Criterion Validity

Criterion validity is whether a test's scores relate to a real-world outcome, or criterion, that the test is supposed to predict or correlate with.

Criterion validity is whether a test's scores relate to a real-world outcome, or criterion, that the test is supposed to predict or correlate with.

Where content validity asks whether a test's items adequately represent a construct, criterion validity skips the construct question entirely and asks something more practical: does the score actually connect to something that matters in the world? A conscientiousness test has criterion validity to the extent that higher scores go with things conscientiousness should plausibly relate to — fewer missed deadlines, better academic performance, lower rates of workplace accidents. If it doesn't track any of that, the test's practical usefulness is in question no matter how well-constructed it looks on paper.

Criterion validity splits into two flavors based on timing. Concurrent validity checks the relationship against an outcome measured around the same time as the test. Predictive validity checks it against an outcome that hasn't happened yet, like performance a year down the line. Both are useful for different purposes: concurrent validity is faster to establish, while predictive validity is often the more meaningful claim for things like hiring assessments.

This is one of the more business-relevant forms of validity, since it's the type most directly tied to whether a test is worth using for a real decision, like hiring or clinical screening, rather than just whether it's theoretically sound.

See also

  • Validity Validity is whether a test actually measures the thing it claims to measure, rather than something else entirely.
  • Concurrent Validity Concurrent validity is a form of criterion validity that tests whether a measure correlates with an outcome assessed at roughly the same point in time.
  • Predictive Validity Predictive validity is a form of criterion validity that tests whether a measure's scores forecast an outcome that hasn't happened yet.
  • Incremental Validity Incremental validity is whether a test adds meaningful predictive power beyond what's already captured by measures already in use.
  • Construct Validity Construct validity is the overarching question of whether a test actually measures the theoretical construct it claims to measure, built from evidence across many separate checks rather than one study.
  • Construct A construct is a theoretical concept, like intelligence or extraversion, that can't be observed directly but is inferred from patterns in observable behavior.
  • Convergent Validity Convergent validity is whether a test correlates strongly with other established measures of the same or a closely related construct.