Multiplicative models for combining information from several sensory experiments: a Bayesian analysis

Abstract
If data are available from a series of quantitative sensory experiments on the same type of product, and the assessors in these experiments are drawn from a common pool, then it is possible to combine information from the experiments on the relative biases and variability of individual assessors, and to examine the influence of possible temporal effects over the series. Such a combination of information is illustrated using a series of apple-tasting experiments conducted with the main aim of monitoring assessor performance over time. Models which include random effects and multiplicative interaction terms have been used for modelling heterogeneous interaction between assessors and products in individual sensory experiments. Such models are extended here to analyse data from series of experiments. A Bayesian approach is used that allows for adjustment for missing observations and for the use of information on assessors' previous performance when analysing future experiments. This use of previous information leads to a reduction in the average variance of product differences.
Year
2008
Category
Refereed journal