HANOVA for Randomized Block Design with more than One Observation per Cell (I)

 

K. Sen (1988).  HANOVA for Randomized Block Design with more than One Observation per Cell (I). Journal of Statistical Research, Vol. 22, No. 1-2, pp.  ??-??.

 

Abstract

HANOVA is developed in this paper for RBD with more than one observation per cell where the error variances ($ \sigma_i^2 $'s) are assumed to vary with one factor. The model under consideration is the two-way fixed-affect model without interaction. On the assumption that the weights $ w_i=\sigma_i^{-2} $, the reciprocal of the error variances are known the results of WANOVA are cited. When $ \sigma_i^2 $'s are unknown in practice, their estimators are derived either by MINQUE of Rao (1970) or by AUE of Horn et al. (1975) or by IUE of Sen (1984). Using suitable estimators of $ \sigma_i^2 $ in WANOVA-$ \chi^2 $ tests, now test procedures are developed for RBD by using ASD approach.

 

Fulltext PDF not available for this article.