Course contentsShow
← Data Science
Lesson 220 of 2,1456. The Normal Distribution in DepthFree lesson

Sample Size Requirements for the CLT

The rule of thumb is n≥30, but skewed or heavy-tailed distributions may require larger samples for normality.

Sample Size Requirements for the CLT

What You'll Learn

You'll discover the practical rule for how large your sample needs to be before the Central Limit Theorem kicks in—and why it's not one-size-fits-all.

The n ≥ 30 Rule of Thumb

The Central Limit Theorem tells us that sampling distributions approach normality, but how many observations do you actually need?

The classic rule is n ≥ 30. With 30 or more observations, the distribution of sample means tends to look normal enough for most statistical work—even if your original data isn't normally distributed.

When 30 Isn't Enough

Here's the catch: 30 is just a guideline, not a magic number.

Think of it like baking: a basic recipe says "bake for 30 minutes," but if your ingredients are unusual (very cold, very dense), you might need more time.

Similarly:

  • Symmetric distributions (like uniform): even n = 10–15 might suffice
  • Moderately skewed data: n ≥ 30 usually works
  • Highly skewed distributions (like income data): you may need n = 50, 100, or more
  • Heavy-tailed distributions (with extreme outliers): even larger samples required

The more your data deviates from normality—especially with long tails or severe skewness—the larger your sample must be for the CLT to produce a reasonably normal sampling distribution.

Practical Implication

Before relying on CLT-based methods (confidence intervals, hypothesis tests), check your data's shape. If it's wildly skewed or has extreme outliers, aim for larger sample sizes to be safe.

Key Takeaway: While n ≥ 30 is the standard rule for the CLT, skewed or heavy-tailed distributions require larger samples—sometimes 50, 100, or more—to achieve approximate normality in the sampling distribution.