Group 3

Biostatistics Assignment
Contents
• Sampling Methods
• Multistage Sampling
• Sampling Error and Convenience Sampling
• Sampling and Volunteer Sampling
• Simple Random Sampling and Judgement Sampling
• Systematic Random Sampling and Quouta Sampling
• Stratified Randosa Sampling and Snowball Sampling
• Sampling Distribution
 Sampling Methods
Cluster Sampling:

• Sampling methods are the techniques • The population is divided into clusters
researchers use to select a manageable subset (groups), and a random sample of
(the sample) from a larger group (the clusters is selected. All members within
population) for study. the chosen clusters are then included in
the sample.
• The goal is to draw conclusions about the
entire population based on the data collected • Useful when the population is naturally
from the sample. divided into groups and it's impractical to
sample individuals directly.
• Example of Cluster Sampling:
Types of Sampling Methods
• Imagine you want to survey student
• Random Sampling: Every member of the satisfaction at a large university:
population has an equal chance of being
chosen. • 1. Define Clusters: Divide the university
population into clusters by academic
• Stratified Sampling: The population is divided departments (e.g., Biology, Engineering,
into subgroups (strata), and samples are History, etc.)
randomly selected from each stratum.
• 2. Select Clusters: Randomly choose a
• Systematic Sampling: Selecting every nth few of these departments.
member from a list or population.
• 3. Sample All Within Clusters: Survey
• Convenience Sampling: Selecting individuals every student within the selected
who are readily available or easy to reach. departments.
 Multistage Sampling
• Multistage sampling is a • Example:
method of selecting a • Let's say you want to survey high
sample in multiple steps, school students about their
using progressively smaller mental health across an entire
country:
groupings at each stage.
• 1. Stage 1: Randomly select a
• This approach is especially few states from the country.
useful for studying large, • 2. Stage 2: From those selected
dispersed populations states, randomly select a few
where a complete list of cities/towns.
individuals is impractical. • 3. Stage 3: From those selected
It's like an extended version cities/towns, randomly select a
of cluster sampling. few high schools.
• 4. Stage 4 (Final): From those
selected high schools, randomly
select individual students to
participate in the survey.
 Sampling error and convenience
Sampling
Sampling Error
Convenience Sampling
• Sampling error is the difference • Convenience sampling is a non-
between the results obtained from a probability sampling method where
sample and the true results that would subjects are selected based on their easy
be obtained from the entire accessibility and availability to the
population. It occurs because a sample researcher.
is not a perfect representation of the • Cause: Convenience sampling is not
population. based on random selection, but is left
• Cause: Sampling error arises due to largely to the discretion of the researcher.
the inherent variability between • Impact: Convenience sampling can lead
samples drawn from the same to sampling bias, where certain groups
population. may be over- or under-represented in the
sample. This can result in findings that
• Impact: Sampling error can lead to are not generalizable to the overall
inaccurate conclusions or estimates population.
about the population. • Suitability: Convenience sampling is often
• Minimization: Sampling error can be used in exploratory research or when
minimized by increasing the sample other sampling methods are not feasible.
size or using more sophisticated However, it is not suitable for research
sampling techniques. that aims to make accurate inferences
about a broader population.
 Sampling and Volunteer
Sampling
What is Sampling? Volunteer Sampling
• Sampling is the process of selecting a subset of A type of non-probability sampling
individuals (or items) from a larger population to where individuals volunteer to
gather data and make inferences about the whole
group. participate in a study. Researchers
• It's a practical and cost-effective way to study rely on people who are willing to
populations, especially large ones, without come forward.
examining every single member.
Example of Volunteer Sampling:
• Two Main Types of Sampling:
1. Probability Sampling: Let's consider a study on the
• Involves random selection, where every member effectiveness of a new educational
of the population has a known chance of being program to improve reading skills:
chosen.
Recruitment: Researchers advertise
Examples: Simple random, stratified, cluster,
systematic sampling.
for parent volunteers (and their
children) to participate in the
2. Non-Probability Sampling:
program.
Does not use random selection, meaning some
members have a higher chance of being selected • Volunteer Sample: The resulting
than others. sample might primarily consist of
Examples: Convenience, judgmental, quota sampling. students whose parents are highly
motivated and engaged in education.
 Simple Random Sampling and
Judgment Sampling
Simple random sampling Judgment Sampling
• Every member of the population has The researcher uses their expertise or
an equal chance of being selected for judgment to select participants they
the sample. This is the gold standard believe are most representative of the
for minimizing bias and ensuring population or will provide the most
generalizability. useful data.
• Method: Random number generators, • Method: The researcher hand-picks
lottery methods, or random number participants based on their knowledge
tables are used to select participants. and experience.
• Advantages: Unbiased, highly • Advantages: Can be quicker and
generalizable. cheaper than random sampling; useful
• Disadvantages: Can be time- when specific expertise is needed.
consuming and expensive, especially for • Disadvantages: Highly susceptible to
large populations; requires a complete researcher bias; results are not
sampling frame (a list of all population generalizable to the wider population.
members). • Example: A researcher studying expert
• Example: Selecting 100 students from opinions on climate change might select
a university's enrollment list by using a only leading climate scientists for
random number generator to choose interviews.
Systematic Random Sampling and Quota Sampling
Both systematic random sampling and quota sampling are methods
used to select samples, but they differ in their approach and how
representative the resulting samples are likely to be.

Quota Sampling
Systematic Random Sampling • A form of non-probability sampling that
• Is a form of probability sampling. ensures the subgroups are represented in
the sample in proportions similar to the
• Method:
population. Method: Divide the population
• Start with a randomly ordered list of the into subgroups based on characteristics
population. (e.g., gender, age).
• Randomly select a starting point within • Set quotas for the number of individuals
the list. needed from each subgroup to match
• Select every ×k×th individual, where ×k× population proportions.
= (Population Size / Sample Size).
• Continue selecting individuals from each
• Uses a fixed interval to choose samples subgroup until the quotas are met.
after random start. Ensures the sample is Example: Let's say we want to survey 500
spread evenly across the population and people in a town of 5000:Systematic
reduces some bias. Random Sampling:
• Systematic Random Sampling:
1. Randomly order the list of residents.
• 1. Randomly order the list of residents.
• 2. Choose a random number between 1 2. Choose a random number between 1 and
and 10 (since 5000/500 = 10). Let's say it's 10 (since 5000/500 = 10). Let's say it's 3.
3. • 3. Select person #3, then person #13, #23,
• 3. Select person #3, then person #13, #33, and so on.
#23, #33, and so on.
Stratified Random Sampling Snowball Sampling

• Divides population into subgroups (strata) • Non-probability method. Existing subjects

based on shared characteristics, then takes a recruit future subjects from their network.
random sample from each.
• Goal: Reach hard-to-reach or hidden
• Ensures representation from all important populations.
subgroups.
Steps:
Steps: 1. Find seed individuals.
1. Identify Strata . 2. Ask for referrals.
2. Divide Population 3. Repeat until sample is reached.
3. Randomly Sample within each stratum Example:
4. Combine Samples * Study experiences of undocumented
Example: workers.
* Study student satisfaction. * Start: Find a few initial individuals.
* Strata: By year (Freshman, Sophomore, etc.) * Referrals: Ask for more participants from
their network.
 Sampling Distributions

• A sampling distribution is the Types of Errors in Sampling

probability distribution of a
statistic (e.g., sample mean, • Sampling Error: Discrepancy
sample standard deviation) between a population parameter
calculated from multiple and its corresponding sample
samples of the same size, taken statistic due to random chance in
from the same population. sampling.
• • It shows how a sample • • Can be influenced by
statistic would vary if we sampling method.
repeated the sampling process • • Non-Sampling Error: Errors
many times. arising from procedural biases:
• • It's crucial for making • • Incorrect responses,
measurement errors, or errors in
inferences about the data processing.
population from sample data.
 Conclusion
Sampling Methods: Techniques for selecting a manageable subset from a larger population to draw conclusions about the
whole. The goal is to obtain representative data that minimizes bias.
Types of Sampling:
• Probability Sampling: Every member has a known, non-zero chance of selection. Examples include:
• Simple Random: Equal chance for every member and every possible sample.
• Stratified Random: Population divided into subgroups (strata), then random sampling within each.
• Systematic Random: Selecting every kth member from an ordered list.
• Cluster: Population divided into clusters, then random selection of clusters, and all members within selected clusters are
sampled.
• Multistage: A combination of sampling methods in multiple steps, using progressively smaller groupings. Often an
extension of cluster sampling.
• Non-Probability Sampling: Does not involve random selection; some members have a higher chance of selection than
others. Examples include:
• Convenience: Selecting readily available individuals.
• Judgment: Researcher uses expertise to select participants.
• Quota: Ensures representation of subgroups based on pre-set quotas.
• Volunteer/Self-Selection: Individuals volunteer to participate.
• Snowball: Existing participants recruit future subjects.
Sampling Error: The difference between sample results and true population results. Increased by non-probability sampling
methods.
Example (Multistage): Surveying high school students nationally:
1. Randomly select states.
2. Randomly select cities within selected states.
3. Randomly select high schools within selected cities.