In the context of pre-employment testing, predictive validity refers to how likely it is for test scores to predict future job performance. Predictive validity is one type of criterion validity, which is a way to validate a test’s correlation with concrete outcomes.
The best way to directly establish predictive validity is to perform a long-term validity study by administering employment tests to job applicants and then seeing if those test scores are correlated with the future job performance of the hired employees.
Predictive validity studies take a long time to complete and require fairly large sample sizes in order to acquire meaningful aggregate data. For this reason, many employers rely on validity generalization to establish predictive validity, by which the validity of a particular test can be generalized to other related jobs and positions based on the testing provider’s pre-established data sets. Alternatively, employers can also perform concurrent validity studies to measure criterion validity; these are done by administering tests to existing employees and comparing results to job performance. Concurrent validity studies are generally much quicker and easier to conduct than predictive validity studies, and they generally do not have the time-range restriction problems often associated with predictive validity studies.