Calculateus

Kurtosis Calculator

Calculate the sample excess kurtosis of a dataset to measure how heavy-tailed its distribution is.

Result

Excess Kurtosis
0.7141
Distribution Shape
Leptokurtic (heavier tails than normal)

About the Kurtosis Calculator

The Kurtosis Calculator computes the sample excess kurtosis of a dataset, a single number describing how heavy or light a distribution's tails are compared to a normal distribution. It's used to check whether a dataset is prone to extreme, outlier-driven swings, which matters in fields from finance to quality control where tail risk differs from average variability.

How It Works

Enter a list of numbers separated by commas or spaces. The calculator finds the mean and sample standard deviation of the data, standardizes each value, raises it to the fourth power, and combines those fourth-power deviations into a bias-corrected excess kurtosis figure. It then labels the result as leptokurtic (heavier tails), platykurtic (lighter tails), or approximately mesokurtic (normal-like) based on how far the number sits from zero.

Excess Kurtosis = [n(n+1) / ((n-1)(n-2)(n-3))] x sum((x - mean)/sd)^4 - [3(n-1)^2 / ((n-2)(n-3))], using the sample standard deviation (n-1 denominator). A normal distribution has an excess kurtosis of 0 by definition.

Formula & Methodology

By hand, compute the mean and sample standard deviation of the dataset, then for every value subtract the mean, divide by the standard deviation, and raise the result to the fourth power. Sum those fourth powers, multiply by n(n+1) divided by (n-1)(n-2)(n-3), and subtract 3(n-1)^2 divided by (n-2)(n-3). That subtraction is what converts raw kurtosis into excess kurtosis, centered at zero for a normal distribution.

Examples

Everyday Dataset

Ten values (2, 4, 4, 4, 5, 5, 7, 9, 3, 6) have a mean of 4.9 and a sample standard deviation of about 2.025. The calculator returns an excess kurtosis of roughly 0.71, labeling the distribution leptokurtic - modestly heavier-tailed than normal.

Dataset With One Outlier

Twelve values clustered near 8 to 13, plus a single outlier of 50, produce an excess kurtosis of about 11.40 - far above the normal baseline of zero, illustrating how one extreme point can dominate the fourth-power calculation.

Advantages

  • Flags outlier-prone datasets that a simple mean and standard deviation would not reveal on their own.
  • Applies the same bias-corrected formula used in common statistical software, so the result is comparable to published sample kurtosis figures.
  • Translates a raw number into a plain-language shape label (leptokurtic, platykurtic, mesokurtic), making the result interpretable without a stats background.

Common Mistakes

  • Confusing excess kurtosis with plain kurtosis - a normal distribution has plain kurtosis of 3 but excess kurtosis of 0, and mixing the two conventions leads to misreading the result.
  • Treating kurtosis as a measure of how 'peaked' a distribution looks visually, when it's actually driven mainly by the frequency and size of extreme values in the tails.
  • Applying the formula to fewer than 4 data points or to a dataset with no spread, both of which make the statistic undefined.

Edge Cases to Watch For

  • Requires at least 4 numbers; the (n-2) and (n-3) terms in the denominator make the formula undefined below that count, and the calculator returns an error instead.
  • If every value in the dataset is identical, the standard deviation is zero and kurtosis is mathematically undefined, which the calculator flags directly rather than dividing by zero.
  • The shape label uses a 0.5 threshold on excess kurtosis (in either direction) as a practical cutoff for leptokurtic versus platykurtic versus mesokurtic, not a formal statistical test - values close to that threshold should be read as borderline, not decisive.
  • A single extreme outlier can dominate the fourth-power sum and push excess kurtosis sharply positive even when the rest of the data looks unremarkable.

Common Use Cases

  • Analysts screening financial return series for heavy-tail risk before applying models that assume normality.
  • Quality control teams checking whether a manufacturing measurement has more extreme defects than a normal distribution would predict.
  • Students and researchers validating a normality assumption before running a parametric statistical test.
Written & fact-checked by the Calculateus TeamLast updated August 5, 2026How we verify our formulas

Frequently asked questions

What does excess kurtosis measure?

It measures how much of a distribution's variance comes from rare, extreme outliers versus frequent, moderate deviations - a normal distribution has excess kurtosis of 0 by definition, positive values indicate heavier tails and more outlier risk, and negative values indicate lighter tails than normal.

Related calculators

Confidence Interval CalculatorCalculate the confidence interval for a sample mean.Sample Size CalculatorCalculate the minimum sample size needed for a survey or study at a given confidence level and margin of error.Margin of Error CalculatorCalculate the margin of error for a survey result given sample size and confidence level.Correlation Coefficient CalculatorCalculate the Pearson correlation coefficient between two paired sets of numbers.

Conclusion

Excess kurtosis is a quick, well-defined way to flag tail-heavy or tail-light data, but it works best alongside a look at the raw values, since a handful of outliers can dominate the number entirely. Use it as a screening step before deciding whether a normal-distribution assumption is reasonable.