Likelihood inference for small variance components

In this paper, we develop likelihood-based methods for making inferences about the components of variance in a general normal mixed linear model. In particular, we use local asymptotic approximations to construct confidence intervals for the components of variance when the components are close to the boundary of the parameter space. In the process, we explore the question of how to profile the restricted likelihood (REML), show that general REML estimates have a lower probability of being on the boundary than maximum likelihood estimates, and show that the likelihood-ratio test based on the local asymptotic approximation has higher power against local alternatives than the likelihood-ratio test based on the usual chi-squared approximation. We explore the finite sample properties of the proposed intervals by means of a small simulation study