Datafication of Education and Machine Learning Techniques in Education Research: A Critical Review

Abstract

The aim of the study is to discuss the technical and ethical considerations regarding the current applications of the emerging data science techniques in the field of education. Based on the literature review, this study provides an overview of machine learning approaches that have been used to answer to educational research agenda in South Korea. By comparing the logical features of these computational techniques to the conventional statistical methodologies, the authors highlight the unique challenges and opportunities associated with educational data science and discuss what that means to developing scientific knowledge in the field of education research. The authors argue the importance of the awareness of the differences in data conditions and causal inferences between computational approaches and conventional statistical modelings. Computational techniques using big data are not a magical tool to discover knowledge. This study explains why researchers’ domain knowledge and rigorous data preparation have bigger impacts on drawing reliable and meaningful information for educational intervention. This study encourages critical inquiry into the implications of data science for education. By revisiting the goals of public education, the authors call for the future research to host open discussion of the potential of educational data sciences.