Nelson needs to read the research and try something radical with education

The recent entry by Federal Education Minister Brendan Nelson into the school-quality debate has shown that he is willing to wrestle with his state colleagues over issues traditionally outside the ambit of the Federal government.

In a carefully researched speech, Nelson touched on issues of closing failing schools, implementing nationwide standards and publishing more data on schools.

But what we've seen so far can only be described as a "half-Nelson", a collection of ideas that may make some small difference but are unlikely to really boost the quality of our schools.

If Nelson is serious about being remembered as one of Australia's great reformers, he should use his position as the nation's most senior education policy-maker to present some truly radical ideas about education.

He should tackle the issue of low school completion rates. Today, only six out of 10 Australians have completed high school, compared with eight out of 10 Canadians, Germans and Americans.

The school-leaving age in most Australian states is 15, unchanged since the 1960s, and below the leaving age in most other developed countries. Researchers have shown that teenagers often make short-sighted decisions to drop out of school. Where compulsory schooling laws bind, an extra year of school boosts lifetime earnings by 10-15 per cent. Australia should follow the lead of several other developed countries and raise the school-leaving age to 17.

Next, he should improve school quality by engendering a culture of policy experimentation. Australia has never conducted a randomised trial of a serious education-policy initiative. Forced to choose between inaction and full-scale implementation, many promising policies are simply scrapped. Yet finding out what works, and what does not, allows us to better direct scarce education resources where they can do most good. The Federal government should provide fiscal incentives to state governments to test creative policies and assess them through randomised evaluations.

Politicians should also speak the truth on class-size reductions. Nelson ought to encourage Australian politicians of all stripes - particularly his NSW Liberal colleagues - to read the research on class sizes. Although intuitively appealing to voters, careful studies have shown that once class sizes drop below 30, there is little benefit from further reductions.

To see how this might be the case, imagine an analogous situation in the medical world. If we reduced the doctor-patient ratio in a ward from 1:25 to 1:20, we might improve patient care but we could merely increase doctors' leisure time.

Likewise, research suggests that lower class sizes make teachers happier but do little to improve student learning. In one of the leading class-size studies, Harvard's Caroline Hoxby looked at variation in student achievement from natural fluctuations in Connecticut class sizes and found that the impact of smaller classes was precisely nil.

Another prominent study, by Princeton's Alan Krueger, looked at a randomised trial in Tennessee. While Krueger found that smaller classes improved student achievement, the magnitude of the effect was minimal. Reducing class sizes by 25 per cent brought about a tiny gain in test scores - only one-fifth as large as replacing a low-performing teacher with a high- performing one. It is time Australian politicians climbed off the class-size-cutting bandwagon.

A more effective reform would be to make teaching more rewarding by raising pay rates for our best-performing teachers. Each of us remembers the teacher who inspired us to read a new author, helped us understand calculus or made history come alive.