Results evidenced a 75% of the correctness
of the panellist were able to discriminate samples based on
the aroma type (NS and NIS) with a risk error lower than 0.1%,
independently from the structuring agent when candies added
with the highest concentration (0.30%) were tested.