No Child Left Behind - the politics and practice of school accountability. Edited by Paul Peterson and Martin R West. The Brookings Institution, pound;24; pound;16.50 pbk.
Choice is the educational mantra of our times: based on the conviction that parents, through the operation of market forces, will drive up school standards. In the UK, it's accompanied with regulation; in the United States, which holds to its belief in the unfettered market, regulation raises hackles.
It raises questions, too. Education vouchers, paid for out of public funds, have long been canvassed as the classic choice solution. Parents who cannot simply move - increasingly, the poor and inner-city minority groups - could receive a grant to spend at the private school of their choice. State schools, faced with the loss of market share, would pull their socks up or go under.
Since 1990 the city of Milwaukee has been running such a programme - the longest established and by far the largest in the US. Initially it was confined to 1 per cent of the city's pupil population; by 1999, it applied to 15 per cent. The original proposal was that participating schools had to meet the same performance targets that were applied to state schools, but in the face of a storm of protest, that was dropped.
So does the Milwaukee voucher system work? School Choice and the Question of Accountability reports a five-year study of the programme, and bluntly answers "no". School improvement, the authors say, depends on two factors: well-informed parents, and schools that are striving to succeed. The Milwaukee experiment has neither. Well-informed parents are scarce, particularly among the target group - and even the best-informed lack adequate information about the performance of the private schools in the system. Nor, apparently, do the parents have any significant effect on those schools. There is some evidence that schools which could not survive on private fees turn eagerly to vouchers, but their performance seems unrelated to the number of vouchers they attract.
The same holds true, conversely, for the state schools. They appear to be "making changes to retain and attract students", but these changes apparently have nothing to do with any loss of market share. Funding isn't the answer because a state school's per-pupil budget allocation isn't reduced when a pupil from its catchment area takes up a voucher. How could it be? Some vouchers go to children already in private education.
No surprises there, then. Parental choice simply isn't working. No surprises, either, in the book's conclusion. What is needed, the authors say, is hard-nosed accountability: full and accurate information about every school in the programme disseminated through a public-private agency not dissimilar to Ofsted. Only with accountability, the argument goes, can market forces work.
No Child Left Behind addresses the same issue, this time over the whole spectrum of US state education. Its starting point is the 2002 Act of Congress of that name that made all federal grants to education conditional on the states determining a set of standards and tests. Any state district or school that didn't meet its standards would face - echoes of Ofsted - "corrective action". It could have been, the editors say, the most important piece of education legislation for 35 years: their purpose here is to track its genesis, as a rarely bi-partisan political compromise, and to evaluate its likely impact.
The first book, fascinating as it is, is mainly for students of politics.
The second, though, is for students of education policy: a revealing indication of how accountability, rather than markets, has become the current mantra.
The thesis is that the act falls short on accountability. What is needed, the editors say, is "tough, coercive, high-stakes accountability". That's what all the states promise, but as coercive teeth begin to bite, self-interested pressure groups (teachers, mostly) lean on politicians, so standards are lowered and penalties postponed.
Failing schools also have to be closed and "reconstituted" - but annual fluctuations in state test results always yield a let-off for such schools.
And accountability at student level is also vital: tests that incur real penalties for inadequate performance. Something like the GCSE in England or the baccalaureate, the researchers say, is what is needed.
Given the echoes of what's been happening in England, it makes intriguing reading - because of the hints in several chapters that the issues are less clear-cut than the summary suggests. There is evidence, for instance, that where teachers have signed up to it, nice, soft, low-stakes accountability produces as much improvement as the other sort. Revealingly, even the central premise - that there has been and is "systemic failure" in the nation's schools - is hardly watertight. "It is as easy to make the claim," contributor Jennifer Hochschild argues, "that American schools have been improving over several decades as it is to claim that they have been worsening."
The one issue that isn't challenged - and it's as American as apple pie - is high-stakes testing. At the end of the day, it's testing that lies at the heart of accountability. It may be imperfect - but imperfect accountability, the consensus is, is better than none at all.