ABSTRACT
Canonical correlation is a very general multivariate method that allows multiple independent and multiple dependent variables. In this sense, CC subsumes other methods discussed so far and at one time was considered the penultimate multivariate method. With the development of even more encompassing multivariate methods, such as structural equation modeling, CC has been superceded in its status. Still, CC is an excellent exploratory method when investigating whether two sets of variables are related. It is also a good procedure for highlighting many of the statistical features we see in most multivariate methods. If we understand CC, we are much closer to having an overarching understanding of all the multivariate methods discussed in this book. More description of CC is presented in a number of excellent references (e.g., Campbell & Taylor, 1996; Cohen, Cohen, West, & Aiken, 2003; Fan, 1997; Tabachnick & Fidell, 2001; Takane & Hwang, 2002; Thompson, 2000; Thorndike, 2000).
