Calculator
Formula Used
Let n be the number of values, x̄ be the mean, m2 be the second central moment, and m4 be the fourth central moment.
m2 = sum((xi - xbar)^2) / n
m4 = sum((xi - xbar)^4) / n
Pearson kurtosis = m4 / m2^2
Excess kurtosis = Pearson kurtosis - 3
Adjusted excess = ((n - 1) / ((n - 2)(n - 3))) * ((n + 1) * excess + 6)
The adjusted formula is available only when the sample size is greater than three. It is close to the bias corrected form used by common R packages.
How to Use This Calculator
- Paste values as an R vector, comma list, space list, or separate lines.
- Choose whether to ignore extra text and missing labels.
- Select decimal places for clean reporting.
- Press the calculate button to place results above the form.
- Use the CSV or PDF buttons to save the result table.
Example Data Table
| Example | Input | Use Case | Expected Shape Note |
|---|---|---|---|
| Heavy tail sample | c(12, 18, 19, 20, 22, 23, 24, 25, 28, 31, 55) |
Outlier review | Often positive excess |
| Compact values | c(8, 9, 10, 10, 11, 12) |
Small variation check | Often flatter tails |
| Sequence data | 1:15 |
R range test | Uniform style spread |
Article: Understanding Kurtosis in R
Understanding Kurtosis in R
Kurtosis measures how strongly data concentrates around the center and tails. In R, analysts often calculate it while checking distribution shape, model assumptions, or outlier influence. A normal distribution has Pearson kurtosis near three. Its excess kurtosis is near zero. Values above zero suggest heavier tails. Values below zero suggest lighter tails and a flatter peak.
Why Kurtosis Matters
Kurtosis is useful when the mean and standard deviation do not tell the full story. Two datasets can have similar averages, yet show very different tail behavior. Financial returns, quality measurements, exam scores, and sensor readings can all contain rare extreme values. Kurtosis helps reveal those patterns. It should not be used alone. It works best beside histograms, box plots, skewness, and domain knowledge.
R Style Input
This calculator accepts numbers written like an R vector, such as c(12, 18, 21, 25). It also accepts comma, space, or line separated values. Missing text can be ignored when the option is enabled. The tool then parses numeric values and calculates central moments from the cleaned dataset. This makes it practical for quick checks before writing or testing R scripts.
Interpreting Results
Pearson kurtosis reports the fourth standardized moment directly. Excess kurtosis subtracts three, making normal data easier to compare. Adjusted excess kurtosis reduces small sample bias when enough observations are available. When the sample is tiny, results can move sharply after one unusual value. For that reason, this calculator also displays sample size, mean, variance, and standard deviation.
Good Practice
Use kurtosis as a guide, not a verdict. Large positive values may indicate heavy tails, spikes, or extreme observations. Negative values may indicate broad shoulders and shorter tails. Always inspect raw data before removing points. If a value is real, it may be the most important part of the dataset. Report the method used, because R packages may use different definitions. Clear notes help readers reproduce your work and compare results correctly. The same dataset can be tested again after cleaning, grouping, or transforming values. Log transformations often change tail behavior. Standardized reporting makes these changes visible. That habit supports better decisions, cleaner teaching examples, and stronger statistical summaries for practical R based workflows in many reports.
FAQs
What does kurtosis measure?
Kurtosis measures tail weight and peak behavior through the fourth standardized moment. It helps show whether values cluster normally, spread broadly, or include unusual extremes.
What is Pearson kurtosis?
Pearson kurtosis is the direct fourth standardized moment. A normal distribution has Pearson kurtosis near three, though sample data can vary.
What is excess kurtosis?
Excess kurtosis equals Pearson kurtosis minus three. This makes normal data close to zero, which is easier to compare across datasets.
Can I paste an R vector?
Yes. You can paste values like c(4, 8, 9, 12). The calculator also accepts comma, space, and line separated values.
Does the calculator handle 1:10?
Yes. When sequence expansion is checked, simple integer ranges like 1:10 are expanded before calculation.
Why is adjusted kurtosis unavailable sometimes?
The adjusted formula needs more than three values. With fewer values, the denominator is invalid and the estimate is not stable.
Should high kurtosis remove an outlier?
No. High kurtosis can flag extreme observations, but it does not prove an error. Review the source and context first.
Which result should I report?
Report the method required by your course, package, or study. Include whether the value is Pearson, excess, or adjusted excess kurtosis.