Semi–Sequential Nonparametric Tests for two Populations with one Sample Size Fixfd

Abstract

The problem or testing the identity of two populations given a sample of fixed size from the first population, is considered . The given sample is used to demarcate a suitable region in the observational space. Observations are then drawn sequentially from the second population. At each step the number of observations lying in the demarcated region is counted and rejection is indicated if this number deviates too much (in the right direction) Otherwise, observation is continued till a prefixed number of observations in the demarcated region is obtained. If rejection is not indicated even at that stage, acceptance is recommended. Different large sample procedures based on this approach are formulated and examined.