The DFD Method for Analyzing Data in Unreplicated Factorial Designs

The DFD Method for Analyzing Data in Unreplicated Factorial Designs

Unreplicated factorial designs are commonly used in industrials and laboratories. Because of traits of these designs, the analysis of such designs is a pretty échallenging problem. There are a number of methods proposed for analyzing data, in unreplicated factorial designs, most of which are nonparametric. Although there are some advantages h these proposed methods, however, non of them is doing well in terms of power in testing the factorial effect. in overall. In this paper, we propose a parametric method, say double F distribution (DFID), which is able to control the error at IER, or EER, level. This method, like step-up and step-down methods is based on sparsity principle in factorial designs. By DFD method, we first define a test statistic and obtain the exact distribution of the test statistic. Then, we employ the method in analyzing the data and finally it is compared with some current methods through a simulation study.