---
What is Kurtosis?
Kurtosis is a statistical measure that describes the shape of a distribution’s tails and peak compared to a normal distribution. While skewness measures the asymmetry of a distribution, kurtosis indicates whether the data are heavy-tailed or light-tailed relative to a normal distribution. In simpler terms, it tells us how much of the data is in the tails and how sharply or flatly the data peaks around the mean.
---
Types of Kurtosis
Kurtosis is mainly categorized into three types:
1. Mesokurtic:
This represents a normal distribution. The peak is moderate, and the tails are neither too heavy nor too light. It has a kurtosis value of 3 (excess kurtosis = 0).
2. Leptokurtic:
A distribution with a sharp peak and fat tails. It has more extreme values (outliers), and the kurtosis is greater than 3 (excess kurtosis > 0). It indicates high concentration around the mean with heavy tails.
3. Platykurtic:
A distribution with a flatter peak and thin tails compared to a normal distribution. It has a kurtosis value less than 3 (excess kurtosis < 0), indicating low central concentration and fewer extreme values.
---
Measures of Kurtosis
Kurtosis is calculated using the fourth central moment of the data.
Formula for Kurtosis:
\text{Kurtosis} = \frac{n \sum (X_i - \bar{X})^4}{(\sum (X_i - \bar{X})^2)^2}
Where:
= mean
= number of observations
Most commonly, the excess kurtosis is used:
\text{Excess Kurtosis} = \text{Kurtosis} - 3
This formula helps to compare the kurtosis of a dataset to that of a normal distribution (which has a kurtosis of 3).
---
Interpretation of Kurtosis Values
Kurtosis Type Value Description
Mesokurtic = 3 Normal distribution, moderate peak
Leptokurtic > 3 Sharp peak, heavy tails, more outliers
Platykurtic < 3 Flat peak, light tails, fewer outliers
---
Importance of Kurtosis
It helps assess the risk of extreme outcomes in data, especially in fields like finance.
Kurtosis is useful in quality control, business forecasting, and statistical modeling.
Understanding kurtosis aids in selecting appropriate analytical techniques and predictive tools.
---
Conclusion
Kurtosis, along with skewness, provides deeper insight into the shape of a data distribution beyond just the average and spread. By identifying how peaked or flat a distribution is, and how likely it is to produce outliers, kurtosis helps analysts and decision-makers understand the reliability, variability, and risk in the data. The use of excess kurtosis simplifies comparison with the standard normal distribution and makes it easier to classify a dataset as leptokurtic, platykurtic, or mesokurtic.
