Higher grades, better students? Or higher grades, lower standards? When more students achieve high exam grades, some claim the credit for supposedly better education systems.

Others, however, say requirements must have been lowered. They believe there is a natural ceiling to overall performance in education. But they are wrong.

We can lift our sights in education, just as in other fields. International comparisons make this clear. By showing how much better some countries do than others, they provide evidence that improvement is possible. It’s a matter of raising expectations - and expectations matter for progress, both for individuals and countries.

Ask a parent or teacher in the United States why a student is doing badly in maths, and you’re likely to hear that it’s to do with intelligence. Ask the question in Japan or Korea, and a student will be blamed for lack of effort. And in some countries, parents blame the teachers.

International comparisons can help to raise expectations. The Office for Economic Co-operation and Development’s Programme for International Student Assessment (PISA), which tracks the achievements of 15-year-olds in reading, maths and science, has done much to open up this issue.

PISA 2000 showed the UK’s average results were among the world’s best. It also showed that gaps between high and low performers, and between students from socially advantaged and disadvantaged backgrounds, were greater than in many other countries, including some with even better average performance.

UK results were “high-quality, low equity”. Other countries such as Finland, Canada, Japan and Korea showed that “high-quality, high-equity” results are possible.

Reactions to PISA have varied. Germany’s poor performance has provoked intense debate, particularly about streaming students at 11 into different types of school. To understand its PISA results better, Germany commissioned a multi-lateral study among countries with which it wanted more detailed comparisons.

Denmark reviewed its education policies in relation to those of Finland, a higher performer. Others have enhanced the monitoring of their systems.

Mexico has established a new evaluation institute, which is independent of its education department. Canada now uses PISA for monitoring language, maths and science and has switched its domestic cycle to cover other subject areas.

Some countries assess all students rather than a sample, as provided for by PISA. Thus, they can monitor the system as a whole and individual schools, where many key decisions affecting students’ learning are made.

In parallel, countries are defining expectations more clearly. Since PISA 2000, Germany has developed a curriculum framework with performance benchmarks. Spain, a relatively poor performer on average but with relatively equitable outcomes, wants to improve its higher performers by pursuing high quality, but accepts this will be at the expense of equity in the first phase.

Finland, the highest achiever in PISA 2000, defines targets centrally, provides support and monitors schools, but lets schools choose how targets are to be met.

England, after developing centralised strategies for initially successful efforts to improve performance in English language and maths, now wants to give more freedom to teachers and schools to determine the means to achieve improvement.

The US gives special emphasis to low-performing groups. It has a larger percentage of high performers than many countries which outperform it, on average, in PISA. This is due to a large percentage of low performers.

Under new federal laws, schools and states must produce overall improvements, but also for disadvantaged ethnic minorities.

So, is the sky the limit? Or is there a ceiling for quality?

Comparisons show improvement can be made. But we must improve our knowledge base about what works. Education is not a knowledge industry whose practices are transformed by systematic understanding of what works. Other areas of professional practice are influenced by research. Education must be also.

We are not without good research. The literacy strategy was research-based and effective. We need a systematic approach involving teachers in the research which shapes their practice.

This is clear in Finland, where teachers play a key part in its success.

Teaching is a high-status profession there. Entry to teaching is competitive - all involved are graduates with masters degrees. Teachers also have scope to innovate in their professional practice. The Finnish system abolished streaming and grade repetition, and students in difficulty are not passed off to others.

All 30 OECD countries and more than 20 others are already using PISA to monitor performance. Results from 2003 will be available in December this year, and work for 2006 is already under way.

Building a full picture of what works by obtaining improved results is a sure way to raise standards. This will take time and is a complex task, but without the evidence, we are vulnerable to impression and prejudice. Not the best basis for good policy-making.

Barry McGaw is director of education for the Organisation for Economic Co-operation and Development. For more on PISA see: www.oecd.orgedu