Local influence in null intercept measurement error regression under a student_t model

Abstract

In this paper we discuss the application of local influence in a measurement error regression model with null intercepts under a Student_t model with dependent populations. The Student_t distribution is a robust alternative to modelling data sets involving errors with longer than Normal tails. We derive the appropriate matrices for assessing the local influence for different perturbation schemes and use real data as an illustration of the usefulness of the application.

Keywords:

• Influence diagnostic
• student_t model
• likelihood displacement
• pretest/post-test data
• measurement error models

Acknowledgements

This research was supported by grants from Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP).