Abstract
We consider a robust estimator of linear regression for longitudinal data by maximizing marginal likelihood of a scaled t-type error distribution. The marginal likelihood can also be applied to the de-correlated response when the within-subject correlation can be consistently estimated from an initial estimate of the model based on the working assumption of independence. While the t-distributed errors can be motivated from a latent hierarchical model as an extension of Gaussian mixed models, our estimators have asymptotic normal distributions for a wider class of error distributions. The estimators have bounded influence functions and can achieve positive breakdown points regardless of the dimension of the covariates.
Original language | English (US) |
---|---|
Pages (from-to) | 253-269 |
Number of pages | 17 |
Journal | Journal of Statistical Planning and Inference |
Volume | 122 |
Issue number | 1-2 |
DOIs | |
State | Published - May 1 2004 |
Keywords
- Asymptotic normality
- Correlation
- Longitudinal data
- M-estimator
- One-step estimator
- t-type regression
ASJC Scopus subject areas
- Statistics and Probability
- Statistics, Probability and Uncertainty
- Applied Mathematics