Listy Biometryczne - Biometrical Letters vol. 42(2), 2005, pp. 143-155
|
OPTIMIZATION OF NESTED STEP DESIGNS
Célia Fernandes1, Paulo Ramos1, Joăo T. Mexia2 1Instituto Superior de Engenharia de Lisboa, Área de Matemática, Secçăo do Departamento de Engenharia de Electrónica e Telecomunicaçőes e de Computadores, Rua Conselheiro Emídio Navarro, 1 - 1950-062 Lisboa Portugal, cfernandes@deetc.isel.ipl.pt, 2Universidade Nova de Lisboa, Faculdade de Cięncias e Tecnologia, Departamento de Matemática, Monte de Caparica 2829-516 Caparica Portugal, jtm@fct.unl.pt |
The number of treatments in a balanced nested design is the product of the number of levels in each factor. This number may be too large. As an alternative, in nested step designs there are as many sub-designs as factors plus one last sub-model used for error estimation. In each sub-design the branching is done only for the corresponding factor. The numbers of treatments is now the sum of the factor levels plus one. Moreover the amount of information for the different factors is more evenly distributed. We present a method to minimize the sum of the estimated variances of the estimators of the variance components.
Random effects, variance components, nested factors, nested step designs, optimization.
