Random Coefficient Repeated Measures Model
Published Online: 15 JUL 2005
Copyright © 2005 John Wiley & Sons, Ltd
Encyclopedia of Biostatistics
How to Cite
Goldstein, H. 2005. Random Coefficient Repeated Measures Model. Encyclopedia of Biostatistics. 6.
- Published Online: 15 JUL 2005
Biostatistical data often have a hierarchical structure. Typically, these structures are naturally occurring ones: animal populations are characterized by individuals nested within parents, themselves often nested within groups or herds that may also be nested within spatial entities. In other cases, the structure may result from research designs, as in multicenter clinical trials. Other examples are repeated measure designs, where measurements are “nested” within individual subjects and multivariate response data, where measurements are “nested” within individuals. In addition to nesting relationships among data units, we may also have cross-classifications. For example, an individual cow may be nested within a herd of cattle, but also be the offspring of parent stock where any parent may contribute to several herds: individual cows are thus cross-classified by parents as well as nested within their herds. A further complexity is also often present whereby individual units at one level of a data hierarchy may be nested within more than one higher-level unit as in spatial data.
The article describes models for all these kinds of data structures, increasing in complexity as they move from simple hierarchies with continuously distributed responses, to cross-classifications and multivariate data and to discrete responses. Various extensions and special cases will also be considered. The emphasis is on model specification with a brief section on estimation.
- data hierarchies;
- meta analysis;
- multilevel models;
- multiple membership model;
- random effects model;
- repeated measures;
- survival model