On the efficiency of some supplemented (α1, α2, ..., αR)-resolvable block designs

Statistical properties of designs with two kinds of treatments: basic and supplementary, are examined. The basic treatments are arranged randomly in an (alpha(1), alpha(2), . . . , alpha(R))-resolvable block design. This basic design is orthogonally supplemented by some orthogonal addition of the supplementary treatments. Mixed linear models of observations following two- or three-step randomizations are considered. The final design under these models is generally balanced and that allows obtaining its stratum efficiency factors for both cases. (C) 2002 Elsevier Science B.V. All rights reserved.