7.1 The Central Limit Theorem for Sample Means (Averages)

The Central Limit Theorem for Sample Means

The Central Limit Theorem (CLT) explains why the normal distribution shows up so often in statistics: when you take sample means from any population, those means form a distribution that's approximately normal, regardless of the population's original shape. This is what makes it possible to use normal distribution techniques for inference, even when the underlying data is skewed, uniform, or otherwise non-normal.

more resources to help you study

practice questions

The Central Limit Theorem for Sample Means

What the CLT States

The CLT makes three specific claims about the sampling distribution of the sample mean (the distribution you'd get if you took every possible sample of size $n$ and plotted all their means):

1. Shape: As sample size increases, the sampling distribution of $\bar{x}$ approaches a normal distribution, no matter what the population distribution looks like (uniform, skewed, bimodal, etc.).

2. Center: The mean of the sampling distribution equals the population mean: $\mu_{\bar{x}} = \mu$

3. Spread: The standard deviation of the sampling distribution, called the standard error of the mean, equals the population standard deviation divided by the square root of the sample size: $\sigma_{\bar{x}} = \frac{\sigma}{\sqrt{n}}$

A common rule of thumb is that a sample size of $n \geq 30$ is sufficient for the CLT to produce a good normal approximation. However, if the population is already roughly normal, smaller samples work fine. If the population is heavily skewed, you may need a larger $n$.

This theorem is closely related to the law of large numbers, which says that as sample size increases, the sample mean converges to the true population mean. The CLT goes further by describing the shape of the distribution of those sample means.

Why the Standard Error Matters

Notice what the formula $\sigma_{\bar{x}} = \frac{\sigma}{\sqrt{n}}$ tells you:

• The sampling distribution is always less variable than the original population because you're dividing $\sigma$ by $\sqrt{n}$, which is always greater than 1 (for $n > 1$).
• As $n$ increases, the standard error shrinks. This means sample means cluster more tightly around $\mu$ with larger samples.
• Quadrupling the sample size cuts the standard error in half (since $\sqrt{4} = 2$). You get diminishing returns as $n$ grows.

This decreasing spread is why sampling error (the difference between a sample statistic and the population parameter) tends to get smaller with larger samples.

Z-Scores Using Standard Error

When working with individual data points, you calculate a z-score using $\sigma$. When working with sample means, you use the standard error instead. The logic is the same: you're measuring how far a value is from the mean, in units of the relevant standard deviation.

Z-score formula for a sample mean:

$z = \frac{\bar{x} - \mu}{\frac{\sigma}{\sqrt{n}}}$

To calculate it step by step:

1. Find the difference between the sample mean and the population mean: $\bar{x} - \mu$

2. Calculate the standard error: $\frac{\sigma}{\sqrt{n}}$

3. Divide the result from step 1 by the result from step 2

Example: Suppose a population has $\mu = 100$ and $\sigma = 15$. You take a sample of $n = 25$ and get $\bar{x} = 106$.

1. $106 - 100 = 6$

2. $\frac{15}{\sqrt{25}} = \frac{15}{5} = 3$

3. $z = \frac{6}{3} = 2.0$

The sample mean of 106 is 2 standard errors above the population mean. You can now use a z-table or calculator to find the probability of getting a sample mean this extreme.

Statistical Inference and Confidence Intervals

The CLT is what makes most of statistical inference possible. Because sample means are approximately normal, you can:

• Calculate probabilities about where a sample mean will fall
• Construct confidence intervals that give a range of plausible values for the population mean, accounting for sampling variability
• Perform hypothesis tests comparing observed sample means to claimed population values

The width of a confidence interval depends directly on the standard error. Since the standard error decreases with larger $n$, bigger samples produce narrower (more precise) confidence intervals.