Construct Validity

Abstract

Content validity reflects the degree to which an empirical measure captures the full range of an attribute’s potential manifestations. Direct measures reflecting the “gold standard” by which a given concept can be gauged are often expensive for researchers to obtain. Construct validity is related to, but distinctive from, two other major forms of validity: content validity and criterion validity. Appropriate measurement of a construct must therefore be inferred using indirect, but observable, indicators of the construct. Researchers sometimes divide criterion validity into two types: concurrent validity, and predictive validity. Whereas criterion validity can, in principle, be objectively established by comparing an indirect measure of a given concept to some gold standard, construct validity can never be definitively established. Using the logic of a multitrait-multimethod matrix, construct validity may be divided into two types: convergent validity and discriminant validity.