Quantitative research using the PLS-SEM method (Partial Least Squares – Structural Equation Modeling) has become a popular analytical approach across various disciplines, particularly in management, psychology, and social sciences. This method allows researchers to design and test relationships between latent variables—those variables that cannot be directly measured, such as customer satisfaction, loyalty, and motivation. With this capability, PLS-SEM serves as a powerful tool in the analysis and validation of research models.

During an online training session, Dr. Prita Prasetya highlighted one of the main advantages of PLS-SEM: its ability to handle data that do not follow a normal distribution and to work effectively with relatively small sample sizes. These characteristics make PLS-SEM highly suitable for studies aimed at theory development or exploring complex relationships within survey data, as it can simultaneously evaluate both direct and indirect effects among variables.

Preliminary steps in conducting research with PLS-SEM involve constructing a path model, a graphical representation of hypothesized relationships between variables. The model consists of two components: an outer model, which links indicators to constructs, and an inner model, which depicts causal relationships among constructs. Participants in the training learned how to accurately develop both models to ensure the reliability and validity of their analysis.

 

Subsequently, researchers must validate the constructs within the outer model to confirm that the indicators adequately represent the latent variables. This validation process involves assessing the validity and reliability of indicators, using reflective or formative approaches depending on the nature of the constructs under study.

The analysis of the inner model is a critical step focused on evaluating the strength and significance of the causal relationships between latent variables. Through this analysis, researchers can assess direct effects as well as mediated effects, thereby gaining comprehensive insights into the phenomena under investigation.

PLS-SEM also offers notable flexibility for testing theoretical models, estimating parameters, and evaluating research hypotheses without the constraints typically associated with covariance-based SEM. This makes it an ideal choice for complex research models and data that do not meet classical assumptions.

To conclude the training session, Dr. Prita encouraged researchers to optimize the use of PLS-SEM in their quantitative studies. This approach not only enhances the accuracy of research findings but also helps generate richer and more meaningful insights for the advancement of knowledge across various fields.

With sound understanding and application, PLS-SEM stands as a powerful method for strengthening the validity of quantitative research, assisting both academics and practitioners in addressing complex and nuanced research questions.