We examine maximum likelihood estimation procedures in multilevel models related to two-level hierarchically structured data. Usually, for fixed effects and variance components estimation, multivariate normal distribution is assumed. Here we consider for random effects multivariate exponential power distribution (MEP), which represents one of the possible generalizations of multivariate normal distribution. We examine robustness of maximum likelihood estimators under normal assumption when, indeed, random effects are MEP distributed. The study is conducted through MC simulation procedures.
Solaro, N., Ferrari, P. (2003). On parameters estimation procedures in multilevel models. In Atti del Convegno "Modelli Complessi e Metodi Computazionali Intensivi per la Stima e la Previsione" (pp.392-397). Venezia : Dipartimento di Statistica, Università Ca' Foscari di Venezia.
On parameters estimation procedures in multilevel models
SOLARO, NADIA;
2003
Abstract
We examine maximum likelihood estimation procedures in multilevel models related to two-level hierarchically structured data. Usually, for fixed effects and variance components estimation, multivariate normal distribution is assumed. Here we consider for random effects multivariate exponential power distribution (MEP), which represents one of the possible generalizations of multivariate normal distribution. We examine robustness of maximum likelihood estimators under normal assumption when, indeed, random effects are MEP distributed. The study is conducted through MC simulation procedures.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
