The exclusion of the United Kingdom from this week’s report on the Pisa 2003 survey is a bitter setback for a Government that has embraced the idea of international benchmarking of our educational performance far more enthusiastically than its predecessors.

The disappointment arises partly because it makes it harder for Tony Blair to rave about the UK’s strong results in Pisa’s first round, carried out in 2000. It is also because ministers and their advisers have genuinely bought into performance monitoring as a driver of improvement. Perversely, our obsession with assessing performance has probably contributed to our exclusion from the survey results.

The Programme for International Student Assessment, run by the Organisation for Economic Co-operation and Development, is the most important regular global survey of student performance. It tests the extent to which 15-year-olds can muster reading, maths and science skills to tackle tasks they will confront in adult life.

It is repeated every three years but, in the latest survey, the UK failed to meet the technical requirements for inclusion in the main results, because more than a third of schools and nearly a quarter of students selected for the survey did not take part. The main performance results for the UK as a whole are therefore not being published, although Scotland’s results are included, showing that it continues to perform above average for the OECD.

What went wrong? The Department for Education and Skills has commissioned a survey to find out, but several things are already clear. Unlike in some other countries where education ministries more or less ordered schools to take part, heads in the UK’s relatively autonomous schools felt able to say no to the survey.

They seem to have become increasingly reluctant to take part in international surveys, given the heavy pressure of UK-based testing. Many British 15-year-olds, unlike those in other countries, were approaching a high-stakes national exam at the time of the Pisa study in March 2003.

Two-thirds of those eligible were Year 11s taking GCSEs.

After insufficient schools took part in March, the UK was given a chance to test students in more schools in July. But by then many Year 11s had disappeared on jaunts to Newquay or beyond. There is no way of knowing whether the students who did take the test are representative of students as a whole.

The irony is that the Pisa results overall suggest that education systems tend to do better if they give schools managerial autonomy and monitor student performance through regular assessment. The very things that the OECD has praised in countries such as the UK are making it harder to collect international data here. It is no coincidence that the one country failing to meet the sampling requirements in the earlier Pisa 2000 survey was the Netherlands, another country where school autonomy and accountability are both emphasised.

The DfES can now learn from the Dutch authorities, which made a great effort to convince their schools of the benefits of Pisa, and succeeded in getting the required response to the 2003 survey. The UK has 15 months to achieve the same in time for Pisa 2006.

In the meantime, there is still much that we can learn from the results of other countries published this week. In 2003, the main assessment was of mathematics. Here are four important lessons:

* Levelling up - to narrow the gap between the best and worst - is an achievable goal. In general in Pisa, high overall performance does not come at the expense of equality.

Among the four top-scoring OECD countries in mathematics in 2003, Finland has the narrowest gap between the best and worst students of any country.

Korea and the Netherlands have greater than average equality, and Japan an average distribution.

While country scores have not generally changed much since 2000, the best improved in mathematics, Poland, is a case study in levelling up. Having been one of the poorest performers in earlier international surveys, it is now almost at the OECD average, and on a par with the United States.

The recent improvement has been achieved entirely through a rise in the performance of Poland’s lower achievers - the better students were already scoring close to their equivalents elsewhere in the OECD, and their scores have not changed.

In the late 1990s, Poland reformed its secondary school system, creating an extra year of comprehensive general education at age 15, rather than dividing up students between school streams after an extended primary education. This has clearly helped under-achieving 15-year-olds to catch up with their peers in other countries.

* The most successful secondary school systems are not selective. The Pisa 2000 results showed that countries such as Germany that divide secondary children into different status schools not only perform unimpressively overall but also tend to end up with the biggest differences in achievement between advantaged and disadvantaged pupils.

Following the storm unleashed in Germany’s education system by its below-average Pisa 2000 results, the country has shown a modest improvement overall, with results now around average. This is likely to result from a greater focus by teachers on performance, rather than on fundamental reforms provoked by Pisa, which are only just getting underway. However, Germany’s divided schools continue to produce wide social differences.

* The gender gap in mathematics is today more important at a psychological than at a cognitive level. While boys continue to score better than girls in many countries, the difference is now very small. In contrast, girls questioned in Pisa showed substantially less interest, less confidence and greater anxiety in relation to mathematics than boys.

Figures on gender differences are among the selected UK results that the OECD has published, and these indicate that in this country there is no significant gender gap in maths performance, but a wider than average gap in attitudes. This underlines the need to focus on building girls’

confidence in the mathematical abilities that they are now clearly displaying.

* Narrow tests of achievement are not the only measures of the outcomes of secondary education. The most innovative aspect of Pisa 2003 was to try for the first time to measure general skills internationally, by testing students on their problem-solving abilities.

Based on tasks such as deciding how to trouble-shoot a malfunctioning fridge using an instruction manual, or planning a travel itinerary under complex constraints, this module demonstrated that it is possible to make robust comparisons of such skills across cultures.

Although in most cases country rankings for problem solving were similar to those for reading, mathematics and science, this was not always so.

Analysis of the results confirms that Pisa as a whole is measuring something other than simple mastery of the curriculum, the focus of the more traditional Trends in International Maths and Science Study being published next week.

For secondary heads with an interest in new approaches to assessing outcomes in 14-19 education, this may be the most important reason for participating in Pisa next time.

Donald Hirsch is an international consultant on education policy