Comparing partial least squares and partial possibilistic regression path modeling to likert-type scales: A simulation study

Abstract

Partial possibilistic regression path modeling (PPRPM) combines the principles of path modeling with those of possibilistic regression to model the net of relations among latent variables through interval-valued coefficients, in order to take into account the vagueness in the model specification. An interval valued coefficient is defined by a midpoint and a range. Through a simulation study, the paper presents a comparison between PPRPM and partial least squares path modeling (PLSPM), when these are used for analyzing questionnaire data, with responses recorded on Likert scales. The estimates of the two models have similar behaviors, with respect to the simulated scenarios. Focusing on a realistic scenario setup, the results highlight the benefit of PPRPM that allows the model to report the component-wise estimation of vagueness in the inner model.