Measurement. A blueprint for theory-building in MIS

This paper provides a methodology for validating measures used in MIS research. In the absence of any referent disciplines upon which to base the development of a measure, the appropriate method would be to refine the instrument through a series of validation checks. We describe a procedure that accomplishes this function. The methodology consists of combining three statistical methods, namely, principal components analysis or exploratory factor analysis, multitrait- multimethod analysis, and confirmatory factor analysis. The use of these three conforms to the notion of triangulation in research methods. The user information satisfaction measure is employed to illustrate the application of our methodology. We feel that such cumulative testing gives information system researchers a valuable tool for theory-building.