Optimal design results for 2nfactorial paired comparison experiments

From MaRDI portal
Publication:3683368





The authors have extended the optimal design results of the first author and \textit{R. A. Bradley} [J. Am. Stat. Assoc. 73, 831-839 (1978; Zbl 0405.62054)] from \(2^ 3\) to \(2^ n\) factorial paired comparison experiments. They have considered four classes of partially balanced designs corresponding to four canonical forms of null hypotheses about main effects and interactions. A class D of designs is considered and an optimal design \(d_ i\) is determined maximising the non-centrality parameter \(\lambda^ 2\) in the ith class \(D_ i\). Conditions have been found under which \(d_ i\) is A- , D- and E-optimal.











This page was built for publication: Optimal design results for 2nfactorial paired comparison experiments

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3683368)