Procedures for quantification of data which make no claim to be representational measurement, but rather assign numbers on the basis of arbitrary units, face validity, or intuition. This often happens when sociologists refer to a concept which seems on commonsense or a priori grounds to be important, but which they do not know how to measure directly, so that a measure is then imposed on the concept arbitrarily.
According to Warren Torgerson (Theory and Method of Scaling, 1958), this form of measurement ‘depends on presumed relationships between observations and concepts of interest. Included in this category are the indices and indicants so often used in the behavioural sciences’. Thus, for example, we might be interested in the concept of socio-economic status, and so decide to measure perceptions of occupational prestige, which we presume to be related to it; or, similarly, we may be interested in the concept of ‘political awareness’ and decide to measure it by scaling such attributes as the ability to name prominent politicians, or propensity to vote at national and local elections. Rating scales are another obvious example of what is the most common form of measurement in the social sciences.
References to, or accusations of, measurement by fiat usually occur in the context of wider philosophical discussions of the epistemological basis of the social sciences. Radical critics of so-called positivist approaches are prone to use the term freely—as, for example, in the ethnomethodological critique of sociological theories and methods offered by Aaron Cicourel (Method and Measurement in Sociology, 1964).