What Does Your Kurtosis Result Mean?

Your kurtosis result helps describe how a distribution’s tails compare with a reference distribution, especially in terms of the frequency and extremeness of observations.

For excess kurtosis, a result near zero is commonly compared with the normal distribution. Positive and negative values indicate different tail behavior.

Kurtosis Interpretation

Negative Excess Kurtosis — Platykurtic

A negative excess kurtosis indicates lighter tails relative to the normal distribution. Extreme observations tend to be less prominent compared with a normal reference distribution.

Near-Zero Excess Kurtosis — Mesokurtic

An excess kurtosis close to zero indicates tail behavior that is similar to the normal distribution. This is commonly described as mesokurtic.

Positive Excess Kurtosis — Leptokurtic

A positive excess kurtosis indicates heavier tails relative to the normal distribution. This suggests that extreme observations may occur more frequently or be more influential than they would under a normal reference distribution.

Important: Kurtosis Is Not Simply “Peakedness”

Kurtosis is often described as a measure of how “peaked” a distribution is, but this explanation is incomplete and can be misleading.

For statistical interpretation, kurtosis is better understood in terms of tail heaviness and extreme observations relative to a reference distribution. A distribution can have a different central shape without kurtosis being determined simply by how tall or flat its peak appears.

Also, kurtosis does not tell you whether a distribution is symmetric or skewed. Skewness measures asymmetry, while kurtosis focuses more strongly on tail behavior and extreme values.

Tip: When interpreting your result, always check whether your calculator reports Pearson kurtosis or excess kurtosis. Pearson kurtosis for a normal distribution is 3, while excess kurtosis for a normal distribution is 0.

Kurtosis Distribution

Kurtosis describes how a dataset’s tails behave compared with a reference distribution, especially how strongly extreme observations contribute to the distribution.

The illustration below is a conceptual visualization. It is not a direct plot of your calculator’s raw dataset. Your actual data may have a different shape, and kurtosis does not require the data to form a smooth theoretical curve.

What to Look For

• Dataset distribution: The overall pattern of observations in your data.

• Mean: The arithmetic average of the dataset, which helps locate the center.

• Reference distribution: A theoretical distribution, such as the normal distribution, used for comparison when interpreting excess kurtosis.

• Tail behavior: The amount and influence of observations far from the center. Heavier tails generally indicate a greater presence or influence of extreme observations.

Conceptual Kurtosis Illustration

Important: This illustration is intended to explain the idea of a distribution and its tails. It does not calculate or represent the kurtosis of your specific dataset.

When interpreting your result, focus on tail heaviness and extreme observations rather than judging kurtosis only by the height or sharpness of the center.

How Is Kurtosis Calculated?

Kurtosis is calculated from the distances between each data value and the dataset mean. The calculation gives particular weight to observations far from the mean because the deviations are raised to the fourth power.

Our Kurtosis Calculator uses your entered dataset to perform these calculations automatically. The exact final formula depends on the kurtosis method you select.

Step 1 — Find the Mean

First, calculate the arithmetic mean of the dataset:

xˉ=∑xin\bar{x}=\frac{\sum x_i}{n}

Where:

• xˉ\bar{x} = sample or population mean

• xix_i = each data value

• nn = number of observations

The mean provides the center from which the deviations are measured.

Step 2 — Calculate the Deviations

For each observation, subtract the mean:

xi−xˉx_i-\bar{x}

This shows how far each value is from the center of the dataset.

A negative deviation means the value is below the mean, while a positive deviation means it is above the mean.

Step 3 — Raise Each Deviation to the Fourth Power

Next, calculate the fourth power of each deviation:

(xi−xˉ)4(x_i-\bar{x})^4

Using the fourth power is important because it gives much greater weight to observations that are far from the mean.

For example, a large deviation can contribute substantially more to the kurtosis calculation than a small deviation.

Step 4 — Calculate the Fourth Central Moment

The fourth central moment summarizes the fourth-powered deviations around the mean.

For a population:

μ4=1N∑i=1N(xi−μ)4\mu_4=\frac{1}{N}\sum_{i=1}^{N}(x_i-\mu)^4

The fourth central moment is then compared with the variance raised to the second power.

For a population, the Pearson kurtosis is:

K=μ4σ4K=\frac{\mu_4}{\sigma^4}

This produces Pearson kurtosis, where a normal distribution has a kurtosis of 3.

Step 5 — Apply the Selected Kurtosis Formula

The final calculation depends on the method selected in the calculator.

For population Pearson kurtosis:

K=μ4σ4K=\frac{\mu_4}{\sigma^4}

For population excess kurtosis:

Kexcess=K−3K_{\text{excess}}=K-3

Therefore, a normal distribution has:

• Pearson kurtosis = 3

• Excess kurtosis = 0

For sample kurtosis, an adjusted formula can be used to account for estimating population characteristics from a sample. Your calculator should display the exact formula used for the selected sample method so the result is transparent.

Step 6 — Interpret the Result

Once the calculation is complete, identify the type of result reported by the calculator:

Pearson Kurtosis

Pearson kurtosis uses 3 as the normal-distribution reference point.

Excess Kurtosis

Excess kurtosis subtracts 3 from Pearson kurtosis, making 0 the normal-distribution reference point.

A positive excess kurtosis generally indicates heavier tails relative to the normal reference distribution, while a negative value generally indicates lighter tails.

Important: Kurtosis should not be interpreted simply as a measure of how peaked a distribution looks. It is more closely associated with tail behavior and the influence of extreme observations.

Detailed Kurtosis Calculation Table

The calculator can also show the intermediate calculations for each observation. This makes the result easier to verify and helps you understand how individual values contribute to the final kurtosis.

The values above are illustrative intermediate values. The calculator should replace them dynamically with calculations based on the user’s actual dataset.

What the Calculation Table Shows

Value: The original observation entered by the user.

Deviation: The difference between the observation and the dataset mean.

Squared Deviation: The deviation raised to the second power.

(xi−xˉ)2(x_i-\bar{x})^2

Fourth-Power Deviation: The deviation raised to the fourth power.

(xi−xˉ)4(x_i-\bar{x})^4

The fourth-power column is especially important because observations farther from the mean can have a much larger effect on kurtosis.

Mobile-Friendly Calculation Table

On smaller screens, the detailed calculation table should use horizontal scrolling rather than squeezing or hiding columns. This keeps every calculation visible and prevents the table from becoming difficult to read.

Kurtosis Formula

There are two common ways to report kurtosis: Pearson kurtosis and excess kurtosis. Both describe the same underlying fourth-moment measure, but they use different reference values.

Pearson Kurtosis

For population data, Pearson kurtosis is calculated using the fourth central moment divided by the fourth power of the population standard deviation:

β2=μ4σ4\beta_2=\frac{\mu_4}{\sigma^4}

Where:

• β2\beta_2 = Pearson kurtosis

• μ4\mu_4 = fourth central moment

• σ\sigma = population standard deviation

For a normally distributed population, Pearson kurtosis equals 3.

This value of 3 is the reference point for interpreting Pearson kurtosis.

Excess Kurtosis

Excess kurtosis is calculated by subtracting 3 from Pearson kurtosis:

γ2=β2−3\gamma_2=\beta_2-3

Where:

• γ2\gamma_2 = excess kurtosis

• β2\beta_2 = Pearson kurtosis

Subtracting 3 changes the reference point from 3 to 0.

For a normal distribution:

Pearson vs. Excess Kurtosis

The two measures are directly related:

Excess Kurtosis=Pearson Kurtosis−3\text{Excess Kurtosis}=\text{Pearson Kurtosis}-3

So, if a dataset has a Pearson kurtosis of 4.2, its excess kurtosis is:

4.2−3=1.24.2-3=1.2

If Pearson kurtosis is 2.6, the excess kurtosis is:

2.6−3=−0.42.6-3=-0.4

The underlying distribution has not changed. Only the reporting convention and reference point have changed.

Important: When comparing kurtosis values, always check which convention is being used. A value of 3 is the normal reference for Pearson kurtosis, while a value of 0 is the normal reference for excess kurtosis.

Sample vs Population Kurtosis

Kurtosis can be calculated differently depending on whether your dataset represents an entire population or a sample from a larger population.

This distinction matters because sample kurtosis may use a bias correction that adjusts the result when estimating population characteristics from sample data.

Population Kurtosis

When your dataset contains the entire population of interest, the population formula can be used.

For population Pearson kurtosis:

β2=μ4σ4\beta_2=\frac{\mu_4}{\sigma^4}

Here, the fourth central moment and population standard deviation are calculated directly from all observations in the population.

For population excess kurtosis:

γ2=β2−3\gamma_2=\beta_2-3

A normal population has a Pearson kurtosis of 3 and an excess kurtosis of 0.

Sample Kurtosis

When your dataset is a sample from a larger population, sample kurtosis may use an adjusted formula.

One commonly used bias-corrected sample excess kurtosis estimator is:

G2=n−1(n−2)(n−3)[(n+1)g2+6]G_2= \frac{n-1}{(n-2)(n-3)} \left[(n+1)g_2+6\right]

where:

g2=1n∑i=1n(xi−xˉ)4(1n∑i=1n(xi−xˉ)2)2−3g_2= \frac{\frac{1}{n}\sum_{i=1}^{n}(x_i-\bar{x})^4} {\left(\frac{1}{n}\sum_{i=1}^{n}(x_i-\bar{x})^2\right)^2} -3

and nn is the sample size.

The adjustment is designed to reduce the bias that can occur when sample data are used to estimate a population parameter.

What Is Bias Correction?

A sample does not contain every member of the population. Because of this, statistics calculated from a sample can systematically differ from the corresponding population parameters.

Bias correction modifies the sample kurtosis calculation to provide an adjusted estimate of population excess kurtosis.

This is one reason why the result from a sample-based formula may not match a simple population calculation using the same data values.

Why Can Different Tools Give Different Kurtosis Values?

You may enter the same dataset into a calculator, Excel, a statistics program, and a textbook example and receive different results.

This does not necessarily mean that one calculation is wrong.

Different tools may use different:

• Population or sample assumptions

• Pearson or excess kurtosis conventions

• Bias-correction methods

• Definitions of the sample estimator

• Rounding rules

For example, one tool may report Pearson kurtosis, where a normal distribution has a reference value of 3. Another may report excess kurtosis, where the normal reference is 0.

A sample-based function may also apply a bias correction that a basic population formula does not use.

Calculator vs. Excel vs. Textbook

If your answer does not match another source, check the formula and convention first.

How to Compare Two Kurtosis Results

Before deciding that two answers disagree, compare these four things:

1. Is the dataset treated as a population or a sample?

2. Is the result Pearson kurtosis or excess kurtosis?

3. Does the method apply bias correction?

4. Are the same observations and rounding rules being used?

Once these settings match, results are much easier to compare.

Tip: When reporting kurtosis in an assignment, research paper, or statistical analysis, state the method you used. For example: “Bias-corrected sample excess kurtosis” or “Population Pearson kurtosis.” This makes your result much easier for others to reproduce.

Pearson Kurtosis vs Excess Kurtosis

Pearson kurtosis and excess kurtosis describe the same underlying fourth-moment property, but they use different reference points.

The main difference is simple: excess kurtosis subtracts 3 from Pearson kurtosis.

Pearson Kurtosis

Pearson kurtosis is the traditional kurtosis measure. For a normal distribution, its value is 3.

A Pearson kurtosis above 3 indicates greater tail heaviness relative to the normal reference, while a value below 3 indicates lighter tails.

Excess Kurtosis

Excess kurtosis is Pearson kurtosis with 3 subtracted:

γ2=β2−3\gamma_2=\beta_2-3

This makes the normal distribution’s reference value 0.

A positive excess kurtosis generally indicates heavier tails than the normal reference distribution, while a negative value generally indicates lighter tails.

Why the Two Values Look Different

The two measures are not describing two different properties. They use different scales for reporting the same underlying kurtosis measure.

For example:

Pearson Kurtosis=4.5\text{Pearson Kurtosis}=4.5

Then:

Excess Kurtosis=4.5−3=1.5\text{Excess Kurtosis}=4.5-3=1.5

So 4.5 Pearson kurtosis and 1.5 excess kurtosis represent the same underlying result.

Which One Should You Use?

Use the convention required by your course, textbook, research method, or statistical software.

If you are comparing results from different sources, always check whether they report Pearson kurtosis or excess kurtosis. A normal reference of 3 means Pearson kurtosis, while a normal reference of 0 means excess kurtosis.

Quick rule: Pearson = 3 for normal; Excess = 0 for normal. The difference between the two is always 3.

Kurtosis Example

Let’s calculate kurtosis using a small dataset:

Dataset: 2, 3, 4, 5, 7, 12

For this example, we use population Pearson kurtosis. Pearson kurtosis uses 3 as the reference value for a normal distribution.

Step 1: Find the Mean

Add all values and divide by the number of observations:

xˉ=2+3+4+5+7+126\bar{x}=\frac{2+3+4+5+7+12}{6} xˉ=336=5.5\bar{x}=\frac{33}{6}=5.5

The mean is 5.5.

Step 2: Calculate the Deviations

Subtract the mean from each value:

The deviations show how far each observation is from the mean.

Step 3: Calculate the Fourth Powers

Raise each deviation to the fourth power:

The fourth-power values give more weight to observations that are far from the mean.

Step 4: Find the Fourth Central Moment

Add the fourth-power values and divide by the number of observations:

μ4=1984.3756\mu_4= \frac{1984.375}{6} μ4=330.7292\mu_4=330.7292

The fourth central moment is approximately 330.7292.

Step 5: Calculate the Population Variance

First, calculate the squared deviations:

∑(x−xˉ)2=65.5\sum(x-\bar{x})^2=65.5

Then divide by the number of observations:

σ2=65.56=10.9167\sigma^2=\frac{65.5}{6}=10.9167

The population variance is approximately 10.9167.

Step 6: Calculate Pearson Kurtosis

The population Pearson kurtosis formula is:

β2=μ4σ4\beta_2=\frac{\mu_4}{\sigma^4}

Using the values above:

β2=330.7292(10.9167)2\beta_2= \frac{330.7292}{(10.9167)^2} β2≈2.7752\beta_2\approx2.7752

So, the Pearson kurtosis is approximately 2.775.

Step 7: Calculate Excess Kurtosis

Excess kurtosis is Pearson kurtosis minus 3:

γ2=β2−3\gamma_2=\beta_2-3 γ2=2.7752−3\gamma_2=2.7752-3 γ2≈−0.2248\gamma_2\approx-0.2248

So, the excess kurtosis is approximately -0.225.

Interpretation

The Pearson kurtosis is about 2.775, which is slightly below the normal-reference value of 3.

The excess kurtosis is about -0.225, which is slightly negative. This indicates lighter tails than the normal reference under this kurtosis measure.

The value 12 is relatively far from the mean, so it has a strong effect on the fourth-moment calculation. This example also shows why kurtosis is especially sensitive to observations far from the center.

Final results:

Note: Different sample formulas can produce a different result. Always check whether the calculation uses population or sample kurtosis and whether it reports Pearson or excess kurtosis.

What Does Kurtosis Tell You?

Kurtosis helps describe the tail characteristics of a distribution. It gives information about how strongly a dataset is associated with observations far from the center.

In practical terms, kurtosis can help you understand whether a distribution has relatively heavy or light tails compared with a reference distribution.

Heavy Tails

A distribution with heavier tails has a greater tendency to produce observations far from its center.

High kurtosis relative to the normal reference is often associated with more extreme observations. These observations can have a strong influence on the fourth-moment calculation because deviations are raised to the fourth power.

Light Tails

A distribution with lighter tails has fewer or less extreme observations relative to the reference distribution.

For excess kurtosis, a negative value indicates lighter tails relative to the normal reference. A positive value indicates heavier tails.

Extreme Observations

Kurtosis is particularly sensitive to values far from the mean.

This happens because kurtosis uses the fourth power of deviations. A large deviation can therefore contribute much more to the result than a small deviation.

For example, an observation that is twice as far from the mean has a fourth-power contribution that is:

24=162^4=16

times as large.

This is why unusual or extreme observations can have a noticeable effect on kurtosis.

Distribution Shape

Kurtosis provides information about an important part of distribution shape, especially its tails and extreme observations.

However, kurtosis should not be treated as a simple measure of whether a distribution is tall, flat, or peaked.

Two distributions can have different shapes near their centers while having similar tail behavior. Kurtosis focuses on the fourth-moment characteristics of the distribution, with strong sensitivity to observations far from the mean.

Kurtosis in Statistical Analysis

Kurtosis can be useful when analyzing financial returns, test scores, measurement data, experimental results, and other datasets where unusual observations matter.

When interpreting a kurtosis result, always check which definition is being used:

• Pearson kurtosis: 3 is the normal reference value.

• Excess kurtosis: 0 is the normal reference value.

• Positive excess kurtosis: generally indicates heavier tails than the normal reference.

• Negative excess kurtosis: generally indicates lighter tails than the normal reference.

Kurtosis is most useful when considered together with other descriptive statistics, especially the mean, standard deviation, and skewness.

Kurtosis vs Skewness

Skewness and kurtosis both describe important characteristics of a distribution, but they measure different things.

Skewness focuses mainly on asymmetry, while kurtosis focuses on tail behavior and fourth-moment characteristics.

Skewness

Skewness describes whether a distribution is asymmetric around its center.

• Positive skewness: the distribution has a longer or heavier right tail.

• Negative skewness: the distribution has a longer or heavier left tail.

• Skewness near 0: the distribution is relatively symmetric, although this does not guarantee perfect symmetry.

Kurtosis

Kurtosis describes the fourth-moment characteristics of a distribution and is strongly influenced by observations far from the mean.

It is commonly used to assess whether the tails are relatively heavy or light compared with a reference distribution.

For Pearson kurtosis, the normal reference is 3.

For excess kurtosis, the normal reference is 0.

Why Both Measures Matter

A dataset can have low skewness but still have notable kurtosis. Likewise, a dataset can be highly skewed without having an unusually large kurtosis value.

Using both measures gives you a better description of the distribution than relying on either measure alone.

In simple terms:

• Skewness asks: Is the distribution asymmetric, and in which direction?

• Kurtosis asks: What does the distribution’s tail behavior look like relative to a reference?

Kurtosis vs Standard Deviation

Kurtosis and standard deviation both describe important features of a dataset, but they answer different questions.

Standard deviation tells you how widely values are spread around the mean.

Kurtosis is more sensitive to observations far from the mean and helps describe tail behavior relative to a reference distribution.

A dataset can have a large standard deviation without unusually heavy tails. This is why the two measures should not be treated as interchangeable.

Common Kurtosis Calculation Mistakes

Kurtosis calculations can become confusing when different formulas or definitions are mixed together. Avoid these common mistakes.

1. Confusing Pearson and Excess Kurtosis

Pearson kurtosis and excess kurtosis use different reference values.

• Pearson kurtosis: normal reference = 3

• Excess kurtosis: normal reference = 0

2. Forgetting the 3-Point Difference

Excess kurtosis is Pearson kurtosis minus 3:

Excess Kurtosis=Pearson Kurtosis−3\text{Excess Kurtosis}=\text{Pearson Kurtosis}-3

For example, Pearson kurtosis of 4 means excess kurtosis of 1.

3. Using the Wrong Sample or Population Formula

Population and sample kurtosis do not always use the same estimator.

Check whether your dataset represents the entire population or a sample. Also check whether the method applies a sample bias correction.

4. Confusing Kurtosis With Skewness

Skewness measures asymmetry. Kurtosis focuses on tail characteristics and fourth-moment behavior.

They describe different features of a distribution.

5. Assuming Kurtosis Only Measures Peak Height

Kurtosis should not be treated as simply a measure of how tall or peaked a distribution looks.

It is strongly influenced by observations far from the mean and is commonly interpreted through tail behavior.

6. Comparing Software Without Checking Definitions

Different calculators, spreadsheets, textbooks, and statistical programs may use different kurtosis conventions.

Before comparing results, check:

• Population or sample method

• Pearson or excess kurtosis

• Bias correction

• Formula used

7. Rounding Intermediate Values

Rounding the mean, deviations, or moments too early can slightly change the final result.

Keep more decimal places during the calculation and round the final answer at the end.

8. Misinterpreting Negative Excess Kurtosis

A negative excess kurtosis does not mean the calculation is wrong.

It generally indicates lighter tails than the normal reference under the selected excess-kurtosis definition.

9. Using an Inappropriate Formula

Not every kurtosis formula applies to every situation.

Always identify whether you need a population measure, sample estimator, Pearson kurtosis, or excess kurtosis before calculating.

Checking the definition first can prevent large differences between otherwise correct-looking results.

Want to Check Asymmetry Too?

Use our Skewness Calculator to calculate and interpret the skewness of your dataset.