Measurement and Instrument Development in the Human Sciences

Abstract

Can measurement provide an independent source of evidence in the human sciences? Recent philosophy of measurement often suggests that it cannot. This essay argues that this conclusion is too strong: even in the human sciences, instruments can sometimes be developed so that the evidence they provide is independent of the substantive theories they in turn put to the test. Psychometric instrument development provides the most challenging case, but the claim survives it. The argument does not deny the roles of models, statistics, judgment, or error; it shows that these need not collapse the distinction between evidence and theory.