“Content” Still Belongs With “Validity”

Abstract

Murphy’s (2009) main thesis, that the predictive value of a test for job performance is not dependent on the degree to which job content matches test content, is an important statement. But this statement and others in his article tend to provoke a “Yes, but . . .” reaction. Here are the main “buts:” “Job relatedness” is distinguished from “validity” in a way that does not match well with conventional usage in either the psychological or legal spheres. There are some situations where tests are differentiated by content, and content relevance is essential to the testing strategy. Content is important for construct validity, and for the concept of validity in general. Moreover, the issue for many non-psychologist stakeholders is that positive manifold or other correlational evidence, rather than content evidence, is neither necessary nor sufficient for validity. In line with these reservations, the major impact of Murphy’s thesis may be, not on content validity, but on validation transportability and synthetic validity.