How this instrument works
The Central Limit Theorem is one of the most useful results in statistics: it says that if you repeatedly draw samples of size n from a population — almost any population, regardless of its own shape — and calculate the mean of each sample, those sample means themselves form a distribution that gets closer and closer to a normal (bell-shaped) curve as n grows. This holds even if the underlying population is skewed, lumpy, or nothing like a bell curve at all, which is exactly what makes the theorem so widely applicable.
This instrument computes the two numbers that pin down that sampling distribution's normal shape. The expected value of the sample mean, E[x̄], always equals the population mean μ exactly — on average, across many repeated samples, the sample mean neither overshoots nor undershoots the true population value. The standard error of the sample mean, SE(x̄) = σ/√n, measures how much any single sample mean is likely to differ from μ due to ordinary sampling variability.
The square root in the denominator is the theorem's key practical lesson: precision improves with sample size, but slowly. Quadrupling n only halves SE, because √4 = 2. That is why a pilot study of 25 people has a much wider sampling spread than a full study of 2,500, but going from 2,500 to 10,000 barely tightens things further — you are fighting diminishing returns the whole way.
- Enter the true population mean into Population mean (μ) — the average value across the entire population you are sampling from.
- Enter the population's spread into Population standard deviation (σ) — how much individual values vary around μ.
- Enter your planned or actual sample size into Sample size (n) — how many observations go into each sample mean.
- Read Expected value of the sample mean, E[x̄] — under the theorem this always equals μ exactly, regardless of n.
- Read Standard error of the sample mean, SE(x̄) — the typical spread of sample means around μ for samples of this size.
Worked example — samples of 25 from a population with μ = 100, σ = 15
Enter μ = 100 into Population mean, σ = 15 into Population standard deviation, and n = 25 into Sample size — a population with mean 100 and standard deviation 15 (a classic IQ-scale example), sampled 25 at a time. Expected value of the sample mean, E[x̄], reads 100.0 — exactly the population mean, by the theorem's first guarantee.
Standard error of the sample mean, SE(x̄), reads 3.0, computed as 15 divided by the square root of 25: 15 ⁄ 5 = 3.0. That means if you drew many independent samples of 25 people from this population and averaged each sample, those sample means would typically land within about 3 points of the true mean of 100 — a far tighter spread than the 15-point standard deviation of individual scores, because averaging cancels out much of the individual variability.
Questions
What does the Central Limit Theorem actually guarantee?
It guarantees that the distribution of sample means (not individual data points) approaches a normal distribution as the sample size n grows, centered on the population mean μ with standard error σ/√n — and this holds regardless of the shape of the original population, as long as it has a finite mean and variance. It does not say individual observations become normally distributed; it is specifically a statement about the sampling distribution of the mean.
How large does n need to be for the theorem to apply well?
A common rule of thumb is n ≥ 30, which works well for moderately skewed populations. Populations that are already close to normal need far fewer, sometimes n = 5 or 10 is enough. Heavily skewed or outlier-prone populations can need considerably more than 30 before the sampling distribution of the mean looks convincingly bell-shaped. The theorem is always technically true in the limit as n grows toward infinity; the practical question is only how large n needs to be for the approximation to be good enough.
How is this different from the standard error calculator that uses sample data?
This instrument uses a known population standard deviation σ, entered directly, to compute SE(x̄) = σ/√n — the theoretical sampling distribution for a hypothetical or planned sample. This site's Standard Error of the Mean calculator instead estimates s from an actual data set you provide and computes SE = s/√n, because in most real studies the population σ is unknown and must be estimated from the sample itself. The formulas look identical; the difference is whether σ is known in advance (this theorem) or estimated afterward from data (that instrument) — the same distinction that separates a z-test from a t-test.
How does this relate to the sampling distribution of a proportion?
It is the same Central Limit Theorem applied to a different statistic. This calculator covers the sampling distribution of a sample mean, x̄, computed from numeric data. This site's Sampling Distribution of Sample Proportion calculator covers the sampling distribution of a sample proportion, p̂, computed from yes/no or success/failure data — its standard error formula, SE(p̂) = √(p(1−p)/n), plays the exact structural role that SE(x̄) = σ/√n plays here.
Why does the sample mean's expected value always equal μ exactly?
Because averaging is a linear operation: the expected value of an average of n independent draws from a population equals the average of their individual expected values, which is just μ repeated n times, divided by n — giving μ back exactly. This holds for any sample size n ≥ 1 and any population shape, which is why E[x̄] never depends on n in the formula, only SE(x̄) does.
Does the Central Limit Theorem mean my raw data becomes normally distributed with a big enough sample?
No — that is the theorem's most common misunderstanding. Collecting more raw data does not reshape the population itself; a skewed population stays skewed no matter how many observations you gather. What becomes normal, as n grows, is the distribution of sample means computed from repeated samples — a genuinely different quantity from the raw data's own distribution.