Stratified Sampling
What you'll learn: How to improve sample accuracy by dividing your population into meaningful groups before sampling.
The Core Idea
Imagine you want to survey students about campus meal plans, and your university has 60% undergraduates and 40% graduate students. If you just use simple random sampling, you might accidentally get 80% undergrads and 20% grads—misrepresenting the true population.
Stratified sampling solves this by dividing your population into strata (homogeneous subgroups) and then sampling proportionally from each stratum. In our example, you'd ensure exactly 60% of your sample comes from undergrads and 40% from grads.
How It Works
- Identify meaningful strata: Divide your population into non-overlapping groups based on important characteristics (age, income, region, education level, etc.)
- Determine proportions: Calculate what percentage of the total population each stratum represents
- Sample from each stratum: Use simple random sampling within each stratum, maintaining the correct proportions
Why Use Stratified Sampling?
Unlike simple random sampling or systematic sampling, stratified sampling guarantees representation of key subgroups. This reduces sampling error and produces more precise estimates, especially when:
- Subgroups differ significantly from each other
- You need to compare specific subgroups later
- Some groups are small but important to include
Example in Plain English
A national health survey wants 1,000 participants. The population is 30% rural, 70% urban. Stratified sampling ensures you get exactly 300 rural and 700 urban respondents—reflecting reality better than random chance might.
Key Takeaway: Stratified sampling divides populations into homogeneous subgroups (strata) and samples proportionally from each, ensuring all important segments are properly represented in your sample.