PLS-SEM for Software Engineering Research
Daniel Russo, Klaas-Jan Stol
Abstract
Software Engineering (SE) researchers are increasingly paying attention to organizational and human factors. Rather than focusing only on variables that can be directly measured, such as lines of code, SE research studies now also consider unobservable variables, such as organizational culture and trust. To measure such latent variables, SE scholars have adopted Partial Least Squares Structural Equation Modeling (PLS-SEM), which is one member of the larger SEM family of statistical analysis techniques. As the SE field is facing the introduction of new methods such as PLS-SEM, a key issue is that not much is known about how to evaluate such studies. To help SE researchers learn about PLS-SEM, we draw on the latest methodological literature on PLS-SEM to synthesize an introduction. Further, we conducted a survey of PLS-SEM studies in the SE literature and evaluated those based on recommended guidelines.