A goodness of fit test based on some graphical representation when parameters are estimated (Q1088333)

We consider a graphical representation of data and a goodness of fit test based on the representation when the family of hypothetical distributions includes unknown parameters. The graphical representation is given by a convex polygon which reflects the pattern of order statistics so that we can judge whether the null hypothesis is true or not by visual inspection. The area of our chart is adopted as the test statistic. Both the asymptotic and exact distributions of our test statistic cannot be derived and hence some computer simulations are performed in order to investigate the null distribution and the power properties.