The Central Limit Theorem
Take samples from almost any distribution, however lopsided or lumpy, average each sample, and those averages form a smooth, predictable bell curve. It is one of the most surprising facts in statistics, and the playground below lets you see it happen.
The setup
Imagine a population: say, the reaction times of every student in a huge psychology study. The shape of that population can be anything. Most real data is not a tidy bell curve; reaction times are usually skewed, with a long tail of slow responses.
Now we do something simple and repeat it over and over:
- Reach into the population and grab a random sample of size n.
- Compute the mean of just that sample: one number.
- Write that number down, and do it all again.
The collection of all those sample means has its own distribution, the sampling distribution of the mean. The Central Limit Theorem tells us what it looks like.
🎮 Sampling Distribution Playground
Pick a population shape, choose a sample size, then draw samples and watch the averages collect below.
① The population you're sampling from
② Your latest sample (orange line = its average)
③ The sampling distribution: every orange average is added here. The dashed curve is the bell curve the CLT predicts.
What you should notice
Try it for a minute and you will see three things. Together they make up the Central Limit Theorem:
1. The averages form a bell curve, even when the population is skewed or lumpy. Switch to the "Skewed" or "Bimodal" population and you'll still get a symmetric bell in panel ③.
2. Bigger samples → tighter bell. Drag n from 1 up to 50 and watch panel ③ get narrower. Averaging more values per sample makes the averages cluster more tightly around the truth.
3. The bell is centered on the population mean. That center is μ, not the middle of the panel: switch to "Skewed" and the whole bell moves left, because that population's mean is about 16, not 50. The dashed μ line moves with it.
The standard error formula
The spread of the sampling distribution has a name, the standard error, and a simple formula (every confidence interval you build uses it):
SE = σ / √n
where σ is the population's standard deviation and n is your sample size. Notice the √n on the bottom: to halve the standard error, you need four times the data. That diminishing return is why the playground's bell tightens quickly at first, then more slowly. Compare the "SE" you measured with the "σ/√n" the theorem predicts. They should match closely once you've drawn enough samples.
The fine print on "almost any"
The theorem needs the population to have a finite variance. Every population you will meet in a psychology study has one. The classic counterexample is the Cauchy distribution, whose tails are so heavy that its variance is infinite. There the theorem fails completely: the average of 1,000 Cauchy values is spread exactly as widely as a single value. In a simulation of 400,000 repeats, the middle half of the averages spanned 2.00 units at n = 1 and 2.02 units at n = 1,000. More data gives no extra precision, because an occasional enormous value is always large enough to pull the average with it.
The condition you actually have to think about is the second one, and it is about n. The theorem describes what happens as n grows without limit, so in practice you need to know whether your n is large enough. The folk answer is n ≥ 30. The table below shows what n = 30 gives on the Skewed population in the playground, an exponential shape with a population mean near 16 and a skewness of 1.84. Each row is 300,000 simulated studies, each computing an ordinary 95% confidence interval for the mean:
| Sample size | Interval covers μ | Entirely below μ | Entirely above μ |
|---|---|---|---|
| n = 10 | 90.2% | 9.4% | 0.4% |
| n = 30 | 92.8% | 6.4% | 0.8% |
| n = 50 | 93.6% | 5.4% | 1.0% |
| n = 100 | 94.3% | 4.4% | 1.3% |
| n = 200 | 94.7% | 3.7% | 1.6% |
At n = 30 the coverage is 92.8%, not 95%, and the misses are lopsided: the interval is entirely below μ eight times for every one time it is entirely above. A right-skewed population has a long thin tail of rare, very large values. A typical sample misses them, so its mean is a little low (52% of samples of 30 are) and its standard deviation is a little small. That makes the interval both too low and too narrow, and both errors push it below μ.
The asymmetry comes from the population's skew, which averaging reduces but does not remove. The skewness of the sampling distribution of the mean is the population's skewness divided by √n: in the simulation it was 0.336 for n = 30, matching the predicted 1.84/√30 = 0.336. So the remaining skew shrinks at the same rate as the standard error, and halving it takes four times the data.
This does not make 30 a bad rule. It is a rule about shape, and the shape does become close to normal quickly, as panel ③ shows. Coverage is a stricter requirement than shape, and a badly skewed population needs a larger n to meet it.
Why it matters: almost everything that comes next, including confidence intervals, t-tests and p-values, relies on sample means behaving this predictably. The CLT is why you can use normal-based methods on real data that are not normal themselves.
Problem 7 of the practice problems takes a clearly right-skewed population of pizza delivery times and asks why a sample of 36 lets you use the normal curve anyway.
Common questions
What sample size is large enough for the Central Limit Theorem?
The folk rule is n ≥ 30, and for mildly skewed data that's usually plenty — sample means settle into a near-normal shape quickly. But it's a rule of thumb, not a law: heavily skewed or outlier-prone populations can need substantially more, while symmetric populations are fine much earlier. When in doubt, look at your data's shape rather than trusting the magic 30.
Does the Central Limit Theorem make my data normally distributed?
This is the classic misreading. The CLT says the distribution of sample means approaches normal as n grows. Your raw data keeps whatever shape it has; skewed data stays skewed no matter how much you collect. The bell curve appears one level up, in the distribution of averages across samples. That distribution is what t-tests and confidence intervals rely on.
My data has a few enormous outliers. Will the CLT still rescue me?
Usually yes, but later than you would like. The theorem needs the population to have a finite variance. Heavy tails with a finite variance (income, some reaction-time distributions, network sizes) make the convergence slow rather than impossible. The sampling distribution of the mean has the population's skewness divided by √n, so a badly skewed population still gives a lopsided sampling distribution at n = 30, and a confidence interval that covers less often than it claims. Two practical steps: check whether the extremes are real data or errors, and prefer a bootstrap interval, which does not assume the sampling distribution is already normal.