Abstract
Brys et al. (2003) introduced some new measures of skewness including the medcouple. They compared the performance of the new measures of skewness among themselves as well as with that of the classical measure of skewness. However, they did not consider the recently introduced competitive measures of skewness which are also robust and satisfy the requirements of a skewness measure. This was the motivation to examine the medcouple. Skewness of some data sets including those considered by Brys et al. (2003, 2004) is examined. The paper explores the situations where the medcouple fails to portray the skewness, while other measures depict the skewness correctly. It is also observed that no measure of skewness can portray the skewness correctly in all situations. The skewness measure given by Groeneveld and Meeden (1984) is much simpler than the medcouple, and it portrays the skewness better than the medcouple in most cases. We put forward our remarks on the use of different measures of skewness.