Title




Doptimal and Defficient equivalentestimation secondorder splitplot designs
 
Author




 
Abstract




Industrial experiments often involve factors which are hard to change or costly to manipulate and thus make it impossible to use a complete randomization. In such cases, the splitplot design structure is a costefficient alternative that reduces the number of independent settings of the hardtochange factors. In general, model estimation for splitplot designs requires the use of generalized least squares (GLS). However, for some splitplot designs (including not only classical agricultural splitplot designs, but also some secondorder splitplot response surface designs), ordinary least squares (OLS) estimates are equivalent to GLS estimates. These designs are called equivalentestimation designs. As an alternative to these equivalentestimation designs, one can use Doptimal designs which guarantee efficient estimation of the fixed effects of the statistical model that is appropriate given the splitplot structure. We explore the relationship between equivalentestimation and Doptimal splitplot designs for a secondorder response surface model and propose an algorithm for generating Defficient equivalentestimation splitplot designs. This approach allows for a flexible choice of the number of hardtochange factors, the number of easytochange factors, the number of whole plots and the total sample size. 
 
Language




English
 
Source (series)




Research paper / UA, Faculty of Applied Economics ; 2010:011
 
Publication




Antwerp
:
UA
,
2010
 
Volume/pages




19 p.
 
Full text (open access)




 
