This paper considers how centrally-available and comprehensive quantitative data can be used as an indication of risk in a risk-based system of quality assurance, as currently implemented in England. This consideration is set within the policy context of expanding higher education and the introduction of a new system of funding undergraduate education through student loans for tuition fees in 2012.
Utilising machine learning techniques this paper demonstrates that the best model utilises three indicators relating to applications, staffing and finance. The paper concludes that the ability of data to predict the outcome of QAA reviews, and hence help prioritise them, is extremely limited
This paper was presented at EQAF and reflects the views of the named authors only.