Central Limit Theorem

The Central Limit Theorem (CLT) is one of the most powerful and surprising results in all of mathematics. It states that when you take sufficiently large random samples from almost any population — regardless of the population's own shape or distribution — the distribution of the sample means will be approximately normal (bell-shaped). This remarkable fact is why the normal distribution appears so frequently in nature, science, and statistics. To appreciate why the CLT matters, imagine measuring the heights of 10 randomly chosen students, computing their average, and repeating this process thousands of times. Even if individual heights were skewed or had an unusual distribution, the collection of those thousands of averages would form a nearly perfect bell curve. This predictability allows statisticians to make confident probability statements and construct reliable confidence intervals without knowing the exact shape of the original population. The CLT has three key conditions: the samples must be drawn randomly, each observation must be independent of the others, and the sample size must be large enough (a common rule of thumb is n ≥ 30, though skewed distributions may require larger samples). When these conditions hold, the sampling distribution of the mean behaves in a mathematically precise way, enabling everything from quality control in manufacturing to clinical drug trials.

Key takeaways
  • Start with Any Population
  • Draw Repeated Random Samples of Size n
  • The Sampling Distribution Becomes Normal
  • Use the Normal Approximation for Probability Calculations
  • Checking the Conditions

Start with Any Population

Consider a population with mean μ (mu) and standard deviation σ (sigma). The population can be uniform, skewed, bimodal, or any other shape — the CLT applies as long as the population has a finite mean and finite variance. For example, the number of minutes a customer waits in line might be right-skewed, with most people waiting briefly but a few waiting a very long time.

Draw Repeated Random Samples of Size n

Select many independent random samples, each of size n, from the population. For each sample, compute the sample mean x̄. Each sample will produce a slightly different x̄ because of random variation. The collection of all these possible sample means forms what we call the sampling distribution of x̄.

The Sampling Distribution Becomes Normal

As n increases, the shape of the sampling distribution of x̄ converges to a normal distribution, no matter what the original population looked like. The center of this new distribution equals the population mean μ, and its spread is measured by the standard error σ/√n. Notice that larger samples produce smaller standard errors — averages from big samples cluster more tightly around the true mean.

Use the Normal Approximation for Probability Calculations

Once you know the sampling distribution is approximately normal with mean μ and standard error σ/√n, you can convert any sample mean x̄ into a z-score using z = (x̄ − μ)/(σ/√n). Standard normal tables or technology then give exact probabilities. This is the bridge connecting raw data to confident statistical inference.

Checking the Conditions

The three conditions must always be verified: (1) Random — the sample must be drawn randomly to avoid bias. (2) Independence — individual observations must not influence each other; for sampling without replacement, the sample must be less than 10% of the population (the 10% condition). (3) Large enough sample — for means, n ≥ 30 is the standard guideline; for proportions, np ≥ 10 and n(1−p) ≥ 10 are required. Violating these conditions makes normal approximations unreliable.

The Key Equation

Xˉ∼N ⁣(μ, σ2n)as n→∞\bar{X} \sim N\!\left(\mu,\, \frac{\sigma^2}{n}\right) \quad \text{as } n \to \infty
  • \bar{X} — Sample Mean (same as original variable): The arithmetic average of n observations drawn from the population.
  • \mu — Population Mean (same as original variable): The true average of the entire population; also the center of the sampling distribution.
  • \sigma^2 — Population Variance (squared units of original variable): A measure of how spread out the population values are around the mean.
  • n — Sample Size (dimensionless (count)): The number of observations in each random sample. Larger n produces a narrower sampling distribution.
  • \sigma/\sqrt{n} — Standard Error (same as original variable): The standard deviation of the sampling distribution of x-bar; it measures how much sample means vary from sample to sample.

Worked Example

Problem. A vending machine dispenses coffee with a mean volume of μ = 236 mL and standard deviation σ = 15 mL. An inspector randomly selects 49 cups. (a) What is the mean and standard error of the sampling distribution of x̄? (b) What is the probability that the sample mean volume is less than 233 mL?

μxˉ=236 mL,SE=1549≈2.14 mL,P(Xˉ<233)≈0.0808\mu_{\bar{x}} = 236\text{ mL},\quad SE = \frac{15}{\sqrt{49}} \approx 2.14\text{ mL},\quad P(\bar{X}<233) \approx 0.0808

Answer. Mean of sampling distribution = 236 mL; Standard Error ≈ 2.14 mL; P(x̄ < 233) ≈ 0.0808 (about 8.1%)

Figure
Diagram
Central Limit Theorem: Effect of Sample SizePopulation (Skewed)n = 5n = 30Right-skewedAny shapeSlightly bell-shapedSome skew remainsNearly normalCLT in full effectKey QuantitiesMean: μ_x̄ = μStd Error: σ/√nShape → N(μ, σ²/n)unchangedshrinks as n growsas n → ∞As sample size n increases, the sampling distribution of x̄ becomes more normalRule of thumb: n ≥ 30 for CLT to apply reliably
Origins · 1 minPierre-Simon Laplace and how this was discovered
PL
Pierre-Simon Laplace
1749–1827 · French

Laplace provided the first rigorous proof of what we now call the Central Limit Theorem in 1810, showing that sums of many independent random variables tend toward a normal distribution. He built on earlier probabilistic work by Abraham de Moivre and Alexis Claude Clairaut, greatly extending the theorem's scope and mathematical precision. His formulation in 'Théorie Analytique des Probabilités' (1812) laid the groundwork for modern statistical inference. Laplace was so confident in determinism that he imagined a hypothetical super-intelligent being — now called 'Laplace's Demon' — that could predict the entire future of the universe if it knew the exact position and momentum of every particle. Despite this deterministic worldview, he spent much of his career developing the mathematics of probability and uncertainty.

  • Introduction to Sampling Distributions — The CLT describes the behavior of sampling distributions; understanding what a sampling distribution is provides the essential foundation for the theorem.
  • Normal Distribution — The CLT concludes that sample means are approximately normally distributed; knowing the properties of the normal curve is required to use CLT results in probability calculations.
  • Standard Error of the Mean — Standard error σ/√n is the key parameter measuring the spread of the sampling distribution and appears directly in CLT-based z-score formulas.
  • Confidence Intervals — Confidence intervals for a population mean are constructed using the CLT result that x̄ is approximately normal, making this a direct application of the theorem.
  • Law of Large Numbers — The Law of Large Numbers guarantees that sample means converge to the population mean, which complements the CLT's description of how those means are distributed around μ.