01 / 20

Random Sampling
and Data Collection

Master the foundations of statistical inference by understanding how to collect data through random sampling methods that produce reliable, unbiased results.

AP Statistics

Why Sampling Matters

In statistics, we rarely have access to an entire population: every individual or object we want to study. Instead, we collect data from a sample: a subset of the population. The quality of our conclusions depends entirely on how we select that sample.

Population vs. Sample

Population (N) Sample (n) Inference about population
Population
The entire group of individuals or items you want to study. Parameters describe populations.
Sample
A subset selected from the population. Statistics computed from samples estimate parameters.

Bias and Variability

High Bias Low Variability Low Bias High Variability

The Principle of Randomness

A random sample is one in which every member of the population has a known, nonzero probability of being selected. Randomness is the key tool for eliminating selection bias and allowing valid statistical inference.

Types of Random Sampling

Simple Random
Every possible sample of size \( n \) has an equal chance of being selected. The gold standard.
Stratified
Divide population into strata (subgroups), then take an SRS from each stratum.
Cluster
Divide population into clusters, randomly select some clusters, then sample all individuals within chosen clusters.
Systematic
Select every \( k \)-th individual from a list after a random starting point. Simple but requires a well-ordered list.

Simple Random Sample (SRS)

Population of 20 individuals ... Random selection Sample (n = 5)

Stratified vs. Cluster Sampling

Stratified Stratum 1 Stratum 2 Stratum 3 SRS from each stratum Cluster Cluster 1 Cluster 2 ✓ Cluster 3 ✓ All from selected clusters

A university population with four visibly different sampling routes

A university campus provides a setting for comparing sampling and study-design routes.The way people or groups enter a study determines its design. Sampling within groups, selecting whole groups and assigning treatments answer different questions.

Compare four sampling rules

Four 12-ID diagrams show a simple random sample, one selection from each stratum, all members of one cluster and a systematic sequence.The 12-ID examples use different selection rules: an SRS selects from the whole list; stratification selects within every group; cluster sampling takes all IDs in the chosen group; systematic sampling uses a random start and fixed spacing. The selection rule, not a single realized sample, identifies the design.

Allocate a sample across every class year

Four rows list class-year populations 2,400, 2,200, 1,800 and 1,600; selected samples 120, 110, 90 and 80; and weights 0.300, 0.275, 0.225 and 0.200.For a university with 8,000 students, allocate 400 selections as 120, 110, 90 and 80 across the four class-year rosters, then choose random IDs within each roster. Each student has a 5% selection chance; the fixed class-year allocations make this a stratified sample, not an SRS of 400.Open full diagram

Sources of Bias in Sampling

Voluntary Response
Individuals choose to participate. Strong opinions are overrepresented: online polls, call-in surveys.
Convenience
Researcher selects easily accessible individuals. Surveying only students in the cafeteria at noon.
Undercoverage
Some groups in the population are left out of the sampling frame entirely.

How to Conduct an SRS

1
Define the population. Clearly identify every individual who belongs to the group you are studying.
2
Create a sampling frame. List every member of the population and assign each a unique number.
3
Use a random mechanism. Use a random number table, calculator, or software to select \( n \) numbers.
4
Collect data. Contact the selected individuals and record their responses. Handle nonresponse carefully.
5
Analyze and infer. Because the sample is random, you can compute valid margins of error and confidence intervals.

Example 1: Identifying the Method

A school principal wants to survey student opinions about the lunch menu. She divides students into four grade levels (9, 10, 11, 12) and randomly selects 25 students from each grade level.

Example 2: Detecting Bias

A TV news station asks viewers to text "YES" or "NO" to the question: "Should the city build a new stadium?" They receive 12,000 responses: 78% say "NO."

Example 3: Designing a Sampling Plan

A researcher wants to estimate the average commute time for employees at a company with 5 office buildings across the city. Each building has about 200 employees. Budget allows surveying only 120 employees total.

Option A: Stratified
Treat each building as a stratum. Randomly select 24 employees from each of the 5 buildings. Ensures every building is represented.
Option B: Cluster
Treat each building as a cluster. Randomly select some buildings (e.g., 3 of 5) and survey all employees within them. This is often more convenient but can be less precise than stratified sampling.

Common Mistakes

✕ Confusing Cluster and Stratified
Stratified: sample from every group. Cluster: randomly choose some groups and sample all within. They are not interchangeable.
✕ Thinking Large Samples Fix Bias
A biased sample of 1,000,000 is still biased. Increasing sample size reduces variability, not bias. Only randomization reduces bias.
✕ Calling Convenience "Random"
Picking the first 50 people you see is not random, even if it feels unplanned. Random means using a chance mechanism.

Sampling Methods Compared

Method How It Works Advantage Limitation
SRS Every subset of size \( n \) equally likely Simple, unbiased Needs full list
Stratified SRS within each stratum Reduces variability Must know strata
Cluster Randomly select clusters, sample all Practical, cost-effective Higher variability
Systematic Every \( k \)-th individual Easy to implement Risk if list has pattern

Question 1 of 5

A researcher numbers all 800 employees at a factory and uses a random number generator to select 50 of them. What type of sampling is this?

Question 2 of 5

A website posts a poll asking visitors: "Do you support the new tax?" This method is most likely to produce which type of bias?

Question 3 of 5

A school district has 40 schools. Researchers randomly select 8 schools and survey every student in those schools. This is an example of:

Question 4 of 5

Which of the following is true about increasing sample size?

Question 5 of 5

A company wants to survey employees about workplace satisfaction. They divide employees into departments (Marketing, Engineering, Sales, HR) and randomly select 15 from each. Why is this approach better than an SRS of 60 employees?