A Statistical Package: An Implementaiton of a Truncated Sequential t-Test

Abstract

This t-test was worked up as a statistical package for the Upjohn Pharmaceutical Company in Kalamazoo, Michigan. It was designed to be used by the researchers in the production area to minimize the time and cost in determining whether or not a new procedure is better than an old one. This test is very well suited to this task, provided that the hypotheses are set up properly. In this presentation, the author hopes to provide a bit of insight into how the test works and how it was implemented. Included are a number of plots, which, it is hoped, will clarify some of the material.