What is Standard Error?
The standard error (SE) measures how much a sample statistic — usually the sample mean — is expected to vary from the true population value if you were to repeat the study many times. In simple terms, it tells you how precise your sample estimate is. A smaller standard error means your sample mean is a more reliable estimate of the population mean.
Standard error is especially important when you want to build confidence intervals or run hypothesis tests. It is the foundation for determining statistical significance.
Standard Error Formula
For the standard error of the mean (SEM), the formula is:
SE = s / √n
where s is the sample standard deviation and n is the sample size.
If you only have the population standard deviation (σ), you can use SE = σ / √n. However, in practice the population SD is rarely known, so the sample SD is used instead.
Difference Between Standard Deviation and Standard Error
| Aspect | Standard Deviation (SD) | Standard Error (SE) |
|---|---|---|
| What it measures | Spread of individual data points around the mean | Precision of the sample mean as an estimate of the population mean |
| Formula | s = √[Σ(xᵢ − x̄)² / (n − 1)] | SE = s / √n |
| Effect of sample size | Stable — does not systematically decrease as n grows | Decreases as n increases (÷ √n) |
| Typical use | Describing data variability | Confidence intervals, hypothesis testing |
In short: SD describes the data; SE describes the estimate. A common mistake is to report SD when SE is needed (or vice versa). Always check which one your audience expects.
Standard Error of the Mean (SEM)
The standard error of the mean is the most common type of standard error. It quantifies how far the sample mean is likely to be from the true population mean. If you were to take many random samples from the same population and calculate the mean of each, the standard deviation of those sample means would be the SEM.
For example, if the SEM is 2.5, it means that if you repeated your study many times, the sample means would typically fall within about ±2.5 units of the true population mean.
Effect of Sample Size
Because the standard error formula divides by the square root of n, increasing your sample size reduces the standard error. However, the relationship is not linear: to cut the SE in half, you need four times the sample size. This diminishing return is a key consideration in study design.
- n = 10 → SE = s / 3.16
- n = 40 → SE = s / 6.32 (about half of n=10)
- n = 160 → SE = s / 12.65 (about half again)
Practical Interpretation
A standard error is often used to construct a confidence interval. For a 95% confidence interval, you would typically use:
CI₉₅ ≈ x̄ ± 1.96 × SE
So if your sample mean is 50 and SE is 2, the 95% CI is roughly 50 ± 3.92, or (46.08, 53.92). This interval is likely to contain the true population mean.
Common Mistakes
- Confusing SD and SE — They answer different questions. SD is about data spread; SE is about estimate precision.
- Using population SD when sample SD is needed — Most real-world data uses the sample formula (n−1 denominator).
- Forgetting to take the square root of n — SE = s / √n, not s / n.
- Applying SE to non-mean statistics without adjustment — The formula SE = s / √n is specifically for the mean. Proportions and other statistics have different SE formulas.
- Reporting SE as if it were SD — This can make results look more precise than they are.
Worked Examples
Example 1: Using SD and Sample Size
Suppose a sample of 25 students has a standard deviation of 10 points on a test. The standard error is:
The sample mean is estimated to be within about ±2 points of the true population mean.
Example 2: Using a Raw Dataset
Dataset: 12, 15, 18, 22, 25, 30, 35, 40, 45, 50
- n = 10
- Mean (x̄) = 29.2
- Sample SD (s) ≈ 13.18
- SE = 13.18 / √10 ≈ 13.18 / 3.162 ≈ 4.17
So the standard error of the mean is approximately 4.17. A 95% CI would be roughly 29.2 ± 1.96 × 4.17 ≈ (21.0, 37.4).
Example 3: Effect of Sample Size
If s = 20 for both samples:
- n = 16 → SE = 20 / 4 = 5.0
- n = 64 → SE = 20 / 8 = 2.5
- n = 256 → SE = 20 / 16 = 1.25
As you can see, quadrupling the sample size halves the standard error.