Additional Results Concerning Estimable Functions and Generalized Inverse Matrices
S. R. Searle · Journal of the Royal Statistical Society Series B (Statistical Methodology) · 1965
Summary A method is derived for calculating the unbiased estimator of any estimable function of the parameters of a linear model. its variance and its covariance with the estimator of any other estimable function are also obtained. An expression is also given for the F-value required in testing a general linear hypothesis.