The Structural Equation Modeling with partial least squares

a statistical technique for defense and international security studies

Abstract

Partial least squares structural equation modeling (PLS-SEM) is a robust multivariate statistical technique, which adjusts to small samples and allows researchers to answer a series of interrelated questions simply, systematically, and comprehensively. This tool was explicitly created to be used in the social sciences, in which studies do not always have large samples, and new theories can be formed from the constant social changes. It achieves this, thus, by modeling the relationships between multiple dependent and independent constructs, considering different types of measures and various variables. This methodological essay aims to present the statistical technique in question, describing the assumptions of the method, its procedures, quality parameters, and limits. This analysis is illustrated with an applied example by using the SmartPLS software. Thus, the article brings a reflection on the potential of this statistical technique for international security and defense research.