A-levels of years ago.

But each year the Government argues that standards are improving, using increased numbers of A-C grades at GCSE and A grades at A-level as evidence. Yet a new A-level grade of A* is being introduced to test the most able students, and universities talk of entrance tests that go beyond the standards of public exams.

Confused? Yes, we are, and who wouldn't be?

A petition organised by the Royal Society of Chemistry has been delivered to Downing Street and signed by 1,600 people, complaining that record-breaking exam results are "illusory". The RSC says even the brightest students are merely taught to "learn to the test".

Pupils are drilled to answer "undemanding questions to satisfy the needs of league tables and national targets," says the RSC, but they are not learning how to solve problems, use critical thinking or apply mathematics in science.

The Royal Society also delivered a report to every MP, as well as chief executives of FTSE-350 companies, schools and universities in which it set out to demolish what it calls the myth of record-breaking science education reports.

The RSC carried out an experiment by setting today's brightest students an exam made of up questions from exams set over the last five decades. Students scored highest on the most recent questions, and worst on those from the 1960s.

Royal Society of Chemistry chief executive Richard Pike said: "The target of our campaign is a failed education system, not the youngsters it's supposed to serve. We know that enthusiastic teachers are being compelled to 'teach to the test', to meet the demands of school league
tables, which involves mainly the recalling of facts, with no reference to logic or mathematics.

"That means the brightest pupils are not stretched, or trained in mathematical techniques, because they can get a grade A* without doing a single calculation. Conversely, the majority get at least a 'good pass' (a C grade) showing merely a superficial knowledge on a wide range of issues, but no understanding of the fundamentals."

The Government continues to insist that standards are rising, after 10 years of investment and improvement in teaching and education generally. And, in a barb aimed presumably at the RSC, a spokesman said: "Exam standards are rigorously maintained by independent regulators, and we would rather listen to the experts whose specific job it is to monitor standards over time."

So the ordinary reader or parent is none the wiser about what truths underlie this war of words over science. Perhaps Professor John Holman, director of the independent National Science Learning Centre based at York University, can help us to extract from these conflicting arguments some residue that is actually scientifically correct and irrefutable.

"The Royal Society got a lot of bright youngsters to volunteer for the tests they did, and yes, they did worst at questions from the 1960s which were quite mathematical, but much better at the questions from recent exams. The conclusion that was reached, therefore, was that exams today are easier.

"But the comparison is not a valid one. The only proper way to compare properly would be to have a time machine and give kids from the 1960s tests done by young people today. The RSC experiment is an interesting starting point for discussion, but one for which you can't make a control."

Prof Holman argues that course content, teaching methods and exams themselves have "got different" rather than easier.

"In 1953, Mount Everest was climbed for the first time, and now it is climbed every year by thousands of people. That doesn't mean the mountain is not as high and difficult... it just means that people have focused and concentrated on methods of doing it.

"It's the same with learning and examinations. I've taught chemistry over several decades, and I don't believe there was ever a 'golden age' of achievement in science learning. Decades ago, teachers taught youngsters 'to the test' as much as they do now, but methods were different and good results were achieved more by the remembering of facts than they are now.

"Where I do agree with the RSC is about the mathematical content of science education now. There should be more of it in all sciences, because the ability to do maths well is crucial."

Rather than contrasting performance at exam questions set in different decades, more valid comparisons of standards are those made internationally, says Prof Holman. Between 2000 and 2007, the UK dropped from fourth to 14th in science achievement by secondary school students, according to assessments carried out by the OECD PISA programme.