On the Construction and Analysis of Some Confounded Asymmetrical Factorial Designs

Publisher Website
Google Scholar
Abstract
The paper points out the necessity of mutual independence of the estimates of the different confounded interactions in confounded asymmetrical factorial experiments together with the requirement of balance so as to ensure equal relative loss of information of each d. f. of any affected interaction. Some of the designs available in the literature on confounded asymmetrical factorial experiments do not satisfy the important condition of providing mutually independent estimates of all the confounded effects. An example of such a design is the class of designs, p x 3 x 2, p being a prime or a prime power given by Kishen, K. and Srivastava, J.N. (1959) In this paper an alternative method of construction of p x 3 x 2 designs with 6-plot blocks, which provide mutually independent estimates of all the confounded effects, is given together with the method of analysis. The method of construction has also been extended to provide designs of the type p x q x k in blocks of qk plots where p >q>k.

This publication has 1 reference indexed in Scilit:

  • Complex Experiments
    Journal of the Royal Statistical Society Series B: Statistical Methodology, 1935