ABSTRACT
In the preceding chapters on composite variables and multiple and partial correlations, we considered how a composite variable could be constructed by a process of adding together several component variables. In those cases, a composite variable was not considered given but was said to be dependent for its existence upon an already established set of component variables. In this chapter on the fundamental equations of factor analysis, we will reverse our point of view about what is an independent variable and what is a dependent variable. For example, in factor analysis we may begin with an already given set of variables, such as test scores on a scholastic aptitude test battery, Likert personality rating scales, the prices of several stocks, the measurements of archeological artifacts, or the measurements of the taxonomic characters of several genera of mammals. But factor analysis will lead us to consider such variables as not necessarily fundamental but rather as derivable from some basic set of “unmeasured” or “latent” component variables. Thus factor analysis, at least in the traditional sense, is concerned with the problem of analyzing a variable into components.
