Using Data Labels to Discover Moderating Effects in PLS-Based Structural Equation Modeling
PLS-based structural equation modeling is extensively used in e-collaboration research, as well as in many other fields of research. Two main types of exploratory analyses are frequently employed in the context of PLS-based structural equation modeling: covariance (or correlation) analyses, where the covariances (or correlations) among all variables are inspected; and model-driven exploratory analyses, where one or more variations of theory-supported models are built and adjusted associations among variables are inspected. These analyses, while useful, provide limited insights about the possible presence of moderating effects. We discuss a general approach through which researchers can employ data labels, implemented as symbols that are displayed together with legends on graphs, to uncover moderating relationships among variables. The discussion is illustrated with the software WarpPLS version 4.0. While the approach is illustrated in the context of e-collaboration research, it arguably applies to any field where PLS-based structural equation modeling can be used.