NORMALITY TESTING METHODS AND THE IMPORTANCE OF SKEWNESS AND KURTOSIS IN STATISTICAL ANALYSIS
Georges Hatem, Joe Zeidan, Herman Goossens, Carla Moreira
Abstract
The normal distribution (bell curve or Gaussian distribution) is a distribution that happens commonly in many circumstances. Real-life data rarely, if ever, follow a perfect normal distribution. Many tests are useful to test normality and more particularly skewness and kurtosis tests assess the comparability of a given distribution from a normal distribution. These tests are widely used in statistics, business, and epidemiological data including blood pressure, heights, IQ scores and measurement errors. This report provides a review assessing the essential methods employed for testing normality and highlighting the importance of skewness and Kurtosis in statistics. Moreover, it gives some examples of the importance of normality in epidemiological health studies analysis.