Blinder–Oaxaca decomposition applied to sensory and preference data

When equivalent regressions (same independent variables, same dependent variable) are applied to two different populations, Blinder–Oaxaca decomposition allows to estimate how much of the difference between means of the dependent variable can be explained by the differences between means of the independent variables. This method has been often applied in sociological studies, i.e., to estimate how much of the difference between salaries of men and women can be explained by differences in age, years of education and so on. The non-explained part of the observed differences can be due to discrimination between sexes. This communication tries to explore the application of the Blinder–Oaxaca decomposition to sensory and preference data, by estimating, for instance, how much of the difference between the means of acceptability of two products can be explained by the differences in the means of sensory attributes also evaluated in the same products.