an extensive

dataset on a country's population. The organization is struggling to segment
the data, as it is scattered and does not provide any valuable knowledge.
To address this issue, ABC plans to categorize the data by city and analyze it
to predict population trends in the coming years.
In this endeavor, the organization will investigate various methods including:
• Exploring data sampling
• Implementing probability sampling
• Utilizing systematic sampling
This approach aims to enhance their analytical work and extract meaningful
insights from the dataset.
Introduction to Sampling and Sampling Errors
Discussion
 Discussion: Data Sampling

Duration: 15 minutes
How does sampling of data help in decision-making?
• What does data sampling mean?
• What are the advantages of sampling?
 Sampling

Sampling is the process of selecting a subset of a population to represent the entire population.

In statistics, sampling allows for the testing of a hypothesis about a population's characteristics.
 Sampling

Consider a sample that is adequate to make certain conclusions about the entire group

Representative sample of the Different samples can provide

population varied numerical results
 Sampling

The process of making an inductive inference about an entire population based on a

sample is called statistical inference.

Sample Produced lot

Average life

Example: The average life of the sample of bulbs tested is taken as an estimate
of the average life of the entire produced lot.
 Sampling

While making conclusions, the following terms are important to understand:

Incorporation of the
Sample collection
possibility of errors
 Sampling Error

The error in the inductive inference from a sample to a population is known as sampling error.

Non-sampling errors encompass all errors not classified as sampling errors. Issues in data collection
or biases in responses may induce sampling errors.
 Sampling Error

Consider a population that comprises three units with associated numerical values of 2, 3, and 4
and with 3 as the average value.
Consider various samples of size 2. These are (2, 3), (2, 4), and (3,4). Their
averages are 2.5, 3, and 3.5.

Population Sample of size 2 Averages

2 2 3 2.5

3 Average value = 3 2 4 3

4 3 4 3.5
 Sampling Error

When the sample average is taken as an estimate of the population average, there is an error in
two of the three possible samples.

Sample
averages

2.5

3 Population average = 3

Error in estimate
3.5

Such an error is referred to as a sampling error.

 Data Collection Errors

Example: Data collected on rentals for paying guest accommodation

The population may consist of tenants who are charged separately for electricity and
those who are not charged.

Electricity charges
included

Electricity charges
excluded

When such details are ignored, the results can be misleading.

 Non-Sampling Errors

It is important to note that non-sampling errors occur both in complete enumeration

and in sampling studies.

Complete enumeration Sampling studies

It is important to select a sample that represents the population.

Advantages and Disadvantages of Sampling
 Advantages of Sampling

Sampling saves time, because it reduces the volume of data and minimizes
the need to go over each item.

The sampling process avoids monotony and eliminates repeated inquiries

for each dataset.

Sampling makes surveys feasible when time and resources are limited.
 Limitations of Sampling

A few of the major limitations of sampling are:

01 02

Chances of bias Data excluded from sampling if it is Difficulties in selecting a

not homogeneous sample

This has an impact on the accuracy of the result.

 Discussion: Data Sampling

Duration: 15 minutes
How does sampling of data help in decision-making?
• What does data sampling mean?
Answer: Sampling is the process of selecting a subset of a population to
represent the entire population.

• What are the advantages of sampling?

Answer: It is the only option when destructive testing is involved. It
becomes necessary to restrict sampling to small sizes. The impact is
particularly significant when a large mass of data is involved.
Probability Sampling Methods
Discussion
Discussion: Probability Sampling

Duration: 15 minutes
Assume you have been given a large data set about a country's population,
and you are tasked with dividing the data by city. Your objective is to analyze
this data to predict the country's population in the next five years.

To accomplish this, you should:

• Comprehend probability sampling and its various methods
 Probability Sampling

Probability sampling is the selection of a sample from a population based on the randomization principle.

It ensures that the sample is representative of the population, allows researchers to estimate the level of
uncertainty in their findings, and permits the results to be generalized across the whole population.
 Approaches to Probability Sampling

There are four approaches to probability sampling:

Simple random sampling Systematic sampling

Stratified random sampling Cluster sampling

 Simple Random Sampling

Simple random sampling is a technique of sample selection in which

every sample has an equal probability of selection.
 Simple Random Sampling

Every unit in the sample also has an equal probability of inclusion in the sample.

Population Sample

Equal probability of inclusion

 Simple Random Sampling: Example

Consider a population consisting of four units: A, B, C, and D.

Population Sample of size 3

A A B C

Each sample 1/4

B A B D

Every unit 3/4

C A C D

D B C D

Each sample has a ¼ probability of being selected. Also, every unit has a ¾ probability of being selected.
 Systematic Sampling

Systematic sampling is a probability sampling technique in which researchers randomly

select members of a population at regular intervals.

Systematic
An alternative A convenient
approach sampling
approach

The process involves systematically selecting elements from a subset

that is arranged in sequential order.
 Systematic Sampling: Example

Suppose a population consists of N = 500 units and a sample of n = 50 units

Population N = 500

Sample will be every kth unit

Sample n = 50

k = N/n = 500/50 = 10

Then, every kth unit, where k = N/n = 500/50 = 10 in the population is chosen.
 Systematic Sampling: Example

The first unit is randomly selected from units 1 to k = 10.

1 to k = 10

1 2 3 4 5 6 7 8 9 10

k is the sampling interval

Every unit in the population has the

probability of 1/k = 0.1 of being chosen.
 Systematic Sampling: Example

In a sequential arrangement of units, systematic sampling maintains both efficiency

and the original order of the units.

Samples are equidistant.

1 2 3 4 5 6 7 8 9 10

The distance is being chosen logically.

 Stratified Random Sampling

Stratified random sampling is a sampling method that divides a population into smaller
subgroups known as strata.

Population (heterogeneous) Strata (homogeneous)

 Stratified Random Sampling

A random sample is drawn from each stratum in a number proportional to the size
of the stratum in comparison to the population.

Strata Unit

Simple random
sampling
of each stratum
 Stratified Random Sampling

In comparison to a sample drawn at random, the chosen sample includes members of all strata,
making it a more accurate representation of the population.

Sample Population

Better representation
 Stratified Random Sampling: Example

An employee satisfaction survey is being conducted in a large company.

The employees’ opinions could vary at different levels in the hierarchy and in different functions, like:

Marketing Sales

Human resources Accounting

 Cluster Sampling

Cluster sampling is a method in which the population is divided into several clusters. All
clusters are almost equally heterogeneous.

Population Clusters
 Cluster Sampling

A sample of clusters is selected using simple or stratified random sampling.

Clusters Sample

Simple random
sampling or
stratified random
sampling

All units from these clusters together constitute the required sample.
 Cluster Sampling: Example

Expenditure patterns of residents in a city:

Population clusters Representation of the clusters

 Uses of Probability Sampling

Some of the key uses of probability sampling include:

Making informed decisions

Incorporating required Establishing rules for

quality standards rejecting items of poor
quality
Discussion: Probability Sampling

Duration: 15 minutes
Consider that you have been given a large set of data about a country's
population and have been asked to bifurcate the data based on each city and
analyze it to predict the country's population over the next five years.

To do so, perform the following:

• Understand the simple random sampling of data
Answer: Simple random sampling is a technique of sample selection in
which every sample has an equal probability of selection.
• Understand probability sampling and systematic sampling
Answer: Probability or random sampling is a sampling approach to select a
sample using the theory of probability. In systematic sampling, samples are
selected from a sequentially arranged lot, at regular intervals.
Non-Probability Sampling Methods
Discussion
Discussion: Non-Probability Sampling

Duration: 15 minutes
• What is non-probability sampling?
• What is convenience sampling?
 Non-Probability Sampling

It is a method of sample selection in which a sample is chosen based on the investigator’s

subjective evaluation.
 Non-Probability Sampling Methods

The four methods of non-probability sampling are:

Convenience sampling Voluntary response sampling

Snowball sampling Judgmental or purposive sampling

 Convenience Sampling

It is a type of non-probability sampling in which the researcher selects the sample based
on convenience rather than random selection.
 Convenience Sampling: Example

Consider the example of a supermarket chain that has prepared a

questionnaire to initiate steps toward improving customer satisfaction.

Questionnaires were handed to

To expedite the
customers who visited the market
process
in the next couple of days.

So here, the sample of customers chosen constitutes a convenience sample.

 Voluntary Response Sampling

It is a type of non-probability sampling in which the researcher selects the sample

from people who volunteer to participate in the study.
 Voluntary Response Sampling: Example

Consider the following example:

• Websites invite individuals to offer responses

• Only a few individuals provide responses

This process of obtaining responses from selected respondents is clearly voluntary response sampling.
 Snowball Sampling

Snowball sampling is a research method in which researchers initially recruit participants and then rely
on referrals to recruit more, thus creating a chain.

The process is continued till a sample of the desired size is obtained.

 Snowball Sampling: Example

Snowball sampling is popular in business studies that focus on specific organizations.

Some of the employees serve as contacts who provide referrals for further contacts
besides serving as respondents for the study.
 Purposive or Judgmental Sampling

It is an approach in which the investigator selects units from the population using his knowledge and
judgment about various units in the population.
 Purposive Sampling: Example

In a qualitative research study, an investigator may target customers who are articulate.

Customer feedback

The investigator could then obtain useful and high-quality information from this sample.
 Uses of Non-Probability Sampling

It provides for a quicker and more cost-effective data collection.

Convenience Willingness to respond

It is
owing to
factors
such as:

Ease of identification

The approach is attractive when time and cost are important considerations.
 Non-Probability Sampling

The approach offers greater scope to collect quality data with factors such as:

Articulation

Willingness

Results from the sample cannot be generalized to the entire population, as the
sample does not necessarily represent the population.
Discussion: Non-Probability Sampling

Duration: 15 minutes
• What is non-probability sampling?
Answer: It is an approach to sample selection wherein a sample is selected
based on the subjective judgment of the investigator.

• What is convenience sampling?

Answer: It is an approach to sample selection in which sample units are
chosen entirely on considerations of convenience.
Sampling
 Sample Indices

The sample indices are usually referred to as statistics, and the population indices are referred to as
population parameters.

Sample Population

Estimate

Statistics Population parameters

Values of sample statistics clearly vary depending upon the sample chosen.
 Sampling: Example

Consider a population comprising four units, with the following values:

Unit ref no. I II III IV

Value 2 3 3 5
 Sampling: Example

The quality of results from the analysis will be poor when several clusters do not adequately represent
the population.

Sample no. Units selected Sample values Sample average

1 I, II 2, 3 2.5

2 I, III 2, 3 2.5

3 I, IV 2, 5 3.5

4 II, III 3, 3 3

5 II, IV 3, 5 4

6 III, IV 3, 5 4
 Sampling: Example

The probability distribution of the sample average is:

Sample
2.5 3 3.5 4 TOTAL
average

Probability 1/3 1/6 1/6 1/3 1

Twice as likely
Sampling Distribution
 Sampling Distribution

The sampling distribution represents the distribution of a statistic, like mean, proportion, and
variance, calculated from multiple random samples taken from the same population.

It provides insights into the characteristics and variability of a statistic across different samples,
allowing for population inferences based on sample statistics.
 Sampling Distribution

The key characteristics of the sampling distribution of the sample mean are:

The distribution of the sample mean follows a

normal distribution with the same mean μ
regardless of the sample size.

However, the standard deviation is inversely

proportional to the square root of the sample
size. Consequently, as the sample size increases,
the standard deviation decreases.
Central Limit Theorem (CLT)
 Central Limit Theorem

The central limit theorem (CLT) states that independent random variables approach a normal
distribution as the sample size increases, regardless of the underlying population distribution. This
applies to both the sum and average of the random variables.

If a random sample of size n is drawn from a normal population with

parameters μ and σ, denoted as X1, X2, ..., Xk, ..., Xn, the sample mean (𝑥 ̅)
follows a normal distribution with parameters μ and σ divided by the square
root of n.
Case Study: Sample and Sampling Techniques
Problem Statement

Let's consider a scenario where a

significant number of students take exams
every year, and the percentage of marks
got in mathematics follows a normal
distribution. The distribution has a mean
of 60 and a standard deviation of 15.

Our objective is to determine the

probability of a student scoring 70% or
above in the examination.
 Problem Statement

Number of samples = 12 Sample size = 16

Determine the sampling distribution of the sample mean

Estimate the proportion of students scoring more than 70%

Calculate the sample mean (x̄) for each of the 12 samples.

 Data Samples

The data comprises 12 samples, with each sample having a size of n = 16.

I II III IV V VI VII VIII IX X XI XII

1 52 87 67 58 54 60 73 62 66 59 50 64
2 57 61 50 61 30 56 64 37 78 62 87 103
3 65 68 90 78 36 73 60 31 83 48 42 40
4 78 68 68 68 78 36 54 84 56 35 47 64
5 96 75 58 64 18 60 66 63 57 80 52 56
6 67 80 66 63 95 67 74 41 45 60 57 39
7 49 46 57 59 87 62 50 66 80 79 72 56
8 39 42 50 48 59 47 69 46 44 48 46 44
 Data Samples

The data comprises 12 samples, with each sample having a size of n = 16.

I II III IV V VI VII VIII IX X XI XII

9 79 68 83 58 40 77 59 59 60 69 50 52
10 80 47 35 69 44 53 66 62 66 49 51 78
11 43 67 74 57 64 54 75 58 31 66 41 51
12 46 63 68 62 42 58 62 51 74 59 70 75
13 77 79 48 62 64 53 73 46 77 74 52 68
14 58 62 56 64 104 76 75 39 55 70 45 71
15 101 37 85 29 73 38 68 68 49 52 50 76
16 43 38 59 50 38 62 73 75 43 78 67 85
 Presentation of Results

Record the results from sample data in the following format:

Sample No. I II III III ………………………………… X XI XII

Value of x̄

No. of observations ≥ 70

Proportion of
observations ≥70
 Presentation of Results

Record the consolidated results in the following format:

Standard deviation Average value of Standard deviation

Average value of
of sample mean sample proportion of sample proportion
sample mean (n=16)
(n=16) (n=16) (n=16)

Theoretical values

Estimated values
from sample data

State the observations

 Solution

Calculate

Mean If p = average proportion

Standard deviation =
Standard deviation of x̄ values
√ ((p* (1-p)/n)
Mean of the proportions

Standard deviation of proportions

 Solution

Sampling distribution of the sample mean:

Mean = 60

Standard deviation = 15/ √16 = 15/4 =3.75 The proportion of students scoring > 70 = 0.2525
 Solution

Sampling distribution of proportion:

Mean = 0.2525

Standard deviation = √{0.2525*(1- 0.2525)/16} = 0.1086

According to the central limit theorem, the provided approximation holds true when
the sample size is 30 or larger.
 Solution

The probability of a student scoring above 70 is as follows:

Score > 70%

0.2525
 Solution

The sample proportion follows a normal distribution.

Sample size = 16 Mean = 0.2525

Standard deviation Number of samples > 70%

= √{0.2525*(1- 0.2525)/16} Proportion =
Total number of samples
= 0.1086
 Solution

Example: Set 1 with 6 samples scoring above 70.

78 96 79 80 77 101

6
Proportion = = 0.375
16
 Results

The results from the sample data are as follows:

Std.
Sample no. 1 2 3 4 5 6 7 8 9 10 11 12 Avg.
Dev.
3.590
Mean 65 62 63 59 58 58 66 55 60 62 55 64 61
815

No. ≥ 70 6 4 4 1 5 3 6 2 5 4 3 6

Proportion 0.255 0.108

0.375 0.25 0.25 0.063 0.313 0.188 0.375 0.125 0.313 0.25 0.188 0.375
≥ 70 2 995
 Inference

Estimated values of parameters have practical applications such as:

Making predictions Making inferences

Example: Determining the Example: Evaluating whether the class

percentage of students who pass average exceeds a specified limit
 Key Takeaways

Sampling is a statistical technique that involves performing a

predetermined number of observations on a population.

Known sampling methods can be used to draw conclusions about

an entire population from a representative sample.

In simple random sampling, every sample of a given size has an

equal probability of being selected.

Non-probability sampling involves selecting a sample based on the

subjective judgment of the investigator.
 Key Takeaways

The four methods of non-probability sampling include convenience,

voluntary response, judgmental or purposeful, and snowball
sampling.
The central limit theorem is applicable in situations involving other
distributions.
Knowledge Check
Knowledge
Check What is the process of making an inductive inference about an entire population based
1 on a sample?

A. Sampling

B. Sampling error

C. Biases

D. None of the above

 Knowledge
Check What is the process of making an inductive inference about an entire population based
on a sample?

A. Sampling

B. Sampling error

C. Biases

D. None of the above

The correct answer is A

Sampling is the process of making an inductive inference about an entire population based on a
sample.
Knowledge
Check What is a referral approach to sampling in which individuals provide referrals for
2 further contact?

A. Non-probability sampling

B. Purposive sampling

C. Judgmental sampling

D. Snowball sampling
 Knowledge
Check What is a referral approach to sampling in which individuals provide referrals for
2 further contact?

A. Non-probability sampling

B. Purposive sampling

C. Judgmental sampling

D. Snowball sampling

The correct answer is D

Snowball sampling is a referral approach to sampling in which individuals provide referrals for
further contact.
Knowledge
Check What is an approach in which the investigator selects units from the population using
3 their knowledge?

A. Non-probability sampling

B. Purposive sampling

C. Snowball sampling

D. None of the above

 Knowledge
Check What is an approach in which the investigator selects units from the population using
3 their knowledge?

A. Non-probability sampling

B. Purposive sampling

C. Snowball sampling

D. None of the above

The correct answer is B

Purposive sampling is an approach in which the investigator selects units from the population using
their knowledge.
Thank You