Sensory Product and Concept Testing: Analyses, Applications and Computer Workshop
March 15-19, 2010

Monday - March 15 (8:00 a.m. - 4:00 p.m.)

Topics
♦ Difference testing methods: m-AFC, triangle, duo-trio
♦ Thurstonian theory underlying all sensory evaluation methods
♦ Estimating a measure of sensory difference, d´ and its variance from discrimination tests
♦ Proportion of discriminators in the population
♦ The method of tetrads
♦ Same-different and degree of difference methods

Cases
♦ Product differences using m-AFC tests
♦ Ingredient supplier change: Texture using 2-AFC, duo-trio, and triangle; The issue of power
♦ Superiority claim: 2-AFC vs. duo-trio
♦ 2-AFC and 2-AC on carbonated water
♦ Review of some cases for the NAD
♦ Evaluating an equivalence claim

Tuesday - March 16 (8:00 a.m. - 4:00 p.m.)

Topics
♦ Power and sample sizes for discrimination methods
♦ Risk management in product testing
♦ Setting action standards
♦ Replicative testing: How to increase power and reduce costs using the beta-binomial model
♦ New model for replicated same-different
♦ Torgerson's method of triads: Comparison of more than two products

Cases
♦ Replicated testing using fragrance preferences
♦ Action standards for product improvement and cost reduction
♦ Multiple comparisons of cookies manufactured using different processes and formulations

Wednesday - March 17 (8:00 a.m. - 4:00 p.m.)

Topics
♦ Advanced concepts and applications: Retasting, memory and sequence effects
♦ Measuring the effect of training
♦ Review of the ASTM Claims Guide
♦ Testing for Equivalence: FDA recommendations, the TOST, improved methods for binary and normally distributed data
♦ Testing to support ratio and multiplicative claims, "Up to" claims
♦ Disposition of no difference/preference votes in claims testing
♦ National Advertising Division (NAD)

Cases
♦ Memory and sequence effects in tests involving orange and apple beverages
♦ Improving discrimination by allowing retasting: A case study using a sports beverage
♦ Relating trained panel and consumer sensitivies using vanilla ice cream
♦ Advertising claim for product equivalence
♦ Malodor multiplicative claims
♦ Review of some cases before the NAD

Thursday - March 18 (8:00 a.m. - 12 noon)

Topics
♦ How to get d´ values from intensity ratings data
♦ Rating means and scale means
♦ Ranking
♦ First choice (discrete choice): Logit and normal assumptions and why they differ
♦ First-last (maxdiff or best-worst) data and how to analyze the data

Cases
♦ Generating a dose response relationship using ranking and rating
♦ Scaling first-last data on file server features
♦ Application of a first choice model to salt taste data
♦ Ingredient change: Getting d´ values from descriptive analysis

Thursday (1:00 p.m. - 4:00 p.m.) and Friday - March 19 (8:00 a.m. - 12 noon)
IFPrograms™ Workshop

Participants will be given a detailed review of the capabilities of IFPrograms™. This includes the organization of the software, types of suitable data, importing data, exporting graphics and use of the report window and interpretation of outputs. Working in teams on projects, participants will use the software to solve practical problems involving the design of difference tests, risk assessment in choosing methods and sample sizes, and the scaling of difference tests, ranks and ratings data. The teams will then prepare presentations and present their findings to the group. In the process of working on the projects, the teams will gain experience in the following areas:

♦ Difference testing methods: 2-AFC, multiple 2-AFC, 3-AFC, triangle, duo-trio, same-different, degree of difference, dual pair, A-Not-A, Torgerson’s method of triads, specified and unspecified methods of tetrads, first-last (MaxDiff) and first choice (discrete choice)
♦ Experimental design and risk issues: Simulation of results from discrimination methods, selecting methods based on power, determining sample sizes to meet risk requirements, generating and using psychometric functions and power curves for different methods
♦ Developing action standards: Compare and combine results from different methods and establish criteria for product quality management decision
♦ Replicated testing: Apply the beta-binomial model to reduce cost and identify sources of heterogeneity in materials and among panelists
♦ Ranking and rating methods: Scale rank data, simulate ratings data from scales with different numbers of categories and use a Thurstonian model for ratings to interrelate methodologies

Download the course brochure here.

Register for the course here.