Finite Population Correction
What You'll Learn
You'll learn when and how to adjust the standard error formula when your sample represents a large fraction of a small population.
The Core Idea
Until now, you've calculated standard error assuming the population is enormous compared to your sample—essentially infinite. But what if you're sampling 50 employees from a company of only 100? Or surveying 80 students from a class of 120?
When you sample without replacement from a small population, each observation you draw changes what's left in the population. This makes your sample less variable than theory predicts—you're actually getting more information than the standard formulas suggest.
The Finite Population Correction (FPC) factor adjusts for this by shrinking the standard error to reflect the extra precision:
FPC = √[(N - n) / (N - 1)]
Where:
- N = population size
- n = sample size
You multiply your standard error by this factor. Notice:
- If n is tiny compared to N, the FPC ≈ 1 (no correction needed)
- If n approaches N, the FPC approaches 0 (almost no uncertainty left!)
When to Use It
Apply FPC when:
- Sampling without replacement
- Your sample is more than 5% of the population (n/N > 0.05)
Analogy
Imagine drawing colored balls from a jar. If the jar has 10 million balls and you draw 10, the contents barely change. But if the jar has 20 balls and you draw 10, you've fundamentally altered what remains—your remaining draws are more predictable.
Key Takeaway: When your sample is a substantial fraction (>5%) of a finite population and you're sampling without replacement, multiply your standard error by the finite population correction factor to get accurate uncertainty estimates.