The Lord Chancellor and the Lord Chief Justice must sleep soundly at night in their grandly wallpapered piles. Neither - to judge by their pronouncements - expects to be woken by thieves prowling around their homes.

Lord Irvine of Lairg weighed in yesterday on the side of Lord Chief Justice Woolf by calling for fewer criminals to be sent to jail. He said they should be given 'community sentences' instead. Both men exude a complacency that goes with a chauffeur-driven lifestyle and a smug contempt for the 'unsophisticated' views of ordinary people.

They believe from the safety of their Establishment bubble, that prison fails to rehabilitate offenders. About 55% of inmates are re-convicted within two years of release from jail and, therefore, according to the noble, lords prison does not work. But they neglect two facts.

First, the re-conviction rate two years after completing community sentences is about the same as the rate for jailed offenders. Second, offenders undergoing community sentences are free to commit crimes during their sentence, whereas prison inmates cannot break into houses or steal cars. Their lordship forget that protecting the public is as essential a part of the justice system as transforming rogues into good citizens.

True, prison has not proved to be a particularly good way of reforming criminals, but it does protect the public. Instead of increasing the use of community sentences, it would be more rational to seek out a method of dealing with crime that achieves both aims: protecting the public and more effectively encouraging criminals to lead a law-abiding life on release.

Lords Woolf and Irvine appear to be unaware that in America more effective methods are being evolved to achieve better rehabilitation results without sacrificing public protection. In America, the first aim has been to get public protection right, and experience over the last 20-30 years has shown that increasing the risk of punishment reduces crime, partly through deterrence and partly by 'incapacitating' offenders.

A study carried out by the US Department of Justice found that, from 1981 to 1995, the risk of imprisonment increased in the USA and the crime rate fell; while in England and Wales the opposite happened: the risk of imprisonment fell and the crime rate increased.

Take burglary as an example: in the USA the number of imprisoned burglars for every 1,000 alleged burglars, increased from 5.5 in 1981 to 8.4 in 1994. In England and Wales the number of imprisoned burglars per 1,000 alleged burglars fell from 7.8 in 1981 to only 2.2 in 1995.

And guess what happened to the burglary rate? In the USA, burglaries per 1,000 households fell by about half from 105.9 in 1981 to 54.4 in 1994. In England and Wales, burglaries per 1,000 households increased from 40.9 in 1981 to 82.9 in 1995. Despite falls in burglary in this country in the last few years, we are still more likely than the Americans to have our houses broken into.

But the American strategy is not about locking people up and throwing away the key. They have made sustained efforts to find ways of changing offender behaviour. The results have not been stunning, but they have been worthwhile. Offenders are often impulsive, aggressive, or unable to put themselves in other people's shoes. They may be the kind of person who shouts 'What are you looking at?' if you accidentally catch their eye in the street. The most effective rehabilitation schemes rely on clinical psychologists working with each individual to change the attitudes that incline them to crime. By working with them over a prolonged period, psychologists have found that they can reduce offending by about ten percentage points that would mean a reduction in this country from about 55% to about 45%. Some of these schemes have been effective outside prison, but they work best in jail where strict compliance can be ensured.

One of the more effective US schemes has been carried out in Delaware. It takes some of the hardest cases, criminals with a drug problem, and puts them through a programme lasting 12-15 months just prior to release from jail. Even so, it alone only made a small difference. About 70% of participants were back on drugs within 18 months, compared with 81% of a control group of similar offenders. Most offenders, without continuing support, soon slipped back into their old ways. However, when combined with a follow-up programme on release from jail, the results were more striking. It required offenders to be in residential care for the first three months, where training concentrated on job readiness, interview technique and preparing job applications. For the next six months, offenders had to return once a week for group support sessions, not unlike Alcoholics Anonymous programmes. Of those offenders who had undergone the in-prison programme of 15 months and the follow-up regime, only 24% were back on drugs within 18 months.

Contrary to Lord Woolf's beliefs, the evidence does not suggest that we should put fewer people in jail, it implies that we should impose longer sentences but simultaneously re-focus the prison authorities on the effective rehabilitation of offenders combined with strict follow-up after release. We should not be jailing fewer criminals, but more of them - and for longer too.