The most well-known procedures in the toolbox of a statistician have their exact inferential optimality properties when sample values come from a normal distribution. When a new methodology is presented, it is usually first tested on the normal distribution. The normal distribution occupies the central place among all distributions in probability and statistics. It is quite common to use a normal distribution as a model for such data. Empirical data on many types of variables across disciplines tend to exhibit unimodality and only a small amount of skewness.
