The conceptualization of validity: Current perspectives

Abstract Virtually all recent discussions of validity in the sociological literature deal with techniques for validity assessment, rather than with clarification of the concept of validity. These new techniques have successfully utilized statistical methods, principally factor analysis and path analysis, in developing sophisticated and rigorous techniques for validity assessment. Unfortunately, however, a great deal of confusion surrounds the basic definition of validity. Thus, the conceptual underpinning lacks the rigor and precision of the statistical techniques. Unfortunately, the statistical rigor of the assessment techniques will not compensate for the conceptual weakness of validity. The present paper attempts to rectify this by clarifying the notion of validity. It is shown that much of the confusion in the extant literature stems from the fact that sociologists virtually always utilize a two-level measurement model (e.g., conceptual/empirical or conceptual/operational), when the actual process of measurement necessarily utilizes three levels (the conceptual, empirical, and operational). A three-level model is proposed which lessens the confusion surrounding validity and which shows that three types of validity exist—Type a validity, Type B validity, and Type C validity. Utilization of these three types greatly clarifies the concept of validity.