In our summer 1999 newsletter, we discussed a method for the analysis of preference data. The goal of that method, called Multivariate Preference Mapping, was to show how preference data can be used to derive maps of products, ideals and their variances so that the basis for preference decisions could be understood. In this report we look at the application of a widely used tool, logistic regression analysis, to model preferential choice data. The logistic model is a staple for modeling risk factors in the field of epidemiology. This model does not allow for probabilistic perceptual variability and covariance as seen in probabilistic unfolding models. Nevertheless, while there are drawbacks to the use of logistic regression as a model for preference data, it can be effectively applied under certain circumstances and it offers a relatively simple technique that lends itself to easy interpretation.
This technical report appears as:
Lampe, R. (2005). Alternative Methods for Understanding Preference. IFPress, 8(4) 2-3.
Colleagues can download the entire technical report here:
Alternative Methods for Understanding Preference
Not a Colleague? Click here to join for free!
This technical report also appears in our book, Tools and Applications of Sensory and Consumer Science.