ABSTRACT
The early use of direct ratio scaling methods with social, and other nonsensory, variables mainly used number matching. That is, subjects were asked to give magnitude estimations, numbers in proportion to the intensity of their opinions about each of a set of items that were felt to vary along a single dimension. At the outset it should be noted that although many direct ratio methods ask subjects to make ratio judgments, they do not necessarily do so nor are scales that result from the data collected necessarily ratio-level scales. Further assessment of scale properties is needed to determine the level of scale achieved. The scales are called direct ratio scales in this paper because the scaling methods involved request subjects to respond directly in terms of subjective ratios of impressions. Through the work of several investigators, mainly Ekman and Künnapas (1960, 1963), it was learned that the magnitude-estimation and ratio-estimation scales obtained using number responses were related in an invariant manner to other types of scales for the same set of items from the social domain. For example, the category scale was related in a concave-downward fashion to the magnitude scale (Perloe, 1963; Stevens, 1966 b). Also, the pair-comparison scale (Thurstone's Case 5) was approximately a logarithmic function of the magnitude scale based on ratio estimation (Ekman, 1962).
