Category-Rating and Magnitude Estimation Scaling Techniques

The application of a system of measurement structures for category-rating and magnitude scales is tested with regard to numerous sensory and social judgment scales. Results indicate that, on the average, category-rating scales yield interval scales and multimodality matching scales yield logarithmic interval scales. However, marked interindividual differences are detected, pointing to different capabilities subjects have in coping with both methods. It is asked, therefore, how suboptimal scale quality affects the results of parameter estimation in structural equation models in which judgment scales serve as indicator variables. Based on a general psychophysical judgment model it is shown that the measurement theoretical properties of magnitude scales of individual respondents account for large proportions of the variation of estimated coefficients and of goodness of fit. For category-rating scales the effects scale properties have on parameter estimation cannot be determined because of nonhomogeneous judgment functions.