Evaluations

The most upstream phase, which consists of collecting the samples, and which will have to represent the products to be evaluated, remains extremely important, even if it is located on a bottling or packaging line on a manufacturing site. All the elements allowing a precise description of the collection conditions must be noted or taken into account in this sampling, products collected during the start-up or end of the process phase may be different and not representative of the entire production.

Products can by nature be very variable, such as the curved part (bottom) or the pointed part (top) of a sausage, the different zones of a ham, the distance from the surface of a cheese, the sides of a fruit, … All these differences must be well known to compare comparable parts or multiply the experiments to evaluate this variability.

One of the main objectives in sensory measurement is to ensure that the subjective information expressed by each panelist can, by the panel and by homogeneous conditions, be transformed into the most objective and repeatable information possible.

Sensory evaluation consists of having panellists to constitute a panel in order to be able to smooth and statistically process the results with tests of hypotheses.

Evaluating in a monadic way (product after product), with few descriptors, and with products that are not very saturating (fruit juices, mineral waters, etc.) can make it possible to evaluate more than ten, or even 20 products in a row, but to be able to evaluate more than 30 samples, even under these optimal conditions, remains excessively difficult by a physiological saturation effect. It is therefore inadvisable, and a fortiori on complex and quite saturating products such as wines, cheeses, chocolates... not to taste more than 30 samples in the same session.

An evaluation session must remain for a moment protected from all external disturbances so as not to induce a change in attention or concentration in the evaluation of a product; this would inevitably cause a perception problem, and therefore a difference in the rating of one product compared to another. Thus, letting people in and out without control is strongly discouraged, and also to prevent information on the products from possibly being communicated between panelists.

An essential point in sensory analysis consists in avoiding any bias that could modify the evaluation by sources other than the actual evaluation of the organoleptic characteristics of the products.

A protocol in incomplete blocks supposes, for a balanced evaluation of all the products, to carry out an evaluation of all the blocks. If this is not done, then the products will not be evaluated identically, and the specific variance analysis treatments cannot be used. On the other hand, provision may be made for a multiplication by a multiple (2 or 3 times, for example) of all the blocks of the protocol in order to increase the reliability of the results by increasing the repetitions, and therefore by reducing the effect of different scoring of judges in the event of a panel with little consensus.

It has been shown in numerous works that the order of presentation and the number of repetitions of a product can influence the evaluation of the products and the ability of a panelist to identify the correct answer.