COMBINING INDEPENDENT TESTS IN LINEAR-MODELS

A class of independent linear models is considered, where the variance s of the different models are unequal but they have a common vector pa rameter theta occurring in their mean vectors. For testing the hypothe sis H-0: theta = 0, some exact test procedures are derived that combin e the information from all the models. The procedures are extensions o f the test suggested by Cohen and Sackrowitz for recovering interblock information in balanced incomplete block designs. Simulation results on the power of the new tests and some standard tests are reported in the context of (1) testing the hypothesis concerning the common mean o f two univariate normal populations and (2) recovering interblock info rmation in a block design that is not a balanced incomplete block desi gn. The numerical results indicate that the new tests have excellent p erformance in terms of power compared to some standard tests.