Linear Estimation & Experimental Designs

Covers the general mathematical theory of linear estimation and illustrates how the theory can be applied to the analysis of experimental data. Offers an integrated approach, placing proportionate emphasis on mathematical theory, its application to standard designs, and computational techniques. Provides the reader with a broad experience of the whole field, and instills the facility to apply the techniques to new designs as they are encountered. A special feature is the use throughout of R. C. Bose's approach to the theory of least squares, which completely avoids the use of side-conditions in the solution of normal equations. Chapters cover linear transformations and projections, multivariate normal and associated distributions, linear estimation, sums of squares, tests of linear hypotheses, the completely randomized design, incomplete block design: general theory, the general factorial experiment, the 3n factorial experiment, analysis of covariance, and more.