ABSTRACT
In the previous chapters we have defined a multitude of test procedures for determining whether a sample may or may not have come from a normal distribution. These procedures have varied widely in a number of ways, including objectivity of the method; optimal conditions for use; power; and computational simplicity and, hence, availability for use. While ideally the first three considerations should determine which procedure(s) should be used, if they cannot be easily accessed then they will not be used.
