SOME STATISTICAL PROCEDURES FOR COMBINING INDEPENDENT TESTS

In many applications available data from several independent studies a ddress the same question, and it is essential to have statistical meth ods for combining the results from the different studies. This article addresses this issue in two setups: (1) a testing hypothesis concerni ng the common mean vector of two independent linear models having diff erent variances, and (2) a testing hypothesis concerning a common vari ance component in linear models involving two variance components. The interblock analysis of a balanced incomplete block design (BIBD) is a special case, of (1) when we are interested in testing the equality o f the treatment effects. Testing the significance of the treatment var iance component in a BIBD with random effects is a special case of (2) . We suggest some new test procedures for the testing problems in (1) and (2) and also give a review of the various existing tests. We numer ically compare the powers of the various tests and make specific recom mendations regarding the choice of the test to be used in practical ap plications.