Predictive validity

In psychometrics, predictive validity refers to empirical evidence that a scale has predictive power over the unobservable construct that it is intended to measure.

For example, the validity of a cognitive test for job performance is the correlation between test scores and, say, supervisor performance ratings. A valid test would have predictive validity if the test were administered before the employee started work, e.g. during a job interview, and job performance were assessed a year later.