| Abstract: | The decomposition of the distribution function into skewness and spread functions serves as a basis for conceptualizing the skewness of a random variable. Measures of skewness are also described in terms of the skewness and spread functions, thereby unifying the measures with the concepts. Results that relate these measures to whether one random variable is more skewed than another are reviewed and extended. Graphical displays are presented for uncovering the nature of the skewness of a variable. The measures are also linked to the issue of symmetrizing a variable. |
