If your professor wants to test your “statistical knowledge,” it would be unfair to give you a single question and call that your final exam. After all, “statistical knowledge” covers a lot of ground; one question cannot possibly assess it! This is why test authors use several questions to measure a single subject area. The idea of most objective tests, then, is to create a representative sample of all possible test questions over the material at hand. All of those questions can be cumbersome to work with, so researchers like to “boil them down” into simpler variables.
Think about your college admission test scores. You had one score for reading, one for math, and another for science. You answered hundreds of questions, but the test folks reduced those down into a small number of variables. These new variables are often called factors by researchers. Factor analysis, then, is a method of examining how these many items interrelate. That is a major departure from other methods we have discussed where there was one dependent variable. Factor analysis examines the relationships between many dependent variables. The factors themselves can be seen as the independent variables. Your answer to ten math questions on the SAT may depend on your mathematical intelligence (a factor).
[ Back | Contents | Next ]
Last Modified: 10/10/2018