×

Optimal crossover designs. (English) Zbl 0911.62073

Ghosh, S. (ed.) et al., Design and analysis of experiments. Amsterdam: North-Holland. Handb. Stat. 13, 63-90 (1996).
This survey on cross-over (repeated measurement) designs represents a broad selection of available results chosen according to the author’s personal interest. The aim is to characterize optimal or efficient designs in situations where repeated measurements are to be performed in order to detect or estimate within subject differences in the effects of treatments. In addition to direct effects also carry-over effects have to be taken into account. The criterion considered here is universal optimality in the sense of J. Kiefer [Surv. Stat. Des. Lin. Models, Int. Symp. Fort Collins 1973, 333-353 (1975; Zbl 0313.62057)] for direct treatment or carry-over effects. In case of uncorrelated errors optimal designs are presented which are uniform in the units and (strongly) balanced in the periods and which can be constructed from suitable orthogonal arrays. Special emphasis is laid on two treatments when construction of optimal designs is considerably simplified. Finally, optimal crossover designs are investigated in the presence of autocorrelated errors.
For the entire collection see [Zbl 0897.00015].

MSC:

62K05 Optimal statistical designs
62K10 Statistical block designs

Citations:

Zbl 0313.62057
PDFBibTeX XMLCite