Factorial Experiments
摘要
Statistically designed experiments are an important tool in data analysis. The objective of such experimentation is to estimate the effect of each experimental factor on a response variable and to determine how the effect of one factor varies over the levels of other factors. Each measurement or observation is made on an item denoted as an experimental unit. Although some ideas of the varying several factors simultaneously appeared in England in the nineteenth century, the first major systematic discussion on factorial designs was given by Sir Ronald Fisher in his seminal book The Design of Experiments (Chapter 6) in 1935.