Heterogeneity in variance-based structural equations modeling: A systematical procedure for applying FIMIX-PLS

Verlag C.H. Beck oHG
Publication Type:
Journal Article
Marketing: Zeitschrift fuer Forschung und Praxis, 2008, 30 (4), pp. 239 - 255
Issue Date:
Filename Description Size
Thumbnail2009001924OK.pdf369.01 kB
Adobe PDF
Full metadata record
Applications of the partial least squares (PLS) path modeling technique are usually based on the assumption that the analyzed data stem from a single population. However, in many real-world applications, this assumption of homogeneity is unrealistic and an analysis on the aggregate data level might result in ambiguous or imprecise findings. The finite mixture partial least squares (FIMIX-PLS) approach by Hahn (2002) and Hahn et al. (2002, 2005), respectively, is an appropriate means of assessing if unobserved heterogeneity significantly affects the inner path model relationships. Although this method is the primary approach for capturing unobserved heterogeneity in variance-based structural equation modeling, the dissemination of FIMIX-PLS is still in its early stages. In view of this approachs recent implementation in the freely available software application SmartPLS (Ringle/Wende/Will 2005a), the aim of this paper is to demonstrate the relevance of FIMIX- PLS in respect of evaluating PLS path modeling results and, thereby, to further support its application in research and business practice. Furthermore, this article presents a multi-stage procedure to systematically conduct FIMIX-PLS analyses, whose application is illustrated by an exemplary brand preference model.
Please use this identifier to cite or link to this item: