A0501
Title: Linear measurement models with replications in independent variable
Authors: Jia-Ren Tsai - Fu Jen Catholic University (Taiwan) [presenting]
Lei Ting-I - Fu Jen Catholic University (Taiwan)
Abstract: The effect of measurement errors increases with uncertainty in parameter estimates. The focus is investigating independent variables with replicated observations in the linear heteroscedastic measurement errors model. The adjusted least squares method can be extended, and the asymptotic covariance matrix of parameters also be constructed by using the sandwich method and the model-based expectation method based on unbiased estimating equations theory. A confidence interval is also constructed for a single slope parameter; and the joint parameters of the intercept and parameter terms. A Monte Carlo simulation study has been performed to assess the relationship and influence of the average reliability ratio, sample size and replicates under the variation of heterogeneous measurement errors. The application of the proposed method is also illustrated with a real dataset.