Why I ...believe dropout rates are too low

We must always be concerned if students drop out of their degree programmes. For each one, we need to understand why they did so, whether we could have done anything to prevent it, whether they were following the right programme for them and, ultimately, whether they should have been offered a place at all. I do not believe that the main purpose of higher education is to serve the needs of employers, although the government clearly does. But even putting the most utilitarian focus on higher education, to assume that a high dropout rate indicates waste in the university system is wrong.

In the first place, a student who drops out may complete the degree later, perhaps in a different subject or at a different university. Second, even if they never complete the degree, they will most likely have acquired skills and knowledge that may be useful in employment or elsewhere. But, beyond this, there are still good reasons for wanting to have a higher dropout rate.

The use of aptitude tests, work samples and other novel forms of assessment will help us improve our ability to select students who can benefit from higher education. We always want to select those who will do well and reject those who will do badly. But under any assessment system, however sophisticated, we will accept students who will drop out and reject those who would have done well (in the jargon, the false positives and false negatives respectively).

We could ensure dropout rates (that is, false positives) close to zero if we denied university education to all but those certain to do well.

Alternatively, we could ensure that everyone who has the capacity to benefit from higher education does so (that is, minimise false negatives), provided we are prepared to allow everyone to go to university. The only tenable policy lies between these extremes. We can't do much about the total number of "mistakes" we make. What we can do is make an informed choice about where between the two extremes we want to be.

Since most students drop out in the first year, the cost to the Treasury of each false positive is just a year's funding - on average about £6,000. The cost of a false negative is more difficult to estimate, but taking the Department for Education and Skills figure of an extra £400,000 in earnings over the lifetime of a graduate, the loss in terms of income tax alone is well over £100,000. Even if we take into account the fact that the Treasury gets that money over the working life of the individual, rather than upfront in one payment, this is an amazing deal.

Given that our selection methods can never be perfect, and the huge difference between the cost of dropouts and the benefits that accrue to both individuals and the Treasury for those who graduate, this suggests that the traditional burden of proof should be reversed. Instead of a student having to show they have the capacity to benefit from higher education, we should accept all those who want to go, unless there is clear evidence that it would not be in their interests.

This will mean that many will drop out, but also that a number who would otherwise not have gone into higher education will graduate. Of course we must do everything we can to cut dropout rates, but we must also accept that the increase in false positives is the price we have to pay for reducing the number of false negatives.

Furthermore, we should not expect dropout rates to be the same in all universities. Different institutions recruit from different strata of the attainment range, so some will have very low dropout rates because they take only the sure bets. Others recruit from that greyer area where we cannot tell whether students will succeed or not. Given this, the fact that no British university has a dropout rate of more than 50 per cent suggests we simply are not taking enough chances.

Dylan Wiliam
Assistant principal and professor of educational assessment, King's College London.
He writes in a personal capacity.