A method for obtaining optimal designs from the class of variance balanced and connected designs was developed for comparing treatment effects with a pre-assigned variance. The properties of the C-matrix of a block design are employed in developing this method. Some new results concerning the design parameters and the non-zero characteristic root of the C-matrix are also presented.
Thomas, Seemon; Thannippara, Alex; Bagui, S. C.; and Ghosh, D. K.
"Determination of Optimal Block Designs With Pre-assigned Variance For Elementary Contrasts,"
Journal of Modern Applied Statistical Methods:
2, Article 13.
Available at: http://digitalcommons.wayne.edu/jmasm/vol5/iss2/13