Listy Biometryczne - Biometrical Letters Vol. 33(1996), No. 1, 1-23
The paper reconsiders the existence of the BLUEs under a randomization model for the randomized block design, questioned by Kaiser (1989; J. Statist. Plann. Inference 22, 63-69). It is shown how two different randomization model for this design imply the existence, under certain conditions, of the BLUEs of the usual linear parametric functions of interest. The role of the double randomization, within and between blocks, is indicated. Finally, the reason for the negative conclusion drawn in Kaiser's paper is discussed.
best linear unbiased estimation, randomization models, randomization block design