Simple linear approximations to the likelihood equation for combining evidence in multiple 2×2 tables: A critique of conventional procedures |
| |
Authors: | Takemi Yanagimoto Eiji Yamamoto |
| |
Institution: | (1) Okayama University of Science, Okayama, Japan |
| |
Abstract: | Summary The conventional procedures for a common odds ratio in multiple 2×2 tables are explored and critiqued. Three types of linear
approximation to the likelihood equations under some models of common measures of association are used to derive the popular
conventional estimators and test statistics. Some of them are derived using the model of the common standardized difference
which is an unacceptable measure. The derivation provides us with some characteristics of the procedures. The advantages of
procedures based on the conditional and unconditional likelihoods are discussed.
The Institute of Statistical Mathematics |
| |
Keywords: | Common odds ratio conditional likelihood half integer correction Mantel-Haenszel estimator quarter integer correction |
本文献已被 SpringerLink 等数据库收录! |