Abstract
With a view to the solution of problems concerned with the consumer's acceptance of foods in which saccharose had been substituted, a discussion is opened on psychologic and physiologic aspects of the measurability of sensory stimulus-sensation relations. The classic method for measuring taste is the determination of the taste discrimination threshold which is also termed least discernible difference. In the lat years, the discrimination threshold has been determined and calculated mainly as the mean concentration (stimulus) difference resulting from the distribution of the tasters according to their efficiency (probit analysis). In contrast, the new method, which is described in detail, is based on the dependence of the probability of a perception of difference upon the intensity which is linear and represents the initial course of a stimulus-response curve. Differences in efficiency in the group of tasters, which is assumed to be homogeneous, are included in the variance of difference frequencies found. The control of the course of the test by sequential analysis and the binomial weighting of the probabilities obtained allow to perform calculations of the discrimination threshold which are largely free from systematic errors; this discrimination threshold is defined as the concentration difference that engenders the arbitrarily defined alternative probability of Pa = 0.75. To minimize the risk of overrating the measurement result, it is suggested to calculate in addition, on the basis of the binomial distribution, the confidence limits of the discrimination threshold. Furthermore, the constant stimulus method for determining the perception of two stimulus intensities of different sweetening agents is adapted, as to performance and calculation, to the new aspects in discrimination threshold determination.
Collapse