Comparing PLS-SEM Statistical Technologies for Educating the Importance of Linearity

Abstract

This study aims to reveal how well two PLS-SEM statistical technologies—WarpPLS and SmartPLS—work for testing attitude theory with different types of relationships: linear and nonlinear. Non-probability sampling collects 786 internet customers from the digital marketplace. In the linear relationships between the theory of reasoned action baseline (attitudes and subjective norms) and online buying intention, SmartPLS and WarpPLS deliver similar loadings, reliability, validity, and path coefficients. WarpPLS is more effective in validating the nonlinear relationship of consumer ethnocentrism to attitudes toward imports, but SmartPLS is more comprehensive in providing robustness and advanced features. Teachers should educate students on the importance of testing linearity in the required preliminary checks. This step helps them choose the suitable software algorithm and get more accurate results and reports. This study is limited to the default parameter and standard application; future works can explore more advanced settings and complex path models.