Management Information Systems Quarterly
Article Title
Abstract
Constructs and indicators are central to the efforts of many researchers who seek to build and test theories and articulate rich narratives about real-world phenomena. For this reason, an extensive discourse exists about their nature. Increasingly, this discourse has become fraught with controversy. Using Bunge’s (1977, 1979) ontology, I examine the nature of constructs and indicators as they are discussed in the extant literature. I define these concepts precisely, disentangle conceptual from measurement issues, and point to ways that discourse about them could better proceed. I show that unidimensional constructs, multidimensional constructs, dimensions, and indicators are all properties in general of a class of things. I also show that only three types of indicators exist—synonyms of the focal construct and succeeding or preceding properties in a pre-order of properties that includes the focal construct. I examine ontologically the notions of content validity, convergent validity, discriminant validity, and internal-consistency reliability and show their problematic nature. I introduce two new concepts, scope validity and the level of concomitance of indicators, that have rigorous ontological foundations. Together, they provide an improved foundation for assessing the construct validity of a set of indicators.
Erratum
12/6/2021: pp. 1660, 1667, 1668; Equations: updated font