This article proposes a pragmatic approach to data justice in education that draws upon Nancy Fraser's theory. The main argument is premised on the theoretical and practical superiority of a deontological framework for addressing algorithmic bias and harms, compared to ethical guidelines. The purpose of a deontological framework is to enable the evaluation of justice claims, not as a burden for individuals or groups, but as part of an ongoing public dialogue about the disparate impacts of algorithmic decisions. In a deontological framework, scrutiny is not limited to differences in treatment but is broader and more demanding, involving the evaluation of design approaches and a wider-ranging political discussion about who might be harmed.