A0378
Title: Restricted maximum likelihood estimation in generalized linear mixed models
Authors: Luca Maestrini - The Australian National University (Australia) [presenting]
Francis Hui - The Australian National University (Australia)
Alan Welsh - the Australian National University (Australia)
Abstract: Restricted maximum likelihood (REML) estimation is a widely accepted and frequently used method for fitting linear mixed models, with its principal advantage being that it produces unbiased estimates of dispersion components. However, the concept of REML does not immediately generalize to the setting of non-normally distributed responses, and it is not always clear the extent to which, either asymptotically or in finite samples, such generalizations reduce the bias of dispersion component estimates compared to standard unrestricted maximum likelihood estimation. We review the various attempts that have been made over the past four decades to develop methods for REML estimation in generalized linear mixed models. We establish four major classes of approaches based on approximate linearization, integrated likelihood, modified profile likelihoods, and direct bias correction of the score function, and show a simulation-based comparison of the approaches.
