Optimal Row-Column Design For Two Treatments
Correlated errors; Neighbor balance; Row-column design
Optimal row-column design for two treatments is investigated for settings where the observations are positively correlated. Universally optimal designs for generalized least-squares estimation are determined for models with and without row and column effects. In most cases considered, the chessboard design is found to be best. © 2002 Elsevier Science B.V. All rights reserved.
Journal of Statistical Planning and Inference
Number of Pages
Source API URL
Morgan, J. P. and Uddin, Nizam, "Optimal Row-Column Design For Two Treatments" (2003). Scopus Export 2000s. 1662.