Applying DebiasEd: A Package for Mitigating Unfairness in Educational Data

Educational data mining has augmented learning processes and improved outcomes within and outside of traditional classrooms. However, due to the diversity of learners, computational methods could be biased against different (often protected) groups of learners. Due to the biases, researchers have become increasingly interested in understanding and reducing the bias within their educational data and models. In this tutorial, we present and discuss how to use several modern unfairness and bias mitigation techniques encapsulated within the DebiasEd Python package with a focus on ease of implementation and generalizability for education researchers.