PARTIAL LEAST SQUARES PATH MODELLING IN MARKETING AND MANAGEMENT RESEARCH: AN ANNOTATED APPLICATION

Since Joreskog (1967) provided an algorithm for maximum likelihood estimation of latent variables structural models, covariance-based approaches (CBSEM) are tautologically synonymous of structural equation modelling. However, when Diamantopoulos and Winklhofer (2001) alerted that formative configuration should be considered as a reasonable alternative to model certain constructs, academic realised that CBSEM had many limitations to deal with this kind of constructs. Partial least squares path modelling (PLSPM) has recently arises as an alternative to CBSEM as its algorithm easily manages formative constructs. Proposed as a soft modelling alternative to CBSEM by Joreskog's thesis advisor Herman Wold, PLSPM has addtional properties that explained its growing use in marketing and management research: Few distribution assumptions and can be estimated with small sample sizes. The objective of this chapter is, on the basis of a practical example, to illustrate and provide guidelines regarding the use of PLSPM as a structural equation modelling (SEM) tool, revealing its strengths and weaknesses, in order to allow researchers to decide on the adequacy of this approach for their particular investigation.