ABSTRACT
In any statistical analysis based on the assumption of a particular distribution for the data at hand, one will naturally be interested in assessing the validity of that assumption. More specifically, one will be interested in testing for the hypothesis that the data has come from that specific distribution wherein only the functional form of the distribution is assumed to be known while it may involve some unknown parameters. One of the simple and most commonly used methods for this task is called Quantile-Quantile Plot. One could repeat the Quantile-Quantile plot for various choices of the shape parameter k and determine a reasonable range of values of k for which the assumed distributional family fit the data well. The exact values of the means of order statistics can be used in the construction of the Quantile-Quantile plot as well as in the determination of the correlation goodness-of-fit statistic R.
