Predictors of Cultural Intelligence: Automated Machine Learning vs. PLS-SEM

Abstract

This study applies a dual-method analytical approach combining Automated Machine Learning (AML) and PLS-SEM to investigate the predictors of Cultural Intelligence (CQ), one of the most commonly used constructs in international business (IB) research. Our research seeks to (1) explore a wide range of potential CQ predictors, and (2) demonstrate how AML and PLS-SEM methodologies can complement each other in IB research. Using a large international sample of 58,784 participants from 160 countries, we find that while international experience and personality traits predict CQ as expected, English language proficiency and emotional intelligence also emerge as significant predictors. The methodological comparison confirms that AML excels in exploratory pattern detection and handling complex datasets, while PLS-SEM enables theory testing and structural validation. The integration of both methods yields richer insights than either method alone. Implications for practice and research are discussed, and guidelines for using AML and PLS-SEM in IB research are provided.