Which property describes whether a test actually measures the intended construct?

Which property describes whether a test actually measures the intended construct?

• A. reliability
• B. validity
• C. bias
• D. norms

Answer: (B)

Validity is about whether a test measures what it claims to measure. It answers the question of accuracy: does the test truly reflect the intended construct, such as a specific skill, trait, or ability, and relate to real-world outcomes you care about? A test can be reliable (giving consistent results across time or items) but not valid if it’s measuring something else entirely. So validity is the best descriptor for “measures the intended construct.” Different forms of validity help build the case: content validity checks that the test covers the full domain of the construct; construct validity looks at how the test relates to other measures and theories about the construct; criterion validity considers how well test scores predict relevant outcomes. Norms describe how scores are interpreted against a reference group, and bias refers to systematic errors that affect certain groups; neither of these by themselves determine whether the test measures the intended construct.

Validity is about whether a test measures what it claims to measure. It answers the question of accuracy: does the test truly reflect the intended construct, such as a specific skill, trait, or ability, and relate to real-world outcomes you care about?

A test can be reliable (giving consistent results across time or items) but not valid if it’s measuring something else entirely. So validity is the best descriptor for “measures the intended construct.” Different forms of validity help build the case: content validity checks that the test covers the full domain of the construct; construct validity looks at how the test relates to other measures and theories about the construct; criterion validity considers how well test scores predict relevant outcomes.

Norms describe how scores are interpreted against a reference group, and bias refers to systematic errors that affect certain groups; neither of these by themselves determine whether the test measures the intended construct.