Although a 2 x 2 contingency table can have both, one or no margins fixed, methods of analysis are asymptotically identical, and the "conditional exact" test, when joined with randomization, is UMPU (universally most powerful, unbiased) in each case. However, randomization, which can be absurd as well as arbitrary, is quite impractical. Without randomization, and when at least one margin is free, the conditional exact test is highly conservative and of low power, and indeed irrelevant.For the comparative trial (one margin fixed), a test based on the maximum likelihood estimator of the single nuisance parameter has certain desirable properties, including haVing greater power than the conditional exact test without randomization. True P-values can be obtained from a Fortran program; approximate values from the x 2 statistic, adjusted rather less conservatively than with Yates' correction.With neither margin fixed, the x 2 statistic does not require adjustment.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.