Refers to the ability to use differences in a measure of a construct to predict future events or outcomes. It may be considered a subtype of criterion validity.