Sampling

Universe
The term "universe" typically refers to the total aggregation of all the items or
units of interest. It's the entire set of possible observations or units that one
could potentially study. It can be finite or infinite

Population:
•A population is the complete set of all items that share a common set of
characteristics. In statistics, we often want to study a specific characteristic of a
population, but it might be impractical or impossible to study the entire
population directly. So, we take a sample. While "universe" and "population" can
sometimes be used interchangeably, "population" is more commonly used in
statistics to denote the entire group from which a sample is drawn.
Research Objective: To determine the average monthly expenditure on online
shopping among university students in the Jalandhar City during 2022.
1.Universe:
1.Definition: All individuals who have the potential to shop online.
2.Contextual Example: This could include individuals of all age groups from
various parts of the world, regardless of whether they've ever attended a
university or live in Jalandhar. It's every person who has ever bought
something online or has the potential to do so.
2.Population:
1.Definition: The specific group about which the researcher wishes to draw
conclusions based on the research objective.
2.Contextual Example: In this case, the population would be all university
students in the Jalandhar City , as they are the specific group whose online
shopping expenditure we want to determine.
Finite and Infinite populations
•Finite Population:
•This refers to a population that has a definite, countable number of
members or units. You can, in theory, enumerate every member of this
population.
•For example, if you wanted to survey the satisfaction level of
employees in a specific company with 500 employees, then the
population size (N) is 500, which is finite.
•When sampling from a finite population, the size of the population
can influence the sampling design and calculations of the sampling
error.
Infinite Population:
•This refers to a population that is indefinitely large and cannot be
feasibly counted. It doesn't literally mean "infinite" in the
mathematical sense, but rather it's so large that for all practical
intents and purposes, we treat it as infinite.

•An example might be if you wanted to survey the satisfaction level of

all passengers ever to fly on any airline. Given that this number is
constantly changing and practically uncountable, we would treat it as
an infinite population.
Sampling Frame

• A sampling frame is essentially a list of all the items or individuals

within the target population from which a sample can be drawn.

• In other words, it's the source material or device from which a

researcher draws the sample. If the sampling frame is accurate and
comprehensive, it closely represents the target population.

• Or Unique address of every respondents in population

Example of a Sampling Frame: Imagine you want to conduct a
survey to determine the favorite fruit of all the students in a
particular school.

1.Target Population: All students in the school.

2.Sampling Frame: A list of names of all students enrolled in
the school for the current academic year.
The sampling unit
• It is the individual element or set of elements considered for selection in
some stage of sampling. In simpler terms, it's the basic unit that is selected
from the population and into the sample during the sampling process.

Let's delve into some examples to clarify the concept of a sampling unit

1.School Survey on Dietary Habits:

1.Target Population: All students in a high school.
2.Sampling Frame: A list of student ID numbers in the school.
3.Sampling Unit: Each individual student selected to participate in the survey.
Market Research on Brand Preferences:
1.Target Population: All adult consumers in a city.
2.Sampling Frame: A list of residents from the city's registry.
3.Sampling Unit: Each individual resident chosen to answer
questions about their brand preferences.
Types of Sampling Techniques

• Probability sampling

• Non Probability Sampling

Probability sampling

It refers to sampling techniques in which every member of the

target population has a known, non-zero chance of being selected in
the sample. Or the selection of a sample from a population, when
this selection is based on the principle of randomization, that is,
random selection or chance.
Types of Probability Sampling Method
1. Simple Random Sampling (SRS)
2. Systematic Sampling
3. Stratified Sampling
4. Cluster Sampling
5. Multistage Sampling
Simple Random Sampling (SRS)

• Simple Random Sampling (SRS) is a probability sampling technique wherein

each member of the population has an equal chance of being included in the
sample.
• It's akin to a random lottery method where individual units are chosen
entirely by chance, and every subset of a given size has the same probability
of being the chosen sample.
Examples of Simple Random Sampling (SRS):
1.School Survey: Suppose you want to survey the opinions of students in a school
about a new cafeteria menu. If there are 1,000 students in the school and you
want to sample 100 of them:
1.Assign each student a unique number from 1 to 1,000.
2.Use a random number generator or a physical method like drawing numbers
from a hat to pick 100 distinct numbers.
3.Survey the students corresponding to those 100 numbers.
Agricultural Study: Let's say you're an agricultural scientist wanting to study the
yield of a particular crop variety in a large farm field. If the field is divided into
1,000 plots:
•Assign a unique number to each of the 1,000 plots.
•Randomly select 50 of those numbers using a random number generator or
another method.
•Collect and assess the yield only from those 50 plots.
1.Quality Control in Production: A factory produces thousands of light bulbs
daily. To ensure quality:
1.Every light bulb produced in a day is assigned a unique serial number.
2.A random sample of 100 light bulbs is selected using their serial numbers to
test for defects.

2.Health Survey: If a researcher wants to know about the general health habits of
residents in a city of 10,000 households:
1.Each household could be given a unique identifier.
2.A random sample of 500 households might be chosen using a random
method.
3.Members of those households would then be surveyed about their health
habits.
Advantages of Simple Random Sampling:

1.Unbiased Representation: Since every member has an equal chance of being

selected, the sample is more likely to be representative of the entire population.
This helps in reducing the chances of selection bias.

2.Ease of Understanding: The concept behind SRS is straightforward, making it

easier to explain to others. Its simplicity is often seen as one of its greatest
strengths.

3.Analytical Simplicity: Statistical analysis with SRS is often simpler and more
direct. Most statistical methods assume random sampling, so SRS can often be
analyzed with fewer complications or adjustments.
Disadvantages of Simple Random Sampling:
1.Not Feasible for Large Populations: For very large populations, obtaining a
list of the entire population might be impractical, making SRS difficult to
implement.

Costly and Time-Consuming: If the population is widespread geographically, it

might be expensive and time-consuming to collect data using SRS. For instance,
if you're randomly selecting households from an entire country, your selected
households might be very far apart.
Systematic Sampling
Systematic sampling is a probability sampling technique where elements are
selected from an ordered sampling frame. In systematic sampling, every kth item
is selected to produce a sample, where the value of K (the sampling interval) is
determined by dividing the total population size by the desired sample size. The
starting point is usually chosen at random from the first k items on the list.
Research Objective:
"To understand the buying habits of customers in a large supermarket chain over a month."
Given the large number of customers who visit a supermarket daily, it would be impractical to
survey every customer. So, a sampling method is needed, and in this case, we will use
systematic sampling.
Random Starting Point:
•Instead of starting with the first customer, pick a random number between 1
and 30 as a starting point. Let's say the random number chosen is 7. So, the
seventh customer on the first day will be the first person you survey.
Selecting the Sample:
•After the seventh customer, every 30th customer will be approached for the
survey. This means the 7th, 37th, 67th customer, and so on, will be included until
you reach the desired sample size of 500 customers over the month.
Conduct the Survey:
•As the chosen customers (based on the systematic sampling method) come in,
you will conduct the survey. It might be a questionnaire asking about their buying
habits, preferred products, how often they shop, and any other relevant
questions that pertain to the research objective.
Analysis:

•After collecting data from the 500 customers, the responses are analyzed to
derive insights about the buying habits of customers in the supermarket chain
over the month.
Advantages of Systematic Sampling

Simplicity: Once the sampling interval (k) is determined, the selection process is
straightforward. There's no need for random number generators after the initial
starting point is chosen.

Easily Implemented: Especially in situations where a process or flow is ongoing,

like on a production line or during customer arrivals, it’s easy to apply
systematic sampling without disrupting the process.
Disadvantages of Systematic Sampling:

1.Not Truly Random: Because the selection is determined by an interval, it's not
completely random after the initial selection, which can introduce some biases.

1.Requires a List: A complete list of the population is needed, similar to simple

random sampling. This can be problematic if the list is hard to obtain.
Stratified Sampling

• It is a method where you divide the entire population into smaller groups, or
"strata", based on specific shared characteristics. After dividing the
population, you then draw random samples from each of these groups.

• The key idea is to ensure that all types or categories of individuals are
adequately represented in the study.
Research Objective:
"To analyze the satisfaction levels of patients at a multi-specialty hospital with
respect to the services offered."
Given a hospital houses various departments such as cardiology, neurology,
orthopedics, etc., the satisfaction levels may vary across these departments due
to the different natures of services provided. Thus, to get a comprehensive view,
we can use stratified sampling.
Process:
1.Identify the Strata:
1.In this case, the strata can be the various departments in the hospital:
Cardiology, Neurology, Orthopedics, Pediatrics, etc. The reason for this
division is to make sure that the satisfaction levels in each department are
adequately represented in the study.
• Determining Sample Size:
1.Suppose you decide to survey 500 patients in total. You would then decide
how many patients to survey in each department. This could be an equal
number from each department or proportional to the size of the
departments
• Random Sampling Within Each Stratum:
•Within each department, randomly select patients to be surveyed. For
example, if you're surveying 100 patients in the cardiology department, you
would randomly select 100 patients from that department's patient list.
• Conducting the Survey:
•A survey would be conducted among the selected patients, with questions
focusing on various aspects of service quality and satisfaction levels - like
doctor's expertise, nursing care, infrastructure, responsiveness, etc.
• Data Analysis:
•After collecting the data, you would analyze the results separately for each
stratum (department), and then combine them for an overall analysis. This
way, you can identify both the overall satisfaction level and the specific
satisfaction levels in each department.
Customer Satisfaction Survey for a Multinational Company:

•A company operates in three regions: Asia, Europe, and North America. They
want feedback on a new product.

•They could divide their customer base into three strata based on these regions.
From each region (stratum), they randomly select a certain number of
customers to survey.
Healthcare Study on a Medication's Effect:

•Suppose a researcher wants to study the effects of a medication across

different age groups. They could divide the population into strata like: 18-30, 31-
50, 51-70, and 70+ years.

•Random samples from each age group (stratum) would then be selected to
participate in the study.
Advantages of Stratified Sampling

1.Improved Accuracy: By ensuring representation from all strata (subgroups),

stratified sampling can provide more accurate estimates than a simple random
sample, especially when the strata are significantly different from each other
regarding the variable of interest.

2. Detailed Analysis: Allows for analysis within individual strata, which can reveal
insights that might be missed in a more general analysis.
Disadvantages of Stratified Sampling

Requires Knowledge of Population: The researcher needs to have knowledge

about the population structure to create strata. This might not always be
available or clear.

Cost and Time: In certain scenarios, especially when there are many strata,
stratified sampling can be more time-consuming and costly compared to other
methods.
Cluster Sampling

• Cluster Sampling is a sampling technique where the entire population is

divided into groups, or clusters, and a random sample of these clusters is
selected. All or a random sample of the members within the chosen clusters
are then surveyed.

• It's different from stratified sampling, where certain members from each
subgroup are chosen. In cluster sampling, entire clusters are chosen.
How Cluster Sampling Works:

1.Divide Population into Clusters: The first step is to divide the entire
population into clusters. These clusters can be naturally occurring or artificially
created.

2.Randomly Select Clusters: Choose a certain number of clusters randomly.

3. Choose Members from Selected Clusters: Survey all members from the
selected clusters or take a random sample from them.
Research Objective:
"To assess the educational outcomes of 8th-grade students in a state."
Given that the state may have hundreds or thousands of schools, surveying every
student or even conducting simple random sampling across all schools could be
logistically challenging and costly. Cluster sampling can be a more feasible
alternative here.

Process:

1.Identify the Clusters:

1.The first step is to identify the clusters for the study. In this case, entire
schools can serve as clusters.
Randomly Select Clusters:
•After listing all the schools in the state, a certain number of schools are
randomly selected to participate in the study. For instance, if there are 1000
schools, we might randomly select 50 schools.

Selecting the Participants:

•Instead of selecting a few students from each school, we survey all 8th-grade
students in each of the 50 selected schools.

Conducting the Survey:

•A survey is administered to all 8th-grade students in the selected schools to
assess various educational outcomes, such as performance in different subjects,
satisfaction with teaching methods, etc.
Data Analysis:
•Once the data is collected, it's analyzed to infer the educational outcomes for
the entire state. You could also analyze the data by individual school or by
groups of schools (e.g., urban vs. rural) to get more detailed insights.
Examples:

1.Surveying Households in a Large City:

1. Step 1: Divide the city into blocks or neighborhoods (clusters).
2. Step 2: Randomly select a few blocks or neighborhoods.
3. Step 3: Survey every household in these selected blocks or randomly select a few
households from them.

2.Research on School Students Across a State:

1. Step 1: Consider each school in the state as a cluster.
2. Step 2: Randomly select a certain number of schools.
3. Step 3: Survey all students in the selected schools or randomly choose a number of
students from each school.
Advantages of Cluster Sampling
Cost-Effective: Often more economical than other methods, especially for
large-scale studies. For example, if you're surveying households in a country, it's
cheaper to pick a few towns and survey all households in those towns rather
than traveling to towns all over the country for a few households in each.

Practicality: In some situations, it's almost impossible to compile a list of the

entire population, but it's feasible to create a list of clusters (e.g., a list of
villages in a region)

Disadvantages of Cluster Sampling

Not as Precise: If individual units within clusters are similar to each other, then
taking all units from a few clusters might not capture the full diversity of the
population.
Multistage Sampling

Multistage Sampling is a complex form of probability sampling. In this method,

the total population is divided into multiple stages to make the sampling
process more manageable. At each stage, a sampling method (like random
sampling) is applied. This is often used when the population is large,
widespread, and not easily accessible in its entirety.
Research Objective:
"To understand the nutritional habits of high school students across an entire
country."
Given the vast number and geographical spread of high schools in a country,
surveying every student directly would be extremely challenging both logistically
and financially. Multistage sampling offers a method to approach this large-scale
objective in stages, each becoming more focused.

Process:
1.First Stage - Select Provinces or States:
1.In the first stage, instead of surveying schools across the entire country
directly, you might start by randomly selecting a handful of provinces or
states within the country. Suppose the country has 50 states, and you choose
to focus on 5 of them.
Second Stage - Select Districts:
1.In each of the chosen states, you'd then further narrow down your study by
selecting a certain number of districts. For instance, if each state has an
average of 20 districts, you might randomly select 4 districts from each
state.
Third Stage - Select Schools:
1.From each of the chosen districts, you now randomly select specific high
schools. Say you decide to study 3 high schools from each district.

Fourth Stage - Select Students:

•Finally, within each of the selected schools, you then conduct a random
sampling of students to respond to the survey. Maybe you choose to survey 50
students from each school.
• Conducting the Survey:
1.With the students identified, a detailed survey can be administered to
understand their nutritional habits, preferences, challenges, and more.

• Data Analysis:
1.After data collection, the results from different stages can be weighted and
analyzed to infer the nutritional habits of high school students across the
entire country.
Advantages:
1.Cost-Effective: Conducting surveys or studies on a large scale can be quite
expensive. Multistage sampling can help to reduce costs as you're not surveying
the entire population, especially in the initial stages.

2. Feasibility: In many cases, surveying the entire population might be practically

impossible due to resource constraints. Multistage sampling allows researchers
to obtain a representative sample in a more feasible manner.

Disadvantages
1. Complexity: Multistage sampling can be complex to design and implement,
requiring careful planning at each stage to ensure that the sample is
representative.
2. Analysis Challenges: Analyzing data from multistage sampling can be more
complex compared to simpler sampling methods. The researcher needs to
account for the different stages and potentially complex survey weights in the
analysis
Non-probability sampling
• It is defined as a sampling technique in which the researcher
selects samples based on the subjective judgment of the
researcher rather than random selection.
• It is a sampling technique in which not all members of the
population have a known or equal chance of being selected in the
sample
Example
In a study aimed at understanding the challenges of start-up
entrepreneurs, a researcher might purposefully select and interview
founders who have been in business for less than two years, ignoring
those who have been operating longer.
Types of non probability sampling
1. Convenience Sampling
2. Judgmental or Purposive Sampling
3. Snowball Sampling
4. Quota Sampling
5. Voluntary Response Sampling
6. Haphazard Sampling
Convenience Sampling
• It is a type of non-probability sampling method where data is
collected from an easily accessible and available group or population.
Instead of relying on random selection, the researcher uses subjects
that are closest and most convenient to them.

• While this method is less expensive and more straightforward, it has

inherent biases and may not be representative of the overall
population.

• Participants are selected based on their easy accessibility and

proximity to the researcher.
•Example: A researcher conducting a quick survey on caffeine
consumption might decide to survey people in the nearest coffee shop
simply because it's convenient
Research Objective Example:
• Research Objective: To understand the preferences of college
students regarding online learning platforms.
• In a convenience sampling approach, a researcher might select
participants as follows:

• Method: The researcher, being a lecturer at a university, decides to

survey their own students after a lecture.

• The researcher prepares a questionnaire about online learning

platforms and distributes it to the students present that day,
collecting their responses.
• In this case, the sampled group (the lecturer's students) is
convenient for the researcher to access.

• The students are readily available, easy to approach, and

likely willing to help the lecturer by filling out the
questionnaire.
However, the convenience sample has limitations:

1.Bias: The opinions of these students might not represent all college
students' views, even within the same university. They might share
similar experiences and backgrounds, especially being in the same
class, and may not offer a diverse range of opinions on the topic.

2. Generalizability: The results obtained might not be generalizable to

the broader population of college students since they only represent a
specific group from one class or lecture.
Judgmental or Purposive Sampling
• It is a non-probability sampling method where the researcher
selects specific individuals or groups for participation in the study
based on their knowledge or judgment about the population.

• These decisions are made because the researcher believes that

these particular participants will be most representative or
informative for the study's purpose.
• The researcher uses their judgment to select participants based on
specific criteria or knowledge about the population.
Example: A researcher studying advanced meditation techniques
might only interview monks who have practiced meditation for over
10 years, based on the judgment that they possess the necessary
expertise.
Research Objective Example:
Research Objective: To understand the challenges faced by start-up
founders in the e-commerce sector during their first year of operations.

Method: The researcher identifies and selects e-commerce start-up

founders who have completed just one year in business. The researcher
might choose founders based on several criteria, like:
•Start-ups that have shown rapid growth in the first year.
•Start-ups that faced significant challenges but managed to navigate them.
•Start-ups located in a specific region or country to keep the context
consistent.
• The researcher believes that by selecting these specific founders, they
can gain in-depth insights into the unique challenges and successes of
the e-commerce sector during the critical first year.
The advantages and challenges of this approach are:

1.In-depth Insights: By selecting specific individuals or groups, the

researcher can get detailed and nuanced insights into the topic of interest.
2.Criteria-based Selection: This method allows the researcher to target very
specific characteristics or experiences that are relevant to the research
objective.
However, there are limitations:
1.Bias: The researcher's biases can influence the selection, which might lead
to overlooking certain perspectives.
2.Generalizability: The results might not be generalizable to all e-commerce
start-ups, as the study focuses on a specific group selected based on certain
criteria.
Snowball Sampling
• It is a non-probability sampling method used primarily when the
target population is hard-to-reach or hidden. In this method, initial
participants (or "seeds") are selected to take part in the study.

• Once they've provided their data, they then refer the researcher to
other potential participants from their network or community, who
in turn can refer the researcher to more participants, and so on.

• This process creates a "snowball" effect, where the sample grows

in size as more and more referrals are obtained.
Example: In a study on a specific rare illness, a patient participating in
the study might refer the researcher to other patients they know who
also suffer from the same condition.
Research Objective Example:
• To understand the experiences and challenges faced by undocumented
immigrants in a city.
• Given the sensitive nature of the topic and the potential risks involved for
the participants, many undocumented immigrants might be reluctant to
participate in a study due to fear of exposure or legal repercussions.
Method:
1.The researcher starts with a few known contacts within the undocumented
immigrant community, perhaps individuals they've worked with in a supportive
capacity or who have been referred by local non-profits or advocates.
2.After interviewing these initial participants, the researcher asks them to refer
other undocumented immigrants they know who might be willing to
participate.
3.The new referrals then participate and provide more referrals, causing the
sample size to grow like a snowball rolling downhill.
Advantages and challenges of this approach:

1.Access to Hidden Populations: Snowball sampling can provide

access to groups that are otherwise difficult to study due to their
hidden or closed nature.

2.Trust Building: The referrals often come from trusted sources within
the community, which can help in building rapport with new
participants.
Quota Sampling
• It is a non-probability sampling method where the researcher aims to
gather a sample that represents certain characteristics of the
population.

• The population is divided into subgroups (also known as strata), and

then participants are non-randomly selected from each subgroup until a
predetermined quota for each subgroup is met. The quotas are typically
set to reflect the distribution of the subgroups in the larger population.

• Example: A marketer might want to gather feedback on a new

product and decides to survey 50 males and 50 females. They might
then approach people in a mall, ensuring they survey the first 50
men and 50 women who agree, without random selection.
Research Objective Example:
Research Objective: To assess the opinions of urban residents on public
transportation improvements in a city.
Suppose the city's demographics are known to be:
•60% adults under 30 years old
•25% adults aged 30-50
•15% adults over 50 years old
Method:
1.The researcher decides to collect opinions from 1000 residents.
2.Based on the city's demographics, the researcher sets quotas to ensure
the sample reflects the population's age distribution:
1.600 residents under 30 years old
2.250 residents aged 30-50
3.150 residents over 50 years old
3.The researcher then goes to various urban locations (e.g., malls, parks,
bus stops) and surveys residents. Once the age-based quota for a particular
group is filled, no more individuals from that age group are surveyed.
Advantages and challenges of this approach:

1.Representation of Subgroups: Quota sampling ensures that specific

subgroups are adequately represented in the sample.

2.Flexibility: Researchers can choose from where and how they will
meet their quotas, giving them flexibility in data collection.
Voluntary Response Sampling
• It is a non-probability sampling method where participants self-
select or volunteer to be part of the study, often in response to a
general appeal or open invitation.

• This method is commonly seen in public polls, call-in radio shows, or

online surveys promoted on websites or social media. Since
participation is voluntary, it's often influenced by strong feelings or
opinions about the topic, leading to potential bias in the sample.

• Example: An online poll posted on a news website where readers

can choose to answer or not. The results are based only on those
who felt compelled to respond.
Research Objective Example:
Research Objective: To gauge public opinion on the introduction of a new
citywide recycling policy.
In a voluntary response sampling approach, the city council might proceed as
follows:
Method:
1.The city council launches an online survey on their official website, asking
residents to give feedback on the proposed recycling policy.
2.They promote this survey through local newspapers, radio spots, and social
media channels, inviting all city residents to participate.
3.Over the course of a month, residents who feel strongly about the policy
(either in favor or against) or those who are more engaged in civic activities
might be more inclined to fill out the survey.
Haphazard Sampling (a variant of convenience sampling)
In this method, participants are selected based on chance or
convenience, without any set pattern or systematic approach.
Essentially, the researcher selects whoever is easiest to recruit
without any specific strategy in mind.

Example: A journalist attending a public event and interviewing any

attendee they come across first, without a specific strategy.
Research Objective Example:
Research Objective: To get initial feedback on the usability of a new
mobile banking application.
In a haphazard sampling approach, a product developer might proceed
as follows:
Method:
1.The developer is at a cafe and decides to gather some immediate
feedback on the new app.
2.The developer approaches a few cafe patrons, asking if they have a
few minutes to try out the app and provide feedback.
3.Those who agree become the participants, and their feedback is
collected.