ABSTRACT
At several points in this book, and particularly in Chapters 4 and 8, the ACV has received some incidental attention. Because its function, the statistical control of irrelevant sources of variation, is frequently necessary in research inference in the behavioral and social sciences, it warrants being treated in a separate chapter. This is particularly true because the conventional ACV, as it is presented in textbooks in experimental design and AV (Edwards, 1972; Hays, 1981; Winer, 1971) is an unnecessarily limited special case of what we will show to be a highly general form of ACV as accomplished by the MRC system as it has been developed in this book.
