in classical test theory, the supposition that a set of correlated variables can be reduced to a smaller set of underlying dimensions or factors. It follows that each person’s observed score can be represented by a factor loading times a factor plus error variance.