A NEW METHOD OF CONSTRUCTION OF 2~3 x 3~3 DESIGN WITH FOUR RANDOMLY SELECTED BLOCKS EACH OF SIZE 12

摘要

In this investigation, we have developed a method for constructing a confounded asymmetrical design of the type 2~3x 3~3 with 4 randomly selected blocks each of size 12 from two replications. Considering different cases, different combination of replications involving simple random sample with and without replacement methods, we have got different block contents. According to Das, the construction of confounded asymmetrical factorial design requires, N =P/R = S~k where S is prime or prime power, K is any positive integer, P is the total number of treatment combination and R is the size of the block in this asymmetrical factorial experiment. But by our new method, it is possible to construct the design when we have, N = P/R≠S~k , simply the product of two primes or prime powers. The achievement of our new method is that, selecting only 4 blocks randomly, we have obtained 50% relative information of the affected effects, taking any two generalized identity relationships.