Title: SO-PLS-PM: From multiblock data analysis to path modeling
Authors: Rosaria Romano - University of Calabria (Italy) [presenting]
Oliver Tomic - Norwegian University of Life Sciences (Norway)
Kristian H Liland - Norwegian University of Life Sciences (Norway)
Age Smilde - University of Amsterdam (The Netherlands)
Tormod Naes - NOFIMA (Norway)
Abstract: A new approach to path modeling named SO-PLS path modeling (SO-PLS-PM) is presented and compared with the more well-known PLS path modeling (PLS-PM). The new method is flexible, graphically-oriented, and allows for handling multidimensional blocks and diagnosing missing paths. Instead of fitting everything at the same time using the full path model scheme, the approach splits the estimation up into separate multi-block regression models for each dependent/endogenous block. In other words, thepath modeling is turned into a series of regression analyses. Since the whole procedure is based on PLS regression and orthogonalization, the method can be used for any dimensionality of the blocks, for collinear variables as well as design data and it is invariant to the relative weighting of the blocks. In order to allow for a thorough comparison between the two methods, new definitions of total, direct and indirect effects in terms of explained variances are proposed, along with new methods for graphical representation. The two PLS methods are tested on two well-known data sets in the PLS-PM literature from customer satisfaction analysis and descriptive sensory analysis. The findings from the empirical applications serve as a basis for recommendations and guidelines regarding the use of the SO-PLS-PM versus PLS-PM.