factor analysis

n. Any of a set of analytic techniques applied to a group of observed variables seeking to discover a smaller set of artificial variables which capture or explain the relatedness of the observed variables. Artificial variables, called factors, are selected; they pass through the densest areas in the multidimensional space created by all the variables and are then rotated, usually to select the ones which collectively account for the most linear covariance, and are most uncorrelated with one another. Some of these procedures can also be used to test whether an a priori factor structure is the best fit.