ABSTRACT
Analysis of variance is often used as a screening technique to determine whether there is any probable qualitative relationship between the treatments before additional effort and resources are spent in an attempt to develop a quantitative relationship. One advantage of ANOVA over other statistical methods to reveal impact is that the treatments do not have to be represented by continuum-valued numbers. Subsampling occurs when the samples (material collected for observation) are divided into subsamples before testing or analysis. Two-way (and higher) analyses of variance can identify interactions provided that the treatment combinations are replicated. ANOVA just seeks statistical evidence of linear correlation. It is very useful as a screening tool to see if there is a relation between treatments and data, which can guide model/relation/cause-and-effect development efforts. Excel Data Analysis has limited power, but it is very convenient for basic cases.
