Correlation Coefficient Goodness-of-Fit Test for the Extreme-Value Distribution

Abstract

Abstract An important part of exploratory data analysis is to determine if the data at hand reasonably satisfy the assumptions underlying the statistical test to be performed. Parametric tests include assumptions about the distribution of the data, which are explored using goodness-of-fit tests. One of the easiest goodness-of-fit tests to use is the “correlation coefficient test.” This test requires special tables that are currently readily available only for testing for fit to the Normal distribution. This article presents tables for the Type 1 (Gumbel) extreme-value distribution.