Goodness-of-fit tests based on a robust measure of skewness (Q626221)

Goodness of fit tests for an i.i.d. sample \(x_1\),\dots,\(x_n\) are considered based on the medcouple statistics \[ MC(F)=\text{med}_{x_k<m_F<x_j}((x_j-m_F)-(m_F-x_k))(x_j-x_k)^{-1}, \] where \(F\) is the distribution corresponding to \(H_0\), and \(m_F\) is the median of \(F\). MC computed by the left and right halves of the sample is also used. The performance of the tests is investigated by simulations. Applications to sociological and economic data are presented.