Ofqual's own rules suggest that it should have prevented controversial changes to the grading of this summer's English GCSEs and allowed a rise in results, TES can reveal.
The exams watchdog argued last week that permitting grade boundaries to remain constant between January and June would have prompted a "big increase" in A*-C grades and breached its statutory duty to maintain standards.
But a letter from the watchdog sent to exam boards in June reveals that, according to its own guidance, it should have made an exception. It lists five separate conditions that must all be met for Ofqual's "comparable outcomes" policy of keeping standards in line with previous years' results to apply. The new English GCSEs fail to meet four of them.
The news came as an unprecedented "alliance" - expected to include headteacher and classroom teacher unions, independent schools, local authorities, at least two academy chains and FE colleges - was forming to call for further investigation.
The first of Ofqual's five conditions was that the cohort in a subject "must be similar, in terms of ability, to those of previous years". The watchdog's report on English GCSEs last week noted that the "attainment profile" for 2012 dropped, as grammar and independent school pupils were replaced with extra, lower-achieving comprehensive pupils.
The second condition was that the qualification must be "fit for purpose". On Monday, education secretary Michael Gove told Parliament that this year's English GCSE was "not fit for purpose".
The third condition was that the "nature of the qualification" must be the same. Ofqual's report last week stated that "these qualifications are different from previous English qualifications in a number of ways".
Finally, Ofqual said in the letter that comparable outcomes must only be applied where "previous grades were appropriate". Ofqual's report said that the English GCSE grades in January were "too generous".
The only condition that appears to have been met is that teaching standards have remained largely consistent.
The watchdog has interpreted its own conditions differently, and is insisting that the comparable outcomes approach that prevented grades from rising in June was correctly implemented. But Russell Hobby, general secretary of the NAHT heads' union, said: "Ofqual doesn't seem to have applied the concept of comparable outcomes properly. In this instance there is a regulatory failure. It is a big mess."
Brian Lightman, general secretary of the Association of School and College Leaders, said: "This would appear to be yet another piece of evidence that the procedures applied to this exam were not fit for purpose." The NUT said they were "very flawed".
Meanwhile, one of the government's favourite heads called for a boycott of AQA, the exam board at the heart of the controversy. Sally Coates, principal of Burlington Danes Academy in West London, who chaired ministers' teaching standards review, told TES that she will not use AQA again after a dramatic fall in pupils achieving a C at GCSE English. "I hope as a result of this AQA loses huge numbers of schools," she said.
But an AQA spokesperson said all boards had raised the English and English language grade boundaries in June and its "overall outcomes" were "very similar to other awarding bodies".
The Education Select Committee will question heads' leaders, Ofqual chief regulator Glenys Stacey and Mr Gove over the controversy early next week, but has not decided whether to report on the issue. Mr Gove has ruled out ministerial intervention.
Key unanswered questions for Ofqual
Why has there been no independent investigation?
Ofqual has produced the only report to date. But many heads, and Sir Mike Tomlinson (see page 9), have said the watchdog is likely to have played an active part in events it is investigating.
If the January grade boundary decisions were `too generous', why were they not changed?
At least one former examiner has said that the discrepancy was easy to spot. These boundaries were then used by schools to decide, at least in part, where to focus their efforts in June.
Why has Ofqual used the smaller numbers of candidates sitting in January and lack of data from previous years to justify `too generous' grade boundaries?
Both issues and the risks involved were predictable for a new qualification and a system that has introduced modular GCSEs before.
If the June grade boundaries were correct, why have so many schools been given results way below their expectations?
Schools have said their predictions were not solely based on January boundaries, but on years of experience among staff.
Did Ofqual breach its duty to maintain standards and a second statutory objective to `promote public confidence' in qualifications in January?
The regulator said that changing the June grade boundaries would have breached its "statutory standards objective". It could be argued that the watchdog is only citing these objectives when it suits.
Original headline: Ofqual rules suggest it should have prevented GCSE grading changes