Likelihood Ratio Testing for Zero Variance Components in Linear Mixed Models
Mixed models are a powerful inferential tool with a wide range of applications including longitudinal studies, hierarchical modeling, and smoothing. Mixed models have become the state of the art for statistical information exchange and correlation modeling. Their popularity has been augmented by the availability of dedicated software, e.g., the Mixed procedure in SAS, the lme function in R and S², or the xtmixed f un c t i on i n STATA.
KeywordsVariance Component Likelihood Ratio Test Linear Mixed Model Null Distribution Nuisance Parameter
Ciprian Crainiceanu's work was supported by NIH Grant AG025553-02 on the Effects of Aging on Sleep Architecture.
- Crainiceanu, C. M. (2003). Ph.D. Thesis: Nonparametric Likelihood Ratio Testing. Cornell UniversityGoogle Scholar
- Fields Development Team (2006). Fields: Tools for Spatial Data. Boulder, CO: National Center for Atmospheric ResearchGoogle Scholar
- Greven, S., C. M. Crainiceanu, H. Küechenhoff, and A. Peters (2008). Likelihood ratio testing for zero variance components in linear mixed models. Journal of Computational and Graphical Statistics Google Scholar
- Zhang, D. and X. Lin (2007). Variance component testing in generalized linear mixed models for longitudinal/clustered data and other related topics. In D. Dunson (Ed.), Model Selection in Linear Mixed Models. Berlin Heidelberg New York: SpringerGoogle Scholar