Journal of Statistics: Advances in Theory and ApplicationsAuthor's: Serpil Aktaş and Melike Kaya Bahçecitapar
Pages: [1] - [15]
Received Date: July 2, 2016
Submitted by:
DOI: http://dx.doi.org/10.18642/jsata_7100121683
Quasi-likelihood estimation methods, an alternative to maximum
likelihood estimation of the regression parameters in marginal models,
are mainly used to deal with non-normal multivariate data. The goal of
this paper is to use quasi-likelihood approach as a method for the
analysis of rater agreement plus linear by linear association model,
which is a type of log-linear models of rater agreement in 
contingency tables with ordinal
variables. We discuss the differences between maximum likelihood and
quasi-likelihood approaches and compared the standard errors.
contingency tables, quasi-likelihood, log-linear models, rater agreement.
