A controversy in scaling of subjective states: magnitude estimation versus category rating methods

Measurement issues surrounding the quantification of subjective states such as sensations and attitudes have received increasing attention in nursing in recent years. It has sometimes been argued that magnitude estimation (ME) procedures are superior to category rating (CR) methods for scaling of subjective states which require judgments of intensity or strength of perception. A critique of the two scaling methods is presented. In contrast to arguments first proposed by Stevens and others for the superiority of ME procedures and the ability to attain "direct" ratio level measurement with ME scaling, there is evidence to suggest that ME procedures do not in fact produce ratio level scales. The "characteristic" exponents associated with ME scaling are not necessarily intrinsic to particular subjective states. CR scales neither necessarily constrain nor lack the requisite sensitivity needed to produce valid responses. It is concluded that ME is not necessarily superior to the commonly used CR method for scaling of subjective states in nursing research. Implications of the controversy for selection of scaling methods for nursing research are discussed. Greater attention to the ordinal or rank-order properties of data is encouraged. A small set of data is used to illustrate the importance of rank order properties in scaling subjective states in nursing.