Algorithmic Bias in Education

In this paper, we review algorithmic bias in education, discussing the causes of that and reviewing empirical literature on specific ways is known to have manifested education. While other recent work has reviewed mathematical definitions fairness expanded approaches reducing bias, our focuses instead solidifying current understanding concrete impacts education—which groups are be impacted which stages agents development deployment educational algorithms implicated. We discuss theoretical formal perspectives connect those machine learning pipeline, metrics for assessing bias. Next, evidence around beginning with most heavily-studied categories race/ethnicity, gender, nationality, moving available less-studied categories, such as socioeconomic status, disability, military-connected status. Acknowledging gaps what been studied, propose a framework from unknown equity. obstacles addressing these challenges four areas effort mitigating resolving problems AIED systems technology.