How to Diagnose the Need for 3D Unfolding
Rousseau, B. (2013). IFPress, 16(3) 3-4.
Multivariate mapping techniques are frequently and commonly used to visualize the large amount of data generated in sensory and consumer testing experiments. Since it is desirable to summarize data using as simple a model as possible, multidimensional solutions that capture the relevant information with fewer dimensions are usually prioritized. Moreover, it is less challenging to communicate results in two dimensions. Thus, many analyses are conducted and summarized in two dimensions and this approach is often appropriate. However, using only two dimensions can ignore important and relevant information contained in higher dimensions. In this report, we illustrate how an extra dimension is sometimes needed to capture relevant information when the multidimensional unfolding method, Landscape Segmentation Analysis® (LSA), is applied, so that the proper dimensionality is used to uncover the drivers of liking space
Figure 1. 2D LSA map of the orange juice data showing product positions and consumer individual ideal points with average liking ratings per product on a 9-point hedonic scale. The white dots represent consumer ideals and the products are labeled according to their source (US and SA).