ABSTRACT
In almost all experiments the investigator would like to reduce the number of observations required for a complete factorial. If certain assumptions can be met the use of fractional factorials is a most efficient technique to reduce the number of observations and still obtain the desired information. The usual fractional factorial is still orthogonal, which means that certain effects are estimated independently of one another; however, fewer effects can be estimated and these may be biased. The usual procedure given by many authors including Kempthorne (1952, Chapter 20), Hicks (1973, Chapter 15, Davies (1971, Chapter 10), and Peng (1967, Chapter 7) is the one that will be discussed in this chapter.
