Student's t Distribution is introduced with background including
comments on its robustness properties. The ad hoc procedure of
pretesting data for normality is discussed in the light of current
advice, and previous work into its effectiveness reviewed. The approach
to the problem is outlined: √b1
for test statistic, the Gram-Charlier
distribution for population, approximations using the Johnson system. [Continues.]
A Doctoral Thesis. Submitted in partial fulfilment of the requirements for the award of Doctor of Philosophy at Loughborough University.
Trent Polytechnic, Nottingham. Nottinghamshire County Council.