Podcast

This paper identifies four political problems inherent in Learning Analytics in higher education contexts that people working within the field generally are not aware of: (1) educational data scientists are not apolitical and that educational data is not neutral, which is due to the fact that knowledge or action cannot be value-free but is shaped by its social contexts; (2) the emphasis on a heightened awareness of ethics in Learning Analytics, which is unlikely to bring social justice, is itself not sufficient to solve the underlying social problems of the technologies; (3) the idea of pursuing Learning Analytics for ‘social good’ could over-simplify issues that are actually politically complex and that there also might not be consensus over what is desirable, because what considers as ‘good’ implies vague and undefine political assumptions; (4) an understanding of the politics of Learning Analytics is needed in order for meaningful social change to be possible. 

 

As data science is concerned with norms, discrete categories, precise definitions and predicable futures, the author points out that it is essentially undesirable to outliers or marginalized groups, such as queer and transgender individuals. To address these limitations of data science, the author encourages practitioners in the field to consider alternative ways of applying data science that support the experiences of non-conforming ‘learners’ and un-categorizable forms of ‘learning’. For this, he mentions about the possibilities of engaging with social research on the issues that Learning Analytics intends to address as well as conducting studies that adopt critical or participatory approaches (e.g. anti-oppressive design).