Validity is the extent to which a test measures or predicts what it is supposed to measure or predict. A valid test is one whose scores can be interpreted as intended.

Content validity concerns whether the items adequately sample the domain of interest. Criterion-related validity concerns whether scores correlate with an external standard, either concurrently or predictively. Construct validity concerns whether the test behaves in ways consistent with theoretical expectations about the underlying construct. Face validity—whether the test looks relevant—is secondary but can affect motivation and acceptance.

Validity is not a fixed property of a test; it is an evidence-based judgment about particular interpretations and uses. A test may be valid for one purpose and invalid for another. Reliability is necessary but not sufficient for validity: consistent scores are of little use if they do not reflect the intended attribute. Ongoing validation research is required as populations and applications change.

A test is valid to the degree that the inferences drawn from its scores are justified.