What you call a school is academic

City academies are remarkably similar to the city technology colleges the Tories launched in the 198

The Hawthorne effect is familiar to most social scientists, particularly to psychologists and management specialists. Hawthorne was a Chicago factory owned by Western Electric. In the 1920s, it tried to raise worker productivity by changing the lighting levels, and hired researchers to monitor the effects. They found that, whether the lights became brighter or dimmer, productivity increased, unless they became so dim the workers couldn't see what they were doing. Intrigued by this result, the researchers tried other changes. For example, they shortened the working day and found output per hour rose. They shortened it further, and output rose again, as you might expect. The researchers then reintroduced the original working day and found output rose even higher.

The explanation is still widely disputed. But new Labour ministers clearly believe the most common one: that, if you tell people a change is for the better, things will duly improve. This surely explains ministers' constant fidget to reorganise health, education and other public services, to the extent that they often come full circle to what the Tories were doing, without conspicuous success, before 1997.

Take city academies, which are remarkably similar to the city technology colleges the Tories launched in the 1980s. I have never understood their point. The academies, like the CTCs, are publicly financed schools independent of local council control, and they are sponsored by the private sector, which contributes to capital costs. Why should changes in governance and financing make a difference to children's education? A council bureaucracy may be a fearful thing, but other reforms over the past 20 years have given schools considerable freedom from local authorities. The national curriculum is by far the most significant restriction. The academies are allowed to vary it somewhat - and are generally encouraged to innovate - but if this is such a good thing, why not extend it to all schools?

The first academies opened only in 2002 and 29 of the 46 (150 are due by 2009) have started in the past two years. So all academy pupils began their secondary education at another school, though it was most often a school on the same site, and nearly always a failing one in a deprived area. According to a new report from the National Audit Office, the academies' GCSE results, though still well below the national average, have improved compared with the predecessor schools and are "broadly comparable" to the results of schools in similar areas. The value-added scores (the GCSE results matched against incoming pupils' attainments) are well above the national average.

Changes in the type of pupil could partly account for it. PricewaterhouseCoopers (PwC), which is monitoring the academies for the education department, found a decline in the proportion of free school meals from 44.5 per cent to 41.6 per cent. It also found that the rate of exclusions from academies is significantly higher than for comparable schools. Moreover, the results are patchy. In the 11 schools studied by PwC, seven scored better at GCSE than before, four worse. So we don't seem to have an effect that can be attributed solely to independence and private sponsorship. If we did, we'd expect it to work across all academies except for the odd case.

What about the freedom to innovate? PwC noted that academies were starting GCSE courses a year early, using electronic whiteboards, and introducing five-term years. But few if any innovations were unique to the academies. Besides, the schools tended to row back from innovation: some initiatives "were not particularly well founded, and were regarded as having a very limited impact".

The clinching argument for academies, supporters say, is their popularity with parents. But that seems to precede their opening. Call a school an academy, give it some flash new buildings, say it's under new management, and parents will come flocking. In the marketing world, it's called rebranding and, as a result, a school should get a better pupil mix, with fewer children from "problem families", and that by itself should raise performance. The problems have to go somewhere but, if they can be spread around a bit, the education system should benefit.

That's the best I can say for city academies. I am sceptical of their supporters' claims, but almost equally sceptical of their opponents' jeremiads. The academies may benefit from a Hawthorne effect. On the other hand, they may suffer from a sort of reverse Hawthorne effect whereby teachers get tired of being continually mucked about, as doctors and nurses are now fed up with NHS reorganisations. The trouble with education is that, after all these years, nobody really knows what works and, if they do, they're not sure why. Maybe they should just try dimming the lights. At least it would be cheap.