A statistical technique known as factor analysis is used to explain variation among associated, observable variables in terms of a conceivably smaller set of unseen variables known as factors. For instance, it's feasible that changes in six variables that have been detected mostly reflect changes in two variables that haven't been measured.