ABSTRACT
In this chapter we will study the method based on inverse-normal p-values
(MINP), in which the test statistic at the kth stage Tk is a linear combina-
tion of the weighted inverse-normal of the stagewise p-values. The weight
can be fixed or a function of information time. MINP can be viewed as a
general method including the group sequential method, the Lan-DeMets
error-spending method, the Lehmacher-Wassmer method, the Fisher-Shen
self-design method, and the Cui-Hung-Wang SSR method as special cases.
5.1 Method with Linear Combination of z-Scores