Scientific Publications with referee

Effect of product knowledge on profiling performance comparing various sensory laboratories.

 

Sensoriske paneler med ulikt nivå av ekspertise og trening ble sammenlignet med hensyn til deres prestasjoner i en sensorisk profileringstest. Trening er viktig for å få et konsistent panel som klarer å skille mellom prøvene og reprodusere bedømmelsen av identiske prøver. Sensorikkdataene ble analysert med to ulike statistiske metoder: Partial least square regression (PLSR) og variansanalyse (ANOVA). Tre ulike paneler profilerte to produktkategorier hver. Signal/støy diagram (PLSR) og p/MSE plott (ANOVA) viste de samme tendensene med hensyn til prestasjonene til hvert enkelt panel. Lave p-verdier og et mye høyere signal enn støyelement indikerte gode prestasjoner i å skille mellom produktene. Men man må huske at disse to metodene er svært forskjellige. p/MSE plotene benyttes hovedsakelig til å finne uteliggere, men kan også være nyttige som i dette studiet, på et panelnivå. Panelene og egenskapene som har lave p-verdier, ligger i en klynge nær X-aksen, mens de ekstreme punktene er tydelig identifiserbare. Signal/støy diagrammene benyttes hovedsakelig til å vise de sensoriske egenskapene som er best modellert. Begge metodene viste imidlertid de samme tendensene.

Facts

Year 2006
Abstract The aim of this study was to determine the best organization/training of sensory panels performing different types of work. Five sensory panels having different types of product knowledge were compared with respect to their profiling performance. Four of the panels were internal panels from various food companies and one was an external panel from a research institute. The panels were categorized in three groups according to their product knowledge: specific product knowledge, unspecific product knowledge, and general product knowledge. Particularly, specific product knowledge from quality control testing was pointed out to be more beneficial for the ability to discriminate between samples, than unspecific product knowledge or general product knowledge from descriptive analysis of a large variety of products. These findings contribute to the explanation that sensory learning evolves from verbal learning and focus of attention rather than perceptual learning. It is most beneficial for food companies with a limited range of products to have panelists that are specifically trained for certain characteristics, rather than have judges that can easily adapt to unfamiliar products and characteristics.
Reference Bitnes, J., Rødbotten, M., Lea, P., Ueland, Ø., Martens, M. 2007. Effect of product knowledge on profiling performance comparing various sensory laboratories. Journal of Sensory Studies, Vol 22, pp 66-80.
Publisher Journal of Sensory Studies,

Related persons

  • Marit Rødbotten

    Sensory Scientist, Nofima Mat | Adj. assoc. Prof., Norw. Univ. of Life Sciences

    Phone: +47 64970170

    Cellphone: +47 975 87 208

  • Per Lea

    Research Scientist, Statistician

    Phone: +47 64970185

  • Øydis Ueland

    Research Director, Consumer and Sensory Sciences

    Phone: +47 64970494

    Cellphone: +47 996 07 621

  • Magni Martens

    Senior Research Scientist / Professor of Sensory Science (h.c.) affiliated to Univ. of Copenhagen, DK

    Phone: +47 64970420

    Cellphone: +47 481 34 856