Expanding Quantitative Approaches in Engineering Education Research: Embracing Empirical Bayes for Longitudinal Person-Centered Analysis

In engineering education research, conventional quantitative methods have been widely used to examine correlation patterns and establish causal relationships from a variable-centered perspective. However, recent critiques highlight the risk of essentializing findings from certain samples to the whole population and the lack of contexts in these methods, which may systematically neglect marginalized perspectives. As a response, there is a growing interest in incorporating person-centered quantitative methods, which prioritize individual perspectives and experiences, aligning with critical (race) quantitative research tenets. Despite the promise of person-centered approaches, they often face challenges such as insufficiently capturing the complexity of individual experiences, difficulty handling missing data and complex data structures, and limited capacity for statistical inference. This article seeks to address these limitations and expand the toolbox of quantitative research methods tailored for person-centered analysis by proposing the use of the empirical Bayes approach to analyze longitudinal data in engineering education applications. The empirical Bayes method offers several advantages over traditional person-centered techniques. By demonstrating the utility of the empirical Bayes approach in a simple example using longitudinal student grade data from a first-year engineering course, this study aims to provide an innovative alternative for more nuanced and inclusive data interpretation in various research contexts.