What are the chances?

I haven’t read the article in the national press about an Open University algorithm that can forecast a student’s final grade “within just a week of their starting their course” (The week in higher education, 30 July), but we first used this kind of forecasting at the OU in 2006.

Using a statistical program, we were able to attach a “predicted probability of success” (PPS) to each new student entering the OU that proved to be surprisingly accurate. In one typical year, the PPS varied from 83 per cent for a few students to just 9 per cent for another.

We used the program only for research purposes, but I proposed that students should have access to their PPS. It seemed wrong that they should be committing a large sum of money while, in some cases, having characteristics that gave them apparently a very moderate chance of success. My idea was that students could complete a self-assessed questionnaire that would tell them their PPS, but – crucially – also tell them how they could improve it.

However, I suppose that the ethical minefield that would have had to be navigated to get to such a system kept the idea in the dock.

Ormond Simpson
Visiting fellow
University of London Centre for Distance Education