Traditionalists need have no fear: Sir Ron Dearing has made it plain that he does not want A-levels watered down.

Action should be taken to bring all A-level syllabuses up to the standard of the most difficult, he says, and the number of syllabuses should be reduced to make comparisons easier. Also, a five-year study should be carried out to make sure that standards are being maintained over time and between awarding bodies. The recommendations are based on research showing that fears that some subjects are more difficult than others are well-founded.

It may be hard for geographers to swallow, but the old jibe that their subject is easy compared to physics is, according to the investigation, true. Physics, the report says, “is not only different in content from geography, but requires different skills and almost certainly requires understanding of inherently more difficult concepts”.

The research, described in an appendix to the main report, found that the most difficult subjects are the physical sciences, mathematics, economics, history, French, German and general studies.

Easier-than-average subjects are computer studies, business studies, home economics, design and technology, classical studies, English, communication studies and art.

The study, which included an analysis of A-level results over three years, an examination of 1995 admissions data by the Universities and Colleges Admissions Service and a review of higher education entry requirements, also confirms that pupils with better-than-average GCSE results tended to opt for sciences, and that students starting higher education courses in the more difficult subjects are among the most highly qualified at A-level.

Yet while candidates in some of the more difficult subjects enjoy the compensation of relatively low entry requirements for higher education courses, no such allowances are made in other subjects - for example history and French. Mathematics and science courses are the most difficult to fill.

The appendix, Quality and rigour in A-level studies, also quashes fears that the increasingly popular “modular” syllabuses, which are broken down into separately-examined units, are a soft option compared to the more traditional “linear” courses in which a set syllabus leads to a final examination.

Although on average students doing modular A-levels achieve better results than those doing linear courses in the same subject, the researchers say, there is little evidence that this is because of a difference in grading standards between the two systems.

The gap is partly explained by the fact that on modular courses students have more opportunities for resitting tests or retaking modules. New rules will lessen this effect by stipulating that at least 30 per cent of marks on modular courses must be accounted for by a final examination.

A study of one A-level modular examination in 1995 showed that almost all students improved their marks by resitting, and around one-third increased their eventual A-level grade.

Modular and linear A-levels should be retained, says Sir Ron, but the final exam should test understanding of the syllabus as a whole. There should be a limit on the number of resits a student can take, and a common timetable for modular exams based on two sittings a year.

In the longer term, traditional and modular A-levels could be combined, with perhaps three modules in the first half of the syllabus and a final examination covering the second half, with questions to test understanding of the whole syllabus.

Sir Ron recommends that every school and college should have a formal procedure to be followed before switching between A-level awarding bodies.

An archive should be kept of examination papers, scripts, marking schemes, coursework statistics and awards to establish a better basis for assessing standards over time. A similar archive should be kept for GNVQs.

A-level regulatory bodies should consider prescribing specific required outcomes in syllabuses.