Tids- og kostnadseffektiv sensorikk for forbedret produktutvikling i matindustrien

The RapidCheck project has focused on the development, testing and application of new and faster sensory methods as well as statistical techniques. The results are achieved through combinations of user-oriented courses, research, development and specific applications in industry. Results from the project are published in scientific articles, book chapters as well as lectures and postings. A master student, a PhD student and a post doc have been linked to parts of the project. There has been cooperation with Univ.De la Republica, Uruguay, Univ.Of Texas, Dallas, and Univ.Stellenbosch, South Africa, as well as the FoodSmack and "Children and Taste" (NFR) projects. Course: The focus has been on making the industry better equipped to apply the methods in the project. Some courses and method demonstrations have been corporate, while most have been available to all participants. The topics that have been focused are the most relevant options in the field: projective mapping (PM), polarized sensory positioning (PSP), check-all-that-apply (CATA) and TCATA, as well as the Xlstat program that is important for analyzing data from tests. Relationship Industry/Research: All industrial companies have been active in the project, but with different focus. A long series of studies have been conducted on issues that have been important to industry. Some of these studies have also been very important for method development and method understanding. Particular emphasis has been placed on PM and CATA, which appear to be the methods that will be of greatest importance in industry, at least in the short term. These are techniques that are relatively simple, fast, easy to understand and can be used by non-trained referees. They are thus of great importance for the efficiency of product development, product improvement and quality control in industry. In addition, they are very useful as the basis for internal discussions about the products you are dealing with. These methods are now in active use in several of the companies in the project. There have also been demonstrations of PM and CATA in the SSG (Sensory Study Group) forum with members coming from a number of Norwegian food companies. Research: The research in the project has been linked to the development of sensory and statistical methods as well as the link between them. On the overall level, emphasis has been placed on the following topics: Methods for studying individual differences between judges, better understanding of what the methods actually do and can be used, tools for interpretation, validation of results, stability / reliability of results, new ways to use the methods and segmentation of judges for better understanding and results. The methods that have been emphasized are: PM, PSP and CATA (discussed separately below) but there has also been activity related to just-about-right (JAR) and TCATA (CATA used multiple times in the chewing process). Comparison of methods for achieving better insight into their characteristics has also been central. Projective mapping (PM). This is a method that has been strongly focused on the project, both because of its importance and its simplicity. Both intrinsic (sensory) and extrinsic (packaging, info, etc.) properties of the products have been studied. A new way of using PM based on preference has been developed in collaboration with industry. The method expands the potential and application of PM in both industry and research. A series of studies have also been conducted where focus has been on stability and interpretation potential associated with the use of several so-called components in the statistical methods used. Alternative statistical methods have been tested and it turns out they can give some different interpretation when the number of components is increased from two, which is common to three. This has consequences for the choice of method and points in the direction of caution and study of stability and individual differences. Methods for this have been evaluated. A new method of segmentation of judges in situations where they are very disagreeable, have been developed and tested. The method is particularly useful if groups of judges have different perceptions of the products. This again is important for segmentation of consumer groups later. A study has also been conducted and indicates that the judges' cognitive style may affect how the test is performed. Polarized Sensory Positioning (PSP). This is a relatively new method for analysis of relatively small differences between products. In this project, particular emphasis has been placed on comparing results from this method with results from other methods for gaining insight. There has been work on how many poles one needs and how the poles/reference should be chosen. The main result is that in situations with simple products, there may be only two poles, while in more complicated situations one needs three. A higher number has not given improved results.

Sensorisk analyse spiller en avgjørende rolle i mange stadier i de fleste produktutviklings- og produktforbedringsprosesser. Tradisjonell sensorisk analyse utført ved såkalt sensorisk profilering (QDA) er den beste metoden man kjenner i dag med hensyn på presisjon og den mengde informasjon man kan oppnå. Hovedproblemet med denne metoden er imidlertid at den er tidkrevende, kostbar og må utføres av trenede sensoriske dommere. For mange utviklingsprosjekter i industrien blir QDA derfor for kostnads- og tids krevende å bruke så ofte og på de stedene man har bruk for denne typen informasjon. Det er derfor et stort behov for mer effektive, raskere og brukervennlige metoder som kan anvendes til å identifisere de viktigste sensoriske egenskapene til produktene. Innovasjonen i dette prosjektet vil være utvikling av nye og forbedrede metoder for sensorisk analyse som på en rask og effektiv måte kan skjelne mellom produkter og samtidig gi en enkel og valid beskrivelse av produktforskjeller. Dette vil innebære utv ikling av metoder som også kan brukes av dommere uten lang trening (i noen tilfeller også av forbrukere) samt tilhørende statistisk metodikk for analyse og visualisering av resultater. Hovedgevinsten for industrien av dette prosjektet vil være at sensor isk analyse kan anvendes på mange flere steder og i flere sammenhenger enn det som har vært tilfellet inntil nå. Dette innebærer at man vil ha et kraftfullt sensorisk verktøy tilgjengelige for å bestemme produktkvalitet på alle kritiske kontrollpunkter, t idlig eller sent, hvor denne informasjon er viktig. Demne typen data kan igjen på en enkel måte linkes til data fra standard forbrukertester som fokuserer på forbrukernes preferanser, noe som er av avgjørende betydning før man sender produkter ut på marke det. Et viktig resultat av prosjektet er at man får en raskere, billigere og mer presis produktutvikling. Metodene som utvikles er generiskes og kan brukes både til produktutvikling og til kvalitetskontroll.