Currently, measuring cause and effect relationship for an intangible situation such as customer satisfaction has been gaining momentum. For this, two statistical techniques are used: (i) structural equation modeling-SEM with co-variance matrix, and (ii) partial least squares-PLS, which determine, as multivariate technique, the relationship between observable and latent variables in order to test a series of associations. In this sense, this paper is aimed at presenting each of the techniques (SEM and PLS) from an interpretative perspective, by means of a case study. As a result, one of the differences found between the methods is the estimation procedure, since SEM is oriented towards theory, emphasizing the transition from exploratory analysis to confirmatory, whereas PLS is focused on the causal-predictive analysis in high complexity situations, though with little theoretical information. Both methods pursue different goals and should not be exclusionary but complementary, based on the interests of the researcher and the objectives of the study for the implementation of one method or another.
Bibliographical notePublisher Copyright:
© 2016, Universidad Nacional de Colombia. All rights reserved.