Kurtosis

Calculates how outlier-prone the distribution is.

Interpretation

=0 has outlier character similar to normal.
<0 the lower the flatter the more light-tailed the distribution is.
>0 the higher the Kurtosis value is the more heavy-tailed the distribution is.

Data file

What can cause light-tailed/heavy-tailed data

1. Light-tailed: The data set is only a part of all the data and all the data outside the tolerance borders is filtered.
2. Heavy-tailed: A mix of more distributions.

How statistically deal with light-tailed/heavy-tailed data

See What to do with not normally distributed data.

Formula

Legend

n = n
=Mean

External links

• http://en.wikipedia.org/wiki/Kurtosis
• http://www.itl.nist.gov/div898/handbook/eda/section3/eda35b.htm