Visual EDF Software to Check the Normality Assumption

Maria E. Calzada and Stephen M. Scariano Loyola University New Orleans New Orleans, LA 70118 [email protected] Abstract: A host of materials in an introductory level statistics course relies on the “normality assumption.” However, assessing normality is a very subtle and difficult task, even for expert data analysts. To aid our student’s understanding we have written a program that implements a visual version of the Lilliefors test for normality and the powerful Anderson-Darling test for normality. In this paper we introduce these tests and show their usefulness through examples. The software, which is freely available to colleagues and students at http://www.loyno.edu/~calzada, should be used as a complement to other statistical techniques, like histograms and probability plots. Introduction: The following problem comes from David S. Moore’s The Basic Practice of Statistics [1], page 382: