DETERMINING THE NO-OBSERVED-ADVERSE-EFFECT LEVEL IN CATEGORICAL DATA

We discuss the determination of the no-observedadverse-effect level (NOAEL) from a categorical data. Recently a method which incorpo rated the order regtriction into the multiple testing was proposed in Brown and Erdreich [2]. The test is an exact test and computation ally involved for large samples. Therefore we propose an alternative test which is competitive to their test and is easily used for large samples.