Sampling

BRM
Primary Data Collection
There are many methods to collect or obtain data for statstical
analysis(or qualitative analysis).
• Observation
• Survey
• Experiments
• Interview
survey
A survey is a method of collecting information from people by asking
them questions. It helps understand their opinions, experiences, or
behaviors.
Surveys can be done through
interviews
questionnaires
or online forms.
Survey Response Rate
• Survey response rate is the percentage of people who complete a
survey compared to the total number who received it.

• Formula:
Response Rate=(Completed Surveys/Total Surveys Sent)×100
Questionnaire Design – Key Principles
A well-designed questionnaire helps collect clear and useful information.
Key Design Principles:
• Clarity – Questions should be simple and easy to understand.
• Relevance – Ask only what is necessary for the research.
• Logical Flow – Arrange questions in a natural order.
• Neutral Wording – Avoid leading or biased questions.
• Variety of Question Types – Use both open-ended and closed-ended
questions.
• Consistency – Keep answer choices uniform and clear.
• Easy to Answer – Make it user-friendly and not too long.
Pilot Study in Questionnaire Design
A pilot study is a small-scale test of a questionnaire before using it in
full research. It helps find errors and improve questions.
Key Points:
• Test Run – Checks if questions are clear and effective.
• Identifies Issues – Finds confusing or biased questions.
• Improves Quality – Helps refine wording and structure.
• Saves Time & Effort – Prevents mistakes in main survey.
• Gathers Feedback – Allows adjustments based on responses
Sampling in Research
Sampling is choosing a smaller group (sample) from a larger population to study. It helps
researchers collect information without surveying everyone.
Example:
• If a company wants to know customer satisfaction, it surveys 500 customers instead of all
50,000.
Sampled vs. Target Population:
• Target Population – The full group researchers want to study (e.g., all students in a
school).
• Sampled Population – The selected group from the target population (e.g., 100 students
from different grades).
Why Sampling is Important?
• Saves Time & Cost – Studying the entire population is expensive and time-consuming.
• Better Data Quality – Allows focused, detailed analysis.
• Easier Data Collection – Reduces effort compared to large-scale surveys.
• Practical & Efficient – Works well for research and decision-making.
Term
1. Population
The entire group researchers want to study.
Example: All students in a university.
2. Element(Unit Of Analysis)
An individual unit within the population.
Example: A single student in the university.
3. Population Frame
A complete list of all elements in the population, used for sampling.(errors are
possible)
Example: A database of student names in the university.
Term
Sample
A small group selected from a larger population for study.
Example: Surveying 200 students instead of all 5,000 in a university.
Sampling
The process of choosing a sample from the population.
Example: Selecting students from different departments to represent the
university.
A subject is a single member of the sample, just as an element is a single
member of the population.
 Sampling Plan
A sampling plan is a strategy for selecting a sample from a population for research. It ensures data is
collected effectively and accurately.
1. Probability Sampling (Representative Samples)(Generalizability-Have External Validity)
• Sampling where every member of the population has an equal chance of being selected. Example:
Randomly picking students from different classes in a school.
Types of Probability Sampling:
• (unrestricted)Simple Random Sampling – Every individual has an equal chance(A university selects
50 students randomly from all departments to participate in a survey.).
(restricted)Complex Random Sampling
• Stratified Sampling – Dividing the population into groups and sampling from each.(A company divides
employees into departments (HR, Sales, IT) and selects 10 employees from each department for
feedback.)
• Systematic Sampling – Selecting every nth person from a list.(A hospital has a patient list and selects
every 5th patient to fill out a satisfaction form.)
• Cluster Sampling - A method where the population is divided into groups (clusters), and some clusters
are randomly selected for study.(A company wants employee feedback. Instead of picking random
Sampling Plan
2. Non-Probability Sampling (Non-Representative Samples)
Sampling where selection is not random and some people have a higher chance of
being chosen. Example: Surveying volunteers instead of randomly chosen people.
Types of Non-Probability Sampling:
Convenience Sampling – Selecting easily available participants(A researcher
surveys shoppers at a mall because they are easy to reach.).
Purposive Sampling – Choosing people based on specific criteria.(A study on
heart disease selects only patients diagnosed with the condition to participate.)
Snowball Sampling(refereral method) – Participants recommend others to join
the study.(A survey on freelancers starts with a few participants, who then
recommend others to join.)
(unrestricted)Simple Random Sampling
Characterstics
• Least Bias
• Most Generalizability(highest External Validity)
• But Expensive
• Update list of population may not be available.
• it is less used.
Systematic Sampling
 Stratified Sampling
Mutually Exclusive group
Divide Population In strata/Group
such as:
age group
income group
gender group
Proportionate vs. Disproportionate Stratified
Sampling
Proportionate Stratified Sampling: Each group (stratum) is sampled
in proportion to its size in the total population.
Example: A school has 80% seniors and 20% juniors, so a study selects
80 seniors and 20 juniors to match the real distribution.
Disproportionate Stratified Sampling: Groups are not sampled in
proportion to their population size, often to ensure enough data from
smaller groups.
Example: A company has 90% office workers and 10% managers, but a
survey selects 50 workers and 50 managers to compare both groups
equally.
Cluster Sampling
Cluster sampling May Increase sampling error due to similarities among
cluster member.(cluster should be as internally dissimilar as possible
and different cluster should be very similar.)
• it is less expensive or low cost
• Greater Biases are possible
• Less time consuming
• Convenient, easy to inspect.
 Difference Between

Feature Stratified Sampling Cluster Sampling

Population is divided into groups (strata), and

Population is divided into clusters, and some
Definition individuals are randomly selected from each
entire clusters are randomly chosen for study.
group.

Individuals are selected from each group Entire clusters are chosen, and everyone in
Selection Process
(stratum). selected clusters is surveyed.

Makes data collection easier in large

Purpose Ensures representation from all groups.
populations.

A school survey selects students from each A survey on schools selects 3 entire schools and
Example
grade (e.g., 10 from each grade). surveys all their students.
Multistage Cluster Sampling
A method where sampling happens in multiple steps, refining the
selection at each stage. It helps when populations are large and spread
out.
Example:
A government survey on education:
• First stage: Randomly selects 5 provinces.
• Second stage: Randomly picks 10 cities from those provinces.
• Third stage: Randomly chooses 20 schools from those cities.
• Final stage: Selects students from the chosen schools for the survey.
Advantages of Multistage Sampling
• Efficient for large populations – Saves time and effort.
• Cost-effective – Reduces expenses compared to full population
surveys.
• Flexible – Allows adjustments at different stages.
• Better representation – Ensures diverse samples across multiple
levels.
• Simplifies data collection – Makes large-scale research manageable.
Use Of Non-Probability Sampling:
• It is quick and inexpensive
• Provide useful Information
• It can be externally valid (generalizable) if theory is strong and
measure are sound)
• Most use sampling method in business and social science research.
• No sample frame
Convenience Sampling
• Used during the exploratory phase of research project.
• Getting basic information quickly and efficiently.
• Some study use conenience sampling if theory is strong and measure
are established to enhance knowledge.(Calder et
al.,1982;Lucas,2003).
Purposive Sampling-Two Types
Judgment vs. Quota Purposive Sampling
• Judgment Purposive Sampling: Researchers choose participants
based on their knowledge or expertise in the topic.
Example: A study on heart disease selects experienced cardiologists
instead of general doctors.
• Quota Purposive Sampling: Researchers set a fixed number (quota)
for different groups to ensure balanced representation.
Example: A job satisfaction survey selects equal numbers of employees
from different departments, even if some departments have more
workers.
Snowball Sampling
Sample Size
• Larger the Sample Size is the more accurate we can expect the sample estimate
to be. but too much larger sample size is not easy.
Sample Size: The number of subjects selected from a population for a study.
• How Many to Select? It depends on the total population size. Larger
populations need proportionally smaller samples.
• Krejcie & Morgan (1970): They created a sample size table to help
researchers decide how many subjects to select.
Example:
• If a population has 500 people, the table suggests selecting 217 for a reliable
study.
• If a population has 1,000 people, about 278 should be selected.
 Sampling vs. Non-Sampling Errors

Type of Error Definition Example

Happens when the selected sample does A survey of 100 students may not reflect
Sampling Error
not fully represent the whole population. the opinions of all students in a university.

Mistakes in data collection, analysis, or

A survey question is unclear, leading to
Non-Sampling Error interpretation, not related to sample
wrong answers from participants.
selection.
Errors in Data Acquisition (Non-Sampling
Error)
These errors happen when data is collected incorrectly, leading to
inaccurate results.
Common Causes:
• Wrong measurements – A faulty thermometer records incorrect
temperatures.
• Data entry mistakes – A survey response is typed incorrectly.
• Misinterpretation – A researcher misunderstands a participant’s
answer.
• Technical issues – A broken microphone records unclear responses.
Non-Response Error or Bias
This happens when some people do not respond to a survey, making the
results less accurate.
Causes:
• Busy participants – People with tight schedules may skip the survey.
• Sensitive questions – Some may avoid answering personal topics.
• Survey length – Long surveys may discourage responses.
Example:
A company surveys employees about work stress, but stressed employees
are too busy to respond. This makes the results less reliable because only
relaxed employees answered.