Estimation of the Correlation Coefficient for Triangular Contingency Tables under the Bivariate Normal Distribution
Serpil Aktaş · DergiPark (Istanbul University) · 2010
Triangular contingency tables are a special class of incomplete contin-gency tables. Association and independence models are used to analyzesuch tables. Association models can be described in terms of the as-sociation parameters for the analysis of triangular contingency tableshaving ordered categories. The aim of this study is to show the re-lation between the association parameters of the uniform associationmodel and the sample correlation coefficient under the structural ze-ros. For this purpose, a simulation study based on random contingencytables containing structural zeros is performed. Association parame-ters are estimated under the uniform association models. The samplecorrelation coefficients are computed using these parameter estimatesand compared with the population correlation coefficients. It is shownthat by using the association parameter estimates under the uniformassociation model, better estimates can be achieved for the populationcorrelation coefficient in the case of structural zeros.