Since the 2010 general election the government has made countless policy changes and among the most damaging of these is the decision to decouple AS from A level. This will lead to some students having to stick with courses that prove to be wrong for them, an increased failure rate at A level, a narrowing of the post-16 curriculum of many students, fewer students studying mathematics and modern foreign languages post-16 (which were commonly done as a “fourth” subject by arts and science-based students respectively), a greatly increased number of university entrance tests, and a disproportionate effect on disadvantaged students.

When secretary of state Keith Joseph decided to try to broaden the post-16 curriculum, he introduced AS levels, but not as we know them today. The Joseph AS was a vertical half of A level and his idea was that 16 and 17-year-olds could broaden their qualifications by studying two A levels and two AS levels, each of the latter counting as a half an A level. Students soon found that doing two Joseph ASs involved more work than one A level and numbers never really took off.

As the disadvantages of the vertical AS became increasingly apparent, Lord Dearing considered their future in his review of 16-19 education in the mid-1990s. I was one of those trying to persuade him that the vertical AS should be replaced with a horizontal AS that would be a half-way house to A level and I wrote a paper for him along those lines.

My main reason for wanting a horizontal AS was that, in interviewing 16-year-olds about to enter the sixth form at Durham Johnston Comprehensive School, where I was head, I was often struck by the difficulty that students had in making a decision about what courses - vertical AS and A level - to study for the next two years. How much better, I reasoned, if they were able to make one-year decisions and then review these a year later in the light of their results in horizontal AS at age 17.

When the horizontal AS was introduced, many students chose four AS subjects at 16 and then dropped one a year later. That was a major factor in reducing the failure rate at A level, which was welcome. But the really interesting thing was that a large number of 17-year-olds did not choose the three A levels that they would have chosen a year earlier without the benefit of a year’s study in each subject.

‘Adding urgency to work’

That is the system we have had for the past 20 years or so. It started with lots of complaints that students had to work much harder in the lower sixth, which had previously been regarded as an “easier” year in many schools. I had no problem with adding some urgency to their work in that year.

Later, the complaints built up that the modular nature of the AS and A level meant that students were taking too many examinations over the two years and that this impacted adversely on teaching time. I have some sympathy with this argument, but the solution to this problem is surely to remove the modular element and have a single AS examination series at the end of Year 12 and a single A level series at the end of Year 13, rather than changing the whole structure.

Education ministers Michael Gove and David Laws were both convinced by the “excessive examinations” argument they heard from some heads and I could not persuade either of them of the serious implications of the loss of the AS.

The AS has provided hard evidence for university admissions tutors to guide them in making offers for university places. AS grades are a much more reliable indicator of likely A level grades than teacher predictions. Not surprisingly, an alarming number of university entrance tests are now listed on the Ucas website, as university departments decide that they need more concrete evidence than GCSE grades and teacher predictions.

A recent Sutton Trust report, Oxbridge Admissions, said that the application processes at both Oxford and Cambridge differed significantly from college to college, thus disadvantaging students from poorer backgrounds and those from schools with little experience of the Oxbridge admissions game. The report stated: “Tests vary by college at Cambridge - where many subjects have different tests at different colleges - and by subject at Oxford; Oxbridge admissions identifies almost 400 different possible tests beyond A levels that prospective students could have to take. Cambridge [has] confirmed that they will introduce new university-wide subject tests.”

If the government believes in evidence-based policy-making, my advice is not to return to the vertical AS, which is almost what is happening. We know it doesn’t work and, when we inevitably find that take-up is very low of the new AS, with the consequent narrowing of the study programmes of many students, more reform will be needed, the lives of thousands of students having been made much harder in the meantime.

John Dunford is chair of Whole Education, a former secondary head, general secretary of the Association of School and College Leaders and national pupil premium champion. He tweets as @johndunford

Want to keep up with the latest education news and opinion? Follow TES on Twitter and like TES on Facebook