Mixed-effects models have found broad applications in various fields. As a result, the interest in learning and using these models is rapidly growing. On the other hand, some of these models, such as the linear mixed models and generalized linear mixed models, are highly parametric, involving distributional assumptions that may not be satisfied in real-life problems. Therefore, it is important, from a practical standpoint, that the methods of inference about these models are robust to violation of model assumptions. Fortunately, there is a full scale of methods currently available that are robust in certain aspects. Learning about these methods is essential for the practice of mixed-effects models.
This research monograph provides a comprehensive account of methods of mixed model analysis that are robust in various aspects, such as to violation of model assumptions, or to outliers. It is suitable as a reference book for a practitioner who uses the mixed-effects models, and a researcher who studies these models. It can also be treated as a graduate text for a course on mixed-effects models and their applications.
Contents: - Introduction
- Generalized Estimating Equations
- Non-Gaussian Linear Mixed Models
- Robust Tests
- Observed Best Prediction
- Model Selection
- Other Topics
Readership: Graduate students and researchers in statistics and biostatistics. And also those who routinely use mixed-effects models in the fields of genetics, medicine, agriculture, education, and surveys.GEE;Mixed Effects;Model Selection;Non-Gaussian;Prediction;Robust Estimation;Tests0
Key Features:- It is the first in looking at robust features of existing methods of mixed model analysis
- It covers a wide range of mixed-effects models, including linear mixed models, generalized linear mixed models, semi-parametric and non-parametric mixed models
- It considers different aspects of mixed model analysis, ranging from estimation to tests, and to prediction and model selection
