A new family of methods for data analysis has been recently developed in the United States. With these methods it is possible to explain an ordinal scale (preference, choice) by means of nominal variables (characteristics of the stimuli). The basis of these methods is the research of an optimal quantification of the qualitative variables. At the end of 1974 we had the opportunity to apply them to several actual problems of marketing and media research. In the first case, the problem was to measure the consumer's reaction to price. Several brands are sharing the market of a current and rather common consumer good. It may therefore be supposed that price changes have rather important effects on the sales. In this application, the dependent variable was the consumers' preference among brands x price combinations and the independent variables were the brand and the price. The second case was a packaging problem and the third application has been done on a media problem. We can already draw first conclusions from the application of this new family of methods to actual problems.