THE USE OF STRUCTURAL EQUATION MODELING OF PARTIAL LEAST SQUARES (PLS-SEM) FOR DEFENSE STUDIES

The applicability of PLS-SEM to defense and operational decision support studies is  presented, emphasizing its potential contribution to problems frequently addressed in Operations Research and military analytical studies. Structural Equation Modeling of Partial Least Squares (PLS-SEM) makes it possible to integrate human, organizational, behavioral, and operational variables into unified predictive models, allowing the quantitative evaluation of relationships that are difficult to be captured by traditional linear approaches. In defense contexts, this capability is relevant to model operational performance, threat perception, military organizational behavior, mission effectiveness, cognitive and behavioral factors, command dynamics, and strategic adaptation in scenarios of high uncertainty. PLS-SEM is a second-generation multivariate statistical technique developed to estimate complex predictive relationships between latent and observable variables. Unlike covariance-based approaches, PLS-SEM has predictive guidance and employs iterative ordinary least squares algorithms to maximize the explained variance of endogenous variables. Its statistical flexibility allows the simultaneous estimation of formative and reflective models, favoring the analysis of direct, indirect and mediation effects in multidimensional systems characterized by uncertainty, limited theoretical consolidation, non-normal distribution of data and relatively small samples