This work considers the implementation of PLS regression principles within the PLS Path Modeling framework in order to cope with two problems often encountered in component-based structural equation modeling: correlated formative indicators and latent variable scores. The proposal shows to be profitably usable in application related to business and related areas.
ESPOSITO VINZI, V. and TRINCHERA, L. (2008). Formative Indicators in Structural Equation Models and Highly Correlated Latent Variable Scores: a Full PLS-based Approach. In: International Symposium on Business and Industrial Statistics with emphasis on Quantitative Analytics for Banking, Finance and Insurance.
Mots clés : #Collinearity