Optimal repeated measurement designs for a model with partial interactions

Pierre Druilhet (1, 2) , Walter Tinsson (3)

1 CREST - Centre de Recherche en Économie et Statistique
2 ENSAI - Ecole Nationale de la Statistique et de l'Analyse de l'Information [Bruz]
3 LMAP - Laboratoire de Mathématiques et de leurs Applications [Pau]

We consider circular crossover designs for models with partial interactions: we assume that the carryover effect of a treatment on itself is different from the carryover effect on other treatments. We generalize Kushner (1997) and Kunert and Martin (2000) methods to find optimal designs when the parameters of interest are total effects, which corresponds to the use of a single treatment, once selected. We give a lower bound of $\Phi_p$ criteria and propose efficient designs of small sizes.

Mots clés: Crossover design; optimal design; universal optimality; total effects

Domaines: Statistiques [math.ST]

Pierre Druilhet, Walter Tinsson. Optimal repeated measurement designs for a model with partial interactions. Biometrika, 2009, 96 (3), pp.677--690. ⟨hal-00221433⟩