Artificial Intelligence (AI) Use in Data Analysis: A Comparison of ChatGPT and SmartPLS Outputs in PLS-SEM Analysis

Abstract

This study aims to evaluate the usability and reliability of ChatGPT, a large language model, to conduct Partial Least Squares Structural Equation Modeling (PLS-SEM) analysis. For this purpose, a research model based on the Technology Acceptance Model was tested with 523 participants using both SmartPLS and ChatGPT. Key statistical indicators such as internal consistency, convergent and discriminant validity, VIF values, path coefficients, and R-square values were compared. The results show a high level of consistency between ChatGPT and SmartPLS outputs throughout all stages of the analysis. Furthermore, the statistical validity of SmartPLS and ChatGPT outputs is supported by examining their Pearson Correlation (r), Mean Absolute Error (MAE), and Root Mean Square Error (RMSE) values. Both tools identified the same significant and nonsignificant relationships, and similar mediation effects were found. Although ChatGPT lacks visualization capabilities and requires step-by-step guidance, it successfully reproduced statistically valid results. These results suggest that ChatGPT may be a complementary and cost-effective tool for researchers, especially those without access to licensed software. Future research should explore its usability with other statistical techniques and artificial intelligence models.

Ähnliche Arbeiten