Abstract

A test statistic following the work of James (1954) has been discussed for pooled estimation and test of treatment contrasts applied on a group of randomized block design experiments with the same set of treatments in presence of heterogeneous error variances. The distribution of the suggested test statistic has been studied by the Monte Carlo study and it was observed that its distribution did not follow the suggestion of James (1954) and Bhuyan (1986) for both large and small error d. f . of individual experiments. The critical values of the test statistic along with some other distributional characteristics that were found by simulation and numerical illustrations are also included.

Author: A. A. JAMJOOM AND M. A. ALI

Received on: July, 2011

Accepted on: November, 2011