A comparison of iterative methods for obtaining maximum likelihood estimates in contingency tables with a missing diagonal

Abstract

SUMMARY This paper considers maximum likelihood estimation for the parameters of quasiindepen-dence in a contingency table with a miiming or excluded diagonal. Two of the methods discussed derive from the em algorithm of Dempster, Laird & Rubin (1977), the other two methods being Newton-Raphson and iterative scaling. Iterative scaling is found, empirically, to be the least efficient method.