A panel management is integrated into the tool, both for the expert and consumer type panelists; characteristics of these panelists are followed as their identification (age, sex, ...), their availability, or criteria for consumption or particular purchasing as regards consumers.
Entering these criteria can be done very simply by defining the criteria to observe and to enter it. It is of course also possible to import these elements that have been entered by example in Excel© or to establish a questionnaire dedicated to the loading of these criteria. Other editing features are also included in this panel management, such as sorting and utilities segment for treatment, or to identify what are the panelists who have participated in a particular test or out of sheets of identification.
It is of course possible to track these panelists in terms of attendance at meetings or to identify their
participation frequencies in testing. For example, for discrimination testing, accuracy coefficients complete the
presence coefficients of judges.
The number of tests up to a judge is the maximum number of triangles in this case, a judge could do, if he was
able to visit all the proposed sessions.
| Maximum Number of Tests for a judge: 35 Evaluation results for the selected panelists: Identification of the judge: DUPONT Number of days of testing ............: 15 Number of Tests ...................: 15 Number of tests per test day ..: 1.00 ATTENDANCE …………………: 0.43 ACCURACY ……………: 0.73 TOTAL …………………………: 0.31 |
Evaluations of their place in the panel and their performances are also offered by Tastel and Tastel+. Thus, it is possible to see how the judges are placed in the panel or on a particular descriptor.
Performance evaluations may apply when multiple sessions or multiple repetitions were performed with the same judges, and repeated products.
For example, it is possible to have evidence of repeatability index per panelist and per descriptor:| Vegetable | Ripe note | Mushroom | Fruity note | Oxydised | Sweety note | Sweety | Acid | Astringent | Gritty | |
| Nom_5 | .32 | .43 | .51 | 1.05 | .61 | .29 | .66 | .7 | .51 | .46 |
| Nom_4 | .33 | .35 | .52 | .85 | .97 | .34 | 1.15 | .46 | .48 | .57 |
| Nom_3 | .49 | .39 | .47 | .82 | 1.22 | .72 | 1.01 | .26 | .63 | .77 |
| Nom_2 | .76 | .3 | .47 | .56 | .88 | .38 | .94 | .38 | .73 | .5 |
| Nom_1 | .35 | .61 | .6 | .57 | .64 | .37 | .52 | .07 | .28 | .7 |
| Nom_6 | .2 | .35 | .51 | .83 | .59 | .31 | .62 | .61 | .61 | .42 |
| Nom_7 | .29 | .36 | .47 | .85 | 1.06 | .42 | .99 | .43 | .36 | .57 |
| Nom_8 | .27 | .39 | .15 | 1.07 | 1.18 | .66 | .68 | .43 | .65 | .48 |
| Nom_9 | .39 | .34 | .58 | .48 | 1.36 | .43 | .95 | .41 | .26 | .44 |
| Nom_10 | .38 | .27 | .51 | .65 | .8 | .32 | .55 | .1 | .17 | .62 |
| Panel mean | .37 | .37 | .47 | .77 | .93 | .42 | .8 | .38 | .46 | .55 |
| Vegetable | Ripe note | Mushroom | Fruity note | Oxydised | Sweety note | Sweety | Acid | Astringent | Gritty | |
| Nom_5 | .62 | .37 | .58 | .98 | .62 | .02 | .17 | .34 | .62 | .43 |
| Nom_4 | .81 | .17 | .78 | .25 | .71 | 0 | .9 | .33 | .08 | .69 |
| Nom_3 | .43 | .09 | .37 | .76 | .48 | .82 | .4 | .01 | .69 | .33 |
| Nom_2 | .58 | .04 | .15 | .91 | .58 | .38 | .28 | .02 | .49 | .53 |
| Nom_1 | .82 | .55 | .58 | .93 | .26 | .07 | .17 | 0 | .15 | .41 |
| Nom_6 | .19 | .57 | .58 | .1 | .54 | .09 | .24 | .72 | .47 | .85 |
| Nom_7 | .91 | .16 | .62 | .25 | .4 | .04 | .91 | .08 | .03 | .69 |
| Nom_8 | .21 | .09 | 0 | .98 | .85 | .8 | .07 | .02 | .33 | .05 |
| Nom_9 | .4 | .27 | .31 | .82 | .39 | .45 | .36 | .19 | .09 | .28 |
| Nom_10 | .47 | .1 | .57 | .65 | .53 | .26 | .2 | 0 | .09 | .39 |
| Panel mean | .54 | .24 | .45 | .66 | .53 | .29 | .37 | .17 | .3 | .46 |
In terms of interpretation, for repeatability, we can see that the panelist the more accurate repeatable is the one who has a lowest index, an index above 1 indicates a judge not repeatable. For discrimination too, the lower is the index, the more is the ability to discriminate between products. These initial findings panelist per panelist can very quickly and accurately assess these two main qualities of experts.
Similar results are also available for the entire panel with, for example, evidence of discrimination by the panel per descriptor:
| Vegetable | Ripe note | Mushroom | Fruity note | Oxydised | Sweety note | Sweety | Acid | Astringent | Gritty | |
| Panel | 0 | 0 | 0 | .24 | 0 | 0 | 0 | 0 | 0 | 0 |
| Vegetable | Ripe note | Mushroom | Fruity note | Oxydised | Sweety note | Sweety | Acid | Astringent | Gritty | |
| Panel | .2801 | .8006 | .8228 | .7776 | .9809 | .9742 | .9834 | .4031 | .7506 | .9909 |
In terms of interpretation, for discrimination, the lower is the panel index the more has the panel the capacity to discriminate between products. Regarding consistency, the higher is the value, which corresponds to a probability, the group is homogeneous, ie, more goods are not valued differently according to the judges. For the consensus index, the higher is the value, the better is the consensus of the panel. This last concept is based on the results of a multivariate analysis of means per panelist and per product.
Tastel and Tastel+ can also perform follow up on panel trainings. The principle of this monitoring is during training sessions, to identify judges who are the most sensitive in identifying features developed and therefore known on one or more referent products ; monitoring can then take into account different types of tests to monitor various aspects of the sensitivity of each of the judges on these characteristics. For example, it is mentioned a follow up concerning a session with two triangular tests.
A compilation of results from one or more sessions can then very easily and automatically shows a summary of the panel training phase.
