‘Kurtosis’ is a term in statistics that refers to the peak probability distribution of random data variables. In typical statistical data, variables and probabilities may have asymmetric distributions. This type of distribution is also called ‘skewed type of distribution’ wherein data variables may lean towards either the right or left directions. There are also instances wherein the data distribution will present with double peaks. In this case, data variables present with two values that represent the most dominant peaks of distribution. It is through kurtosis, that these peaks of data distribution are quantified and/or measured in relation to the mean value of the data variables.
There are three categories wherein peak data distribution may be measured from a given source of data variables. One method or category is called mesokurtic, and this refers to data distribution that has its peak in the same way as normal data distribution. In this category, the peak distribution of data can also be classified as neither very high nor too low. Instead, the peak of data distribution is considered at the baseline. Another category is called leptokurtic, and this occurs when the peak distribution is above the normal or mesokurtic distribution. In the leptokurtic category, the peaks are typically tall and thin while the tails of the data distribution are usually heavy and thick. Data distribution tails may also lean towards either the left or right. One common example of a leptokurtic data distribution is the student’s T-data distribution. The third category is platykurtic, and this is the opposite of a leptokurtic distribution type wherein the peak data distribution is lower than the normal distribution or the mesokurtic type. Basic characteristics of this third type include thin-shaped tails and flat lines along the peak. Because of these flat lines, platykurtic data distributions may also be associated or are referred to as ‘uniform data distribution.’