ABSTRACT
People experience and express evaluations, likes and dislikes, for social objects and concepts. Such likes and dislikes are called attitudes, and they may be measured by means of a number of methods available for scaling subjective experience. In other chapters of the present volume, much attention is devoted to scaling problems, and here I simply assume that such scaling is possible. The quantitative details may give rise to problems such as those encountered in comparisons of magnitude and category scales, but the fact that the assumption of unidimensional preference continua usually is supported by data is quite robust (Sjöberg, 1968), although exceptions do exist (Tversky, 1969). In particular, in the empirical work reported here I have used category ratings. A monotonic transformation, approximately exponential, to a magnitude scale would probably not affect the gross trends in data that are at present the focus of interest. This is not to say, of course, that the scaling method is not of crucial importance in other applications.
