Computation and Use of the Asymptotic Covariance Matrix for Measurement Error Models

摘要

The measurement error model assumes that errors occur in both the response variables and the predictor variables. In using this model, it is of interest to compute confidence regions and intervals for the estimators of the model parameters. An asymptotic form for the covariance matrix is used to construct approximate confidence regions and intervals. The solution of the minimization problem resulting from the use of the measurement error model is discussed, and a procedure for accurately computing the covariance matrix is developed. Then the quality of the confidence regions and intervals constructed from this matrix is assessed via a Monte Carlo study.