On Estimation and Influence Diagnostics in Linear Mixed Measurement Error Models

SYNOPTIC ABSTRACT

It is not uncommon that the independent variables of linear mixed models are subject to measurement errors in practice. In this paper, we suggest a method for parameter estimation in linear mixed measurement error models, based on the corrected score function. The proposed parameter estimates are shown to be statistically consistent to the order of n^−½ and normally distributed asymptotically. Diagnostic measures for the model such as the leverage measure, Cook's distance and local influence measures are constructed. A data set on hedonic housing price is analyzed.

Key Words and Phrases:

• case-deletion
• corrected score function
• fixed effects
• influence analysis
• leverage
• local influence
• measurement errors
• random effects