Symposium Highlights

Topics

After reviewing the submitted abstracts, the program committee assigned the 332 accepted papers to the following sessions:

·                      Sensory physiology, psychophysics & scaling

·                      Food choice: sensory & cognitive influences

·                      Statistical techniques in sensory analysis

·                      Sensory evaluation of healthy & non-traditional foods

·                      Sensory texture research

·                      Sensory evaluation & market research

·                      Non-food applications of sensory analysis

·                      Sensory interactions in food

·                      Consumer research: laboratory & field

·                      Sensory - instrumental relationships

·                      Sensory evaluation of novel foods & food processes

·                      Food choice: individual differences & social influences

·                      Descriptive analysis of food products

·                      Cross-cultural influences on food acceptance

·                      New tools & techniques for sensory science

·                      Quality evaluation of shelf-life of foods

Except for three plenary sessions, two or three parallel sessions were conducted simultaneously, and so we could attend only a limited number of papers. I should therefore state that the following remarks are only my personal view, based on the limited number of papers I was able to attend. If a reader liked to know more details, he should look up in the Proceedings, to be published soon.

Tradition

Dr. H. Moskowitz, who worked at the Natick Soldier laboratories in the 60's and 70's, is a well-known sensory scientist by his magnitude estimation. In the keynote speech, he presented how quantitative sensory methods were established by unveiling many episodes of his mentor, Dr. S.S. Stevens, who originated the famous Stevens’ power law that offers the psychophysical basis for the quantitative assessment of food quality. As many Nobel laureates mentioned reminiscently, students of a distinguished researcher can inherit his or her extraordinary way of thinking or behaving, the same way we inherit genetic traits from our parents through the genome.

Scales

The presentation of a new line scale incorporating anchors based on psycophysical observation was remarkable, since no ideal scale for measuring human response against stimuli has been previously created. In a workshop focused on the JAR (just-as-right) scale, every single aspect of the pros and cons of this scale were intensively discussed, but various sensory data clearly demonstrated that JAR is less reliable than an ordinary category or line scale both in precision and accuracy. In fact, many papers presented at this symposium clearly suggested that utilizing a line scale for QDA (quantitative descriptive analysis) and a hedonic scale for consumer testing have been recognized as de facto standards among sensory and consumer scientists.

Preference mapping

By integrating QDA data and preference data collected from consumers, we can obtain a so-called preference mapping that can give us useful information, not only on the mutual relationships between current and new or competing products, but on relationships among attributes, products and consumer preference. In a California winery, flavor properties of wines were modified according to the information obtained from preference mapping and succeeded in an increase in sales amounts. A successful application of preference mapping to marketing tomatoes was presented by a researcher at Matforsk. He applied this approach to interpreting the preference of Norwegian consumers for tomatoes in the context of sensory attributes, in order to select the maturity and cultivar of tomato that most meets consumer preferences. He ultimately showed a great increase in both volume and sales amounts brought by the tomato type selected. A French researcher has invested a great deal to obtain a preference mapping on cars. This high cost research project clearly indicates that this approach can supply valuable information to understand what consumers like.

L-PLS

Nowadays, PLS (partial least squares) regression analysis is widely recognized by most sensory analysts as an indispensable tool to analyze sensory data. Dr. H. Martens, who created the first software package of PLS in the mid 1980's, disclosed the idea and application of the advanced form of PLS designated as L-PLS. Using this new PLS, three different types of information: QDA, preference and demographic data of consumers, can be mutually connected. By combining these three data sets, we can obtain a preference mapping depicting mutual relationships between attributes, preference and consumers’ demographic data, such as age, income and gender. Thus, we can extract a consumer cluster that will accept a new product or we can modify or design a product that can meet the preferences of a target group of consumers. Of course, L-PLS can also analyze other combinations, such as QDA, consumer preference and instrumental data. I admire Dr. Martens’ sustaining mental energy that has allowed him to keep working on PLS algorithms for more than a quarter of a century.

Sensory Science in Japan

As easily assumed, based on the small number of Japanese participants, the current situation of sensory science in Japan is far behind many of the countries that sent a number of people to this symposium. In most American and European countries, the evaluation of food using human senses has long been recognized as an indispensable means of collecting information on sensory properties and consumer preference to design a new product that can be well accepted by consumers. In Japan, in contrast, sensory evaluation has not been recognized as a valuable tool for either development of a new product or basic research on food quality.

Some efficient sensory methodologies, such as QDA, magnitude estimation and a line scale, are precisely described in ISO and ASTM guidelines and have been routinely applied to collecting sensory data on various foods, materials and other products in many countries. However, these methodologies are not described in the JIS (Japanese Industrial Standards) Manual. Consequently, most sensory methodologies, generally recognized as authentic in Japan, are greatly different and far behind from those widely used in European and American countries. In Japan, for example, the old-fashioned paired comparison test and semantic differential method are still being used by both academia and industry, in many food research laboratories. In comparing quantity and quality of information obtained from QDA, with those from the paired comparison test and semantic differential method, the results from the latter two are considerably poorer, and advanced multivariate statistical analysis or chemometric techniques cannot be fully applied to the data collected using these two methods.

The 6^th Symposium and Pangborn Scholarship