In November 2013, Michael Gove reminded us the Coalition had “outlined plans for changes to GCSE qualifications designed to address the grade inflation, dumbing down and loss of rigour in those examinations”.

“Rigour” was Gove’s watchword. More rigorous exams in English and Maths would be first examined in 2017. This meant teachers had to be ready to teach the new English and maths GCSE curriculum this September.

Untrialled, untested, unevaluated – the tight timetable for these exams foretold disaster. To make sure schools didn’t opt out of GoveGCSEs, schools were told that IGCSEs would no longer count towards league tables. This overturned a 2010 Conservative manifesto promise that IGCSEs would be included to create a “level playing field2 between the private sector, where IGCSEs had become increasingly popular, and the state sector.

But do GoveGCSEs live up to their promise of injecting more rigour? Possibly not, judging from the row which erupted after two exam boards, Edexcel and OCR, criticised AQA sample papers for not being sufficiently challenging.

As part of its review following this criticism, Ofqual compared sample papers for “reformed” GCSEs offered by AQA, Pearson, OCR and WJEC Eduqas, with question papers for the “current GCSE (2011-2012), similar assessments of pupils aged around 16 taken in ten international jurisdictions and Cambridge IGCSEs and O level (2011)”.

It’s unclear why papers for 2011-2012 are described as “current’. Surely, it was possible to use more recent papers? And what of the international exams used for comparison? The Massachusetts Comprehensive Assessment System is being dropped by half of Massachusetts school districts in favour of a new test (PARCC) which the Commissioner of the state’s Department of Elementary and Secondary Education said would “help the state reduce the stubborn achievement gaps between rich and poor, white and minority, by giving teachers better information about which kids need extra support”. In other words, PARCC will be formative unlike summative GCSEs.

Similarly, the Japanese exam was diagnostic. The results of early National Assessment of Academic Ability tests were used to revise Japanese school curriculum and inform policy making. This is the opposite of what is happening in England where policymaking defines the curriculum and exam changes follow. In South Korea, the National Assessment of Educational Achievement tests are for a “purely informational purpose”. New Zealand’s exam forms part of a cumulative system of credits which culminate in a graduation certificate at age 18; the Hong Kong exam was replaced in 2012 in favour of a final exam taken at age 17.

Despite many of the international exams used for comparison being out-of-date or designed for formative purposes, Ofqual said “current GCSEs are judged to be of lower expected difficulty than similar international assessments”, and the “expected difficulty of the reformed, higher tier GCSE sample assessments is more in line with similar international assessments”.

But reformed GCSEs are not the same as most international tests. Most developed countries have graduation at 18 with few tests at 16. Where tests are taken they are often formative. And they are not used to judge schools.

Gove claimed his reforms would bring England in line with the best international exams. But as we have seen they do not.

And what of rigour? As Schools Week previously reported, three exam boards need to “refine their higher and foundation tier papers to sufficiently differentiate across student abilities” while AQA’s sample paper was too easy. Exam boards have been told to produce rewritten samples and these will be sent to schools before the end of June.

It is likely schools will already have decided which exam board’s syllabus to follow – they need to be sure the chosen exam will accurately assess pupils’ competence. Ofqual says it is “more confident that going into 2017 we will have effective assessments of GCSE mathematics”.

Thousands of teachers, parents and pupils will be crossing their fingers this confidence isn’t misplaced.