Figure 10.7 presents the results of the bootstrap approach. Cartographiesobtained by continuous scale and triadic were very similar and mainlydriven by the global mineralization of waters. Besides, variability (the sizeof ellipses) was comparable for both methods, even if triadic seemed tobetter discriminate some waters (less overlapping of ellipses).Variability remains pretty much the same from 88 to 40 consumers.However, variability largely increases when less than 40 consumers areconsidered, suggesting that for this particular example, a minimum of 40untrained panelists seems necessary for a PSP study.It would be very interesting to study the effect of training on this variability. One could think that training would improve this variability.Considering these results, the two methodologies seem to be equivalent in terms of conclusions about water samples. However, consumersdeclared that the triadic PSP was easier to perform than scale-based PSP.
đang được dịch, vui lòng đợi..
