Vous faites régulièrement des tests consommateurs ? Alors comment choisissez-vous le nombre de panélistes que vous devez recruter ? Sans doute en prenant en compte les résultats attendus, les normes et votre budget.
Comment est-il possible de choisir le bon nombre de panélistes à recruter ?
Mais savez-vous qu’il est possible de vérifier à posteriori quel aurait été le nombre minimal de consommateurs qui vous aurait garanti les mêmes résultats ?
Nous vous proposons donc de découvrir nos travaux sur le sujet, qui peuvent vous permettre d’obtenir une réelle valeur ajoutée pour l’organisation de vos futurs tests consommateurs, lors du prochain Pangborn à Providence.
Voici en avant-première l’abstract accepté à Providence :
“The number of subjects needed for sensory and consumer tests is an important and redundant problematic for most companies using sensory analysis routinely.
Indeed, the cost invested is directly impacted by the size of the panel, both for internal and outsourced studies.
Anyone can find norms and tables in literature proposing guidelines in terms of panel size: for example 100 consumers for a simple hedonic test (AFNOR), 30 subjects for a triangular difference test…
But these advices are finally empirical and “a priori” defined. It cannot really be adapted to the specific objectives and products of each company.
An “a posteriori” approach
In this paper, we propose a systematic “a posteriori” approach. This approach is based on resampling and enabling the optimization of panel sizes, according to specific products and methodologies.
Bootstrapping (Efron, 1979) is in that context a very useful statistical technique. Based on random sampling with replacement, it enables the simulation of thousands “realistic” panels from a single data set. By the way, the stability of the main results (significance of differences and/or similarities between products for example) can be studied while decreasing the size of a panel (e.g. 100, 80, 60…20, 10 subjects).
The paper will propose a methodology where the main criterion is defined as a best compromise between cost and information kept. Indices will be defined for the basic sensory and consumer tests (triangular and other difference tests, similarity tests, hedonic tests). But other indices for other sensory methods (profiling, free sorting task, JAR scales…) will also be proposed.
Examples from cosmetology and food industry fields will be presented, showing that this optimization could help companies getting their own expertise in terms of panel sizes.
Finally, the advantages and limits of this approach will be discussed, as well as the next development considered.”
ACTUALITÉS
Choisir le nombre de panélistes à recruter
Vous faites régulièrement des tests consommateurs ? Alors comment choisissez-vous le nombre de panélistes que vous devez recruter ? Sans doute en prenant en compte les résultats attendus, les normes et votre budget.
Comment est-il possible de choisir le bon nombre de panélistes à recruter ?
Mais savez-vous qu’il est possible de vérifier à posteriori quel aurait été le nombre minimal de consommateurs qui vous aurait garanti les mêmes résultats ?
Nous vous proposons donc de découvrir nos travaux sur le sujet, qui peuvent vous permettre d’obtenir une réelle valeur ajoutée pour l’organisation de vos futurs tests consommateurs, lors du prochain Pangborn à Providence.
Voici en avant-première l’abstract accepté à Providence :
“The number of subjects needed for sensory and consumer tests is an important and redundant problematic for most companies using sensory analysis routinely.
Indeed, the cost invested is directly impacted by the size of the panel, both for internal and outsourced studies.
Anyone can find norms and tables in literature proposing guidelines in terms of panel size: for example 100 consumers for a simple hedonic test (AFNOR), 30 subjects for a triangular difference test…
But these advices are finally empirical and “a priori” defined. It cannot really be adapted to the specific objectives and products of each company.
An “a posteriori” approach
In this paper, we propose a systematic “a posteriori” approach. This approach is based on resampling and enabling the optimization of panel sizes, according to specific products and methodologies.
Bootstrapping (Efron, 1979) is in that context a very useful statistical technique. Based on random sampling with replacement, it enables the simulation of thousands “realistic” panels from a single data set. By the way, the stability of the main results (significance of differences and/or similarities between products for example) can be studied while decreasing the size of a panel (e.g. 100, 80, 60…20, 10 subjects).
The paper will propose a methodology where the main criterion is defined as a best compromise between cost and information kept. Indices will be defined for the basic sensory and consumer tests (triangular and other difference tests, similarity tests, hedonic tests). But other indices for other sensory methods (profiling, free sorting task, JAR scales…) will also be proposed.
Examples from cosmetology and food industry fields will be presented, showing that this optimization could help companies getting their own expertise in terms of panel sizes.
Finally, the advantages and limits of this approach will be discussed, as well as the next development considered.”
Lien vers Les statistiques appliquées chez SensoStat
Vous n’allez pas au Pangborn cette année ? N’hésitez surtout pas à nous contacter pour avoir de plus amples informations sur le sujet !
Il est d’ailleurs possible d’adapter notre méthode à tous types de tests sensoriels. Mais ceci est une autre histoire… ! 😉
Lien vers la vidéo de l’équipe installant le stand SensoStat à Providence
Rechercher
Categories