A pooled standard deviation calculator combines the spread of two or more groups into a single number. Type the raw values of each group, or just the size and the standard deviation or variance of each, and it returns the pooled standard deviation, the pooled variance and the degrees of freedom.
How to use the pooled standard deviation calculator
- Choose whether each line holds the raw values or the n and s of a group.
- Type one group per line, at least two lines.
- Press Calculate.
What it measures
The pooled standard deviation is a measure of the overall dispersion of a combined data set, found when several groups are assumed to share the same underlying variance [1]. It is a weighted average of the groups' variances, turned back into a standard deviation. You meet it in the two-sample t-test with equal variances and in Cohen's d, where the difference between two means is divided by the pooled standard deviation [2].
The formula for two groups
For two groups the pooled standard deviation is the square root of [(n₁ − 1) s₁² + (n₂ − 1) s₂²] ÷ (n₁ + n₂ − 2) [2]. Here n₁ and n₂ are the sizes of the two groups, and s₁² and s₂² are their sample variances, each calculated with n − 1 in the divisor. The numerator is the sum of the squared deviations of the whole set, and the denominator is the number of degrees of freedom.
The formula for many groups
Wikipedia gives the pooled variance for k samples as the sum of (nᵢ − 1) sᵢ² divided by the sum of (nᵢ − 1), which equals n₁ + n₂ + … + n_k − k [1]. The pooled standard deviation is the square root of this [1]. This calculator accepts as many as 200 groups, and the degrees of freedom are shown in the result.
A worked example
A worked example uses the data sets 5, 7, 9, 11, 13 and 4, 6, 8, 10, 12. Each has n = 5, and the squared deviations from its mean add up to 40, so each sample variance is 40 ÷ 4 = 10. The pooled variance is (4 × 10 + 4 × 10) ÷ 8 = 10, and the pooled standard deviation is √10, about 3.1623. The calculator gives the same result, and also lists each group's mean, standard deviation and variance.
Equal group sizes
If all the groups have the same size, the pooled variance is just the arithmetic mean of the sample variances [1]. In the example above, both variances are 10, so the pooled variance is 10. For groups of different sizes, the larger groups count for more, because each variance is weighted by its degrees of freedom. A group of 20 values has about twice the influence of a group of 10.
Entering summary statistics
You do not need the raw values. Choose the n, s mode and type the size and the standard deviation of each group on one line, for example 10, 2 and 20, 3. The pooled variance is then (9 × 4 + 19 × 9) ÷ 28 = 7.3929, and the pooled standard deviation is about 2.7190. The n, s² mode does the same, with variances on the lines. This is our own worked example, not one from the pages.
When to use it
The pooled value is appropriate when you can assume that the groups come from populations with the same variance, as in a two-sample t-test with equal variances, or as the denominator of Cohen's d [2]. If the spreads of the groups are very different, a pooled value is misleading, and a test that does not assume equal variances is the better choice. Compare the group standard deviations in the result to judge this.
Typing the data
Write one group per line. In the raw values mode, separate the numbers with commas, spaces or semicolons. Each group needs at least two values, since a single value has no sample variance. In the n, s modes, n must be a whole number from 2, and the standard deviation or variance cannot be negative. Lines that do not fit are reported by the calculator, so that you can correct them.
Using the calculator
Pick the mode, type your groups and press Calculate. Read the pooled standard deviation first, then the pooled variance and the degrees of freedom, which equal the total number of values minus the number of groups. Use the list of groups to check that each one was read as you meant, including its own mean and standard deviation. Then carry the pooled value into your test or effect size.
Using the pooled value in Cohen's d
Cohen's d divides the difference between two means by the pooled standard deviation [2]. Take the summary example above, with a pooled standard deviation of about 2.7190, and suppose the two group means are 52 and 48. The difference is 4, so d = 4 ÷ 2.7190 = 1.471. This is our own example. A pooled value that is smaller gives a larger d for the same gap between the means, because the groups overlap less. Note that the calculator gives the pooled spread only. You supply the means yourself and do the final division. You can also read the result the other way round. A d of about 1.5 means that the two means are about one and a half pooled standard deviations apart, which is a large gap in most fields. Always say which pooled value you used.
Common mistakes
- Dividing the group variances by n instead of n − 1. Use the sample variance.
- Averaging the standard deviations. Pool the variances, then take the square root.
- Pooling groups with very different spreads. The equal-variance assumption fails.
- Entering a variance in the standard deviation mode. Check which mode you chose.
What next
Use this often? Add Pooled Stand to your Home screen.
Frequently asked questions
What is a pooled standard deviation calculator?
A tool that combines the variances of two or more groups into one pooled standard deviation.
What is the pooled standard deviation formula?
The square root of [(n₁ − 1) s₁² + (n₂ − 1) s₂²] ÷ (n₁ + n₂ − 2) for two groups.
How do you pool more than two groups?
Divide the sum of (nᵢ − 1) sᵢ² by the sum of (nᵢ − 1), then take the square root.
Why not just average the standard deviations?
Groups of different sizes should count for different amounts, and variances, not standard deviations, are the quantities that add.
When is a pooled standard deviation appropriate?
When you can assume the groups share the same variance, as in an equal-variance two-sample t-test or Cohen's d.
Was this tool helpful?
What went wrong?
We only receive the reason you pick, never what you typed into the tool.
Thanks for your feedback.