Skewness and Kurtosis Calculator

Measure data distribution shape, symmetry, and tail behavior with skewness and kurtosis

About the Skewness and Kurtosis Calculator

The Skewness and Kurtosis Calculator is a specialized statistical tool designed to analyze the shape and tail behavior of a dataset beyond simple measures of central tendency. While mean and median tell you where the center is, and variance tells you how data is spread, skewness and kurtosis reveal the specific architecture of your distribution. This tool is essential for data scientists, financial analysts, and researchers who need to verify if their data meets the assumption of normality required for various parametric statistical tests.

Skewness quantifies the asymmetry of the probability distribution. A positive skew indicates a longer or fatter tail on the right side, while a negative skew indicates a tail stretching to the left. Kurtosis measures the 'tailedness' of the distribution. High kurtosis (leptokurtic) signifies that the data has heavy tails or frequent outliers, whereas low kurtosis (platykurtic) indicates a distribution with thin tails and fewer extreme values. By inputs of raw data points, this calculator provides the precise coefficients needed to characterize these geometric properties, helping users determine if data transformations or non-parametric alternatives are necessary for their analysis.

Formula

Skewness (G1) = [n / ((n-1)(n-2))] * Σ[(xi - x̄) / s]^3 ; Kurtosis (G2) = [n(n+1) / ((n-1)(n-2)(n-3))] * Σ[(xi - x̄) / s]^4 - [3(n-1)^2 / ((n-2)(n-3))]

In these formulas, n represents the sample size, xi is each individual value, x̄ is the sample mean, and s is the sample standard deviation. The skewness formula measures the third standardized moment, highlighting the direction and magnitude of asymmetry. The kurtosis formula calculated here is the 'excess kurtosis' for a sample, which adjusts for bias and centers the result so that a normal distribution equals zero. To find the raw kurtosis, you would omit the subtraction of the final term.

Worked examples

Example 1: A researcher analyzes a perfectly uniform dataset of [10, 20, 30, 40, 50].

1. Calculate Mean (x̄) = 30.
2. Calculate Standard Deviation (s) = 15.81.
3. Calculate Z-scores: -1.26, -0.63, 0, 0.63, 1.26.
4. Sum of cubed Z-scores = 0 (yielding 0 skewness).
5. Sum of Z-scores to the 4th power = 5.06.
6. Apply sample kurtosis adjustment formula resulting in -1.20 excess kurtosis.

Result: Skewness = 0.00, Kurtosis = -1.20. Interpretation: The data is perfectly symmetrical but has thinner tails than a normal distribution.

Common use cases

Assessing stock return distributions to determine the risk of extreme 'black swan' market events.
Checking the normality of residuals in a linear regression model to validate the model's reliability.
Evaluating manufacturing process consistency by checking if output dimensions follow a balanced bell curve.
Identifying ceiling or floor effects in psychology test scores that result in highly skewed data.

Pitfalls and limitations

The formulas for sample skewness and kurtosis require at least three and four data points respectively to be mathematically valid.
Outliers have a disproportionately large impact on both values because the calculations involve raising deviations to the third and fourth powers.
Different software packages may use slightly different variations of these formulas, particularly regarding bias correction for small samples.
A skewness of zero does not guarantee symmetry if the distribution is multimodal.

Frequently asked questions

what does a skewness of 0 mean for my data?

A skewness of zero indicates a perfectly symmetrical distribution, meaning the left and right sides of the mean are mirror images. In real-world datasets, a value between -0.5 and 0.5 is generally considered fairly symmetrical.

why do I need to subtract 3 from kurtosis?

Excess kurtosis is calculated by subtracting 3 from the raw kurtosis value. This adjustment allows you to compare your distribution directly to a normal distribution, which has a kurtosis of 3. Positive excess kurtosis indicates heavy tails, while negative indicates light tails.

is a skewness of 1.5 considered high or low?

If your skewness value is higher than 1 or lower than -1, your data is highly skewed. This usually suggests that the mean is being pulled away from the median by outliers or a long tail, and you might need to use non-parametric tests or data transformations.

can a distribution have high kurtosis but low skewness?

No, skewness and kurtosis are independent measures of shape. Skewness tells you about the symmetry (left vs. right), while kurtosis tells you about the 'peakedness' and the weight of the tails (extreme values) relative to the center.

difference between standard deviation and kurtosis?

Standard deviation measures the spread or scale of the data in the original units. Skewness and kurtosis are dimensionless units that describe the geometry or 'shape' of the distribution profile regardless of the scale.