Central Limit Theorem: Statement, Formula, Examples

Key Points

• The central limit theorem says the sampling distribution of the mean is approximately normal when the sample size is large enough.

• This holds even when the population itself is skewed, flat or otherwise not normal.

• The sampling distribution is centred on the population mean μ, with standard deviation σ/√n, called the standard error.

• A common rule of thumb is that samples of 30 or more are large enough, though skewed populations may need more.

• The theorem is why normal-based methods, such as z-tests and confidence intervals, work for so many kinds of data.

Why Averages Look Normal

One die roll is flat, but the average of ten rolls piles up in a bell shape every time.

That pattern is the central limit theorem at work. It is one of the most important results in statistics, because it explains why the normal distribution appears whenever averages are involved.

What Is the Central Limit Theorem?

Take repeated random samples of size $n$ from any population with mean $\mu$ and standard deviation $\sigma$, and compute each sample's mean $\bar{x}$. The central limit theorem says that, for large enough $n$,

$$\bar{X} \approx N!\left(\mu, \frac{\sigma}{\sqrt{n}}\right)$$

The sample means are approximately normally distributed, centred on the population mean, with a spread of $\frac{\sigma}{\sqrt{n}}$.

The figure above shows it with dice. A single roll is equally likely to be any face, which is nothing like a bell. Yet means of 10 rolls cluster tightly around 3.5 in a near-perfect bell shape.

What Is the Standard Error?

The standard deviation of the sample mean is called the standard error:

$$SE = \frac{\sigma}{\sqrt{n}}$$

Because of the square root, quadrupling the sample size only halves the standard error. Going from 25 to 100 observations cuts it from 2 to 1 when $\sigma = 10$.

When Does the Central Limit Theorem Apply?

A few conditions.

The "30 or more" rule is a guideline, not a law. A roughly symmetric population needs fewer; a heavily skewed one, such as incomes, may need many more before the sample means look normal.

Examples of the Central Limit Theorem

Example 1

Test scores have mean 70 and standard deviation 10. For samples of 25 students, what are the mean and standard error of the sample mean?

Mean 70, standard error $\frac{10}{\sqrt{25}} = 2$.

Example 2

For the same scores, what is the probability that a sample of 25 has a mean above 74?

A first attempt uses the individual standard deviation: $z = \frac{74 - 70}{10} = 0.4$, giving a probability of about 0.34.

That treats the sample mean like a single score. Averages vary much less than individuals, so the standard error, 2, is the right spread:

$$z = \frac{74 - 70}{2} = 2, \qquad P(\bar{X} > 74) = P(Z > 2) \approx 0.023$$

Example 3

A population has standard deviation 12. How large a sample gives a standard error of 2?

$\frac{12}{\sqrt{n}} = 2$ gives $\sqrt{n} = 6$, so $n = 36$.

Example 4

Incomes are strongly right-skewed. Can the central limit theorem be used for the mean income of a sample of 10?

Probably not reliably. With heavy skew, 10 observations are too few for the sample mean to be close to normal. A much larger sample is needed.

Example 5

The mean of one die roll is 3.5 and its standard deviation is about 1.71. What is the standard error of the mean of 10 rolls?

$\frac{1.71}{\sqrt{10}} \approx 0.54$, which matches the narrow bell in the figure.

Example 6

Does the central limit theorem make the population normal?

No. It describes the distribution of sample means. The individual values keep whatever shape the population has.

Why the Central Limit Theorem Matters

It is the reason normal-based statistics work so widely.

• It justifies z-tests. A z-test compares a sample mean with a normal sampling distribution.

• It builds confidence intervals. Intervals of the form $\bar{x} \pm 1.96,SE$ rely on the sample mean being normal.

• It works for any population shape. Analysts do not need to know the population's distribution to reason about averages.

• It quantifies precision. The standard error shows exactly how much a larger sample helps.

OpenStax's Central Limit Theorem for Sample Means section states the theorem in this form, with the sample mean's standard deviation equal to $\sigma/\sqrt{n}$.

What Most Guides Get Wrong About the Central Limit Theorem

Two points are commonly blurred.

The first is what becomes normal. Many explanations say "with enough data, things become normal". The data do not change shape. Only the distribution of the sample mean becomes normal, as Example 6 stresses.

The second is the magic number 30. Guides often treat $n \ge 30$ as a guarantee. It is a rough guide for moderately shaped populations. Strong skew or extreme outliers can require far larger samples, and a normal population needs none at all.

Where Students Trip Up on the Central Limit Theorem

Mistake 1: Using σ instead of the standard error

Where it slips in: Probabilities about sample means, as in Example 2.

Don't do this: Dividing by $\sigma$ when the question is about a sample mean.

The correct way: Divide by $\frac{\sigma}{\sqrt{n}}$.

Students reach for $\sigma$ because it is the standard deviation given in the question, and the shift from individuals to averages is easy to miss in the wording.

Mistake 2: Thinking the data become normal

Where it slips in: Interpreting the theorem, as in Example 6.

Don't do this: Assuming individual values are normal because $n$ is large.

The correct way: Apply the theorem only to the sample mean.

Mistake 3: Expecting the standard error to halve when n doubles

Where it slips in: Planning sample sizes, as in Example 3.

Don't do this: Doubling $n$ to halve the standard error.

The correct way: Quadruple $n$, because the standard error depends on $\sqrt{n}$.

Opinion polls apply the theorem directly. A poll of about 1,000 people has a standard error for a proportion near 50 percent of roughly 1.6 percentage points, which gives the familiar "plus or minus 3 points" margin, and halving that margin would need about 4,000 people.

Key Takeaways

• The central limit theorem says sample means are approximately normal for large random samples.

• The sampling distribution has mean $\mu$ and standard error $\sigma/\sqrt{n}$.

• Larger samples shrink the standard error, but only by the square root of $n$.

• The theorem applies to averages, not to individual values.

• To work through statistics with a teacher, explore Bhanzu's high school math tutor sessions or math classes online.

A Practical Next Step

Roll a die five times and record the mean. Repeat 20 times, or pool results with classmates, and draw a histogram of the means. Compare its spread with that of single rolls, and check it against the standard error $1.71/\sqrt{5} \approx 0.76$.

Practice these to solidify your understanding. If σ still slips in where the standard error belongs, return to Example 2. Want a live Bhanzu trainer to walk through more statistics problems? Book a free demo class, online globally, or in person at our McKinney, TX center.

Read More

• Hypothesis Testing

• Z-Score Table

• Methods of Sampling

• Inferential Statistics

• Standard Deviation