Sampling Methods
Learn different sampling techniques: random, stratified, cluster, and systematic sampling for effective data collection.
On This Page
Why Sampling Matters
Key Terms
| Term | Definition |
|---|---|
| Population | The entire group we want to study |
| Sample | A subset of the population we actually collect data from |
| Parameter | A number describing the population (μ, σ, p) |
| Statistic | A number calculated from sample data (x̄, s, p̂) |
| Sampling frame | List of all individuals in the population |
Probability vs Non-Probability Sampling
| Probability Sampling | Non-Probability Sampling |
|---|---|
| Every member has known chance of selection | No known selection probabilities |
| Results generalizable to population | Cannot generalize statistically |
| More expensive, time-consuming | Cheaper, faster |
| Examples: SRS, stratified, cluster | Examples: convenience, voluntary |
Simple Random Sampling (SRS)
In a simple random sample, every member of the population has an equal chance of being selected, and every possible sample of size n is equally likely.
Population: 1000 employees at a company
Method:
- Assign each employee a number (1-1000)
- Use random number generator to select 50 numbers
- Survey those 50 employees
Result: SRS of size 50
Advantages
- Eliminates selection bias
- Simple to understand
- Easy to calculate sampling error
Disadvantages
- Requires complete sampling frame
- May miss small subgroups
- Can be impractical for large populations
Stratified Sampling
Divide the population into strata (groups), then take a random sample from each stratum.
Population: University students Strata: Year (Freshman, Sophomore, Junior, Senior)
Method:
- Separate students by year
- Take SRS from each year
- Combine into final sample
| Year | Population | Sample (10%) |
|---|---|---|
| Freshman | 3000 | 300 |
| Sophomore | 2500 | 250 |
| Junior | 2200 | 220 |
| Senior | 2300 | 230 |
| Total | 10000 | 1000 |
When to Use Stratified Sampling
- Subgroups are important to study
- Subgroups differ from each other
- Want to ensure representation of all groups
- Subgroups vary greatly in size
Advantages
- Ensures representation of all strata
- Often more precise than SRS
- Can analyze subgroups separately
Disadvantages
- Requires knowledge of strata membership
- More complex to implement
- Need to identify relevant strata
Cluster Sampling
Divide the population into clusters (natural groups), randomly select some clusters, then sample everyone (or many) within selected clusters.
Population: All high school students in a state Clusters: Individual schools
Method:
- List all high schools in state (clusters)
- Randomly select 20 schools
- Survey all students in those 20 schools
Advantage: Don’t need list of all students—just schools!
Stratified vs Cluster
| Stratified | Cluster |
|---|---|
| Sample from ALL strata | Sample from SOME clusters |
| Strata are different from each other | Clusters are similar to each other |
| Small sample from each stratum | All (or many) from selected clusters |
| Increases precision | Decreases cost |
Advantages
- Cost-effective for geographically spread populations
- Don’t need complete sampling frame
- Practical for field research
Disadvantages
- Less precise than SRS or stratified
- Clusters may not be representative
- Larger sampling error
Systematic Sampling
Select every kth individual from the sampling frame after a random start.
Sampling interval:
Where N = population size, n = desired sample size
Population: 5000 customers Desired sample: 100
Method:
- k = 5000/100 = 50
- Randomly select start between 1-50 (say, 23)
- Select customers: 23, 73, 123, 173, …
Every 50th customer starting from #23.
Advantages
- Simple to implement
- Ensures spread across population
- No need for random numbers for each selection
Disadvantages
- Can miss patterns if population has periodicity
- Not truly random if starting point biased
Non-Probability Sampling Methods
Convenience Sampling
Select whoever is readily available.
- Survey people at a mall
- Poll students in your class
- Use your social media followers
Problem: Those available may differ from those not available.
Voluntary Response Sampling
People choose to participate on their own.
- Online reviews
- Call-in polls
- Surveys with opt-in response
Problem: Those with strong opinions are more likely to respond.
Snowball Sampling
Ask participants to recruit others.
Useful for: Hard-to-reach populations
- People with rare diseases
- Underground communities
- Specialized professionals
Sampling Errors and Bias
Sampling Error
Random variation between sample statistic and population parameter. Always present; decreases with larger n.
Sampling Bias
Systematic tendency to favor certain outcomes. Does NOT decrease with larger samples!
| Type | Example | Prevention |
|---|---|---|
| Selection bias | Surveying only daytime shoppers | Random selection |
| Nonresponse bias | Only 30% respond to survey | Follow-up, incentives |
| Response bias | Leading questions | Careful wording |
| Undercoverage | Missing homeless in census | Improve sampling frame |
Choosing a Sampling Method
Summary
In this lesson, you learned:
- Simple Random Sample (SRS): Every individual has equal chance
- Stratified sampling: Divide into groups, sample from each
- Cluster sampling: Select groups, sample within them
- Systematic sampling: Every kth individual from random start
- Non-probability methods (convenience, voluntary) cannot support statistical inference
- Sampling error is random; sampling bias is systematic
- Different methods suit different research needs and constraints
Practice Problems
1. A researcher wants to study eating habits of college students. She surveys everyone in her 9 AM class. What type of sampling is this? What’s the problem?
2. To study employee satisfaction at a large company with 4 departments, you randomly select 25 employees from each department. What sampling method is this?
3. A polling organization randomly selects 15 cities, then surveys all adults in those cities. What type of sampling is this?
4. You want a sample of 200 from a list of 10,000 customers. Describe how to implement systematic sampling.
Click to see answers
1. Convenience sampling (whoever was available)
Problems:
- Cannot generalize to all college students
- 9 AM students may differ from others (morning people, certain majors)
- Only one class from one school
- Results cannot support statistical inference
2. Stratified sampling
Departments are the strata, and a random sample is taken from each stratum.
3. Cluster sampling
Cities are the clusters. Some clusters are randomly selected, and then everyone (or many) within those clusters is sampled.
4. Systematic sampling procedure:
- Calculate interval: k = 10,000/200 = 50
- Randomly select a starting number between 1-50 (e.g., 17)
- Select customers: 17, 67, 117, 167, 217, …
- Continue selecting every 50th customer until you have 200
- Final sample: customers #17, 67, 117, 167, …, 9967
Next Steps
Continue learning about statistical inference:
- Sampling Distributions - Distribution of sample statistics
- Point Estimation - Estimating parameters
- Confidence Intervals - Interval estimates
Was this lesson helpful?
Help us improve by sharing your feedback or spreading the word.