HANOVA for Randomized Block Designs with more than One Observations Per Cell (II)

 

K. Sen (1990). HANOVA for Randomized Block Designs with more than One Observations Per Cell (II). Journal of Statistical Research, Vol. 24, No. 1-2, pp.  85-95.

 

Abstract

Heteroscedastic Analysis of Variance(HANOVA) is dealt with considering Randomized Block Designs or two-way cross classifications with interaction where the error variances ($ \sigma^{2}_i $'s) are assumed to vary with one factor. On the assumption that $ \sigma^{2}_i $'s are known , a Weighted Analysis of Variance (WANOVA) is derived for such layouts. Suitable Unbiased estimator are suggested when $ \sigma^{2}_i $'s are unknown. Using such estimators of $ \sigma^{2}_i $'s Studentized test criteria are obtained from WANOVA-$ \chi^{2} $ tests. Finally new test procedures are developed by using the Asymptotic Series Development (ASD) approach of James (1954) and Sen (1984).

 

Fulltext PDF not available for this article.