The question is simple: the answer complex. Could we have an education service in which there are no more failing schools? Put another way, could we say we want “successful schools for all” and really mean it?

The question is complicated not only because creating success in education is a complex matter but also because people in education tend to shy away from serious debate about failure. It has been, at least until recently, what Roland Barth, professor of education at Harvard, would call a “non-discussable”.

Paradoxically, I have chosen this opportunity to promote what I hope will be a frank, open, serious debate about failure because the education service appears to be increasingly successful. Examination results and participation rates have risen steadily for several years. OFSTED has recorded an improvement in the proportion of lessons that are satisfactory or better.

Meanwhile, at all levels in the education service - nationally, locally and in schools - there is a huge growth of interest in the research findings about school effectiveness and school improvement. The work of Peter Mortimore (who gave the first TES Greenwich lecture), Tim Brighouse (who gave the last) and many others in the field has never been in such demand.

This growing interest is hugely encouraging. Maybe cultural attitudes to failure are at last changing. For too long, it has been a powerful strand of culture in this country that failure in education is inevitable and that, like the poor, it will always be with us. There is another strand which defines success by the extent of failure. According to this view, without failure there is no success. Hence the annual attack from some quarters each August on the GCSE examination: if more people have succeeded, standards must have fallen.

We, as educators, might be expected to be in the vanguard of a campaign to challenge this poverty-stricken culture. In fact all too often we reinforce it.

When the education service is accused of failure, the tendency among educators is to deny it or minimise it. When David Blunkett, Labour’s education spokesman, raised the issue of failing schools recently, one union leader was quoted as saying that if schools had sufficient support from government they would never fail. This view is untenable. By suggesting that the only variable is government support, it implies that the quality of teaching is insignificant. It therefore belittles, by implication, the entire profession.

Another union leader argued that “dealing with a small number of schools in deep trouble is not the important agenda”. For the pupils and parents concerned, it is the only agenda. More fundamentally, it needs to be recognised that a small number of failing schools undermine both the teaching profession and education as a public service. Their existence reinforces the negative cultural attitudes that so urgently need changing.

When in the autumn of 1994 Chris Woodhead, the chief inspector of schools, remarked that 30 per cent of teaching was unsatisfactory or poor, I was guilty of a similarly inadequate response. Why doesn’t he point out that 70 per cent of teaching is satisfactory or better, I wondered? On reflection, I could not help thinking that this reaction was both defensive and complacent. Was I really willing to accept that 30 per cent of lessons are inadequate? Is this not playing into the hands of those who believe that failure will always be with us? Does this not fuel what Stephen Ball, professor of education at King’s College, London, has called “the discourse of derision”?

Surely we, as a profession, should be concerned about even 10 or 20 per cent failure, never mind 30 per cent. Our goal should be steadily to reduce the extent of failure. In the language of business, we should aim for ever higher reliability.

We would, after all, be appalled if an air traffic controller attempted to reassure us by saying that the other nine planes had landed safely. While managing a successful school is infinitely more complex than controlling air traffic, failure in education is as catastrophic in its consequences as failure in the airline industry. It differs only in that it happens more slowly and that no one has yet made the movie.

So far, 50 schools have been identified as failing, schools where at least 8,000 children are being failed. Of these, 19 are secondary schools, two are middle, 26 are primary and three are special. By no means all of these are in urban areas; several, for instance, are in Norfolk. At present, the proportion of schools found in their OFSTED inspection to be failing is running at between 1 and 2 per cent. At this rate, by the time OFSTED has completed its four-year cycle, there will be between 250 and 500 failing schools in the country. Another 5 to 10 per cent of schools, according to OFSTED, have serious weaknesses but are not technically failing, a rate which, if maintained, would cover 1,250 and 2,500 schools in total. They are “struggling” schools. The chief distinction between “failing” and “struggling” is in the relative capacity of the school to improve from within. In the former, there will be little or no evidence of this capacity.

There is, for obvious reasons, little research evidence on failing schools. The assumption tends to be that they are simply schools without the characteristics of effective schools. David Reynolds of Newcastle University suggests this is too simplistic. Failing schools, he argues, have characteristics which actively militate against improvement. They have what he calls “antithetical characteristics”. Staff tend to blame pupils and the community for failure. They are fearful of outside intervention and will counter criticism by saying “We’ve always done things this way”. There are personality clashes and feuds among staff. Good new ideas are met with cynicism and there is resistance to change, anxiety about taking risks. Effective staff in these circumstances often hide behind the norms of an ineffective group.

If Reynolds is right, then the improvement strategy must take these “psychological” problems into account. Clearly there is no justification for allowing a school in this state to continue. It is a gross infringement of the entitlement of young people. Equally, the evidence suggests that traditional support and advice, even if it is backed by funding, will not be enough.

The question is, then, what policy for failing schools would enable pupils to receive the education to which they are entitled? This 10-step strategy is an attempt to answer that question. Steps 5 to 7 specifically apply to struggling schools; 8 to 10 are for failing schools.

1 There are several preconditions of a strategy for failing schools. Clearly, a national policy which promotes school improvement and provides steady investment is important. It is also essential to have an effective means of identifying failing schools. This requires improvements in the consistency of OFSTED inspections and as much clarity as possible in the OFSTED framework about what constitutes success. Improved value-added analysis of examination results would also help. OFSTED’s commissioned work on this from the Institute of Education in London needs to be built upon, since uncontextualised raw results are not an adequate guide to either failure or success.

2 OFSTED’s resources need to be targeted more effectively at schools which are less successful. For most schools the inspection system should consist of a validation check on their own self-review, which itself ought to be based on a revised OFSTED framework. A full-scale inspection would then only be necessary in three circumstances: where the check revealed the self-review to have been inadequate; where there was evidence that a school did not have the capacity to drive its own improvement; or where its performance, as revealed through examination results, truancy figures or other hard data, appeared to be inadequate.

The reduced level of inspection for most schools would make it possible for OFSTED to put more resources into the inspection of less effective schools and into assisting schools in the development of constructive post-inspection improvement strategies.

3 Over time, the standards expected of all schools should be steadily ratcheted up. This would be done through periodic revisions of the OFSTED framework. OFSTED would continue to focus on the least successful schools. As David Hargreaves, professor of education at Cambridge, suggested recently: “OFSTED should identify and promote help for the worst schools, since only by progressively eliminating the worst schools - an endless task - can national standards and levels of achievement indisputably rise over a sustained period.” One might call this the Reading Recovery approach to school improvement.

4 The first three steps are designed to ensure effective identification of failing schools and sensible targeting of limited national resources. The remaining steps look at possible interventions in schools which are either struggling (steps 5 to 7) or failing (steps 8 to 10). This one - step 4 - proposes a policy principle on which the remaining proposals are based. It is that external intervention in a school’s affairs should be in inverse proportion to its success. The general assumption behind this principle is that most schools have within them the capacity steadily to improve themselves, as long as national policy provides a sensible policy and funding framework.

Any intervention needs to be carefully planned and based on knowledge of the school effectiveness and school improvement research. Its purpose needs to be clear: improvement cannot be imposed. The role of an intervention is to generate within a school the capacity for sustainable self-renewal, and to unblock the barriers to improvement.

At present there is a problem in the case of schools whose performance appears to be deteriorating between four-yearly OFSTED inspections. While a local education authority may decide to inspect a school at its own initiative, its powers to intervene remain limited - whatever it may find. For this reason LEAs should be given a new power to trigger an OFSTED inspection where there is objective evidence that a school’s performance is deteriorating. This power would give LEAs leverage over such schools and would rarely need to be used in practice.

5 For schools found by OFSTED to be struggling, the quality of the action plan is more important than it is for more successful schools. Yet a struggling school, by definition, is less likely to be able to produce an effective plan. Each school in these circumstances should receive a limited sum of money specifically for the purpose of enabling it to purchase external expertise in the action planning process. The redirection of OFSTED’s resources, suggested earlier, would make this possible.

6 As experience of action planning develops, there should be increasing research evidence of the characteristics of effective post-inspection action planning and implementation. The experience in both this country and abroad suggests that action plans need to be a judicious combination of the latest research evidence on school improvement and knowledge, and detailed understanding of the school in question. For this reason, there can be no simple blueprint. It is vital that a school believes in, understands and is committed to the plan that emerges from the process.

7 In the case of some struggling schools, it is evident from the inspection report that their poor performance is due to the significant underperformance of a relatively small number of members of staff. There should be a formal post-inspection procedure for dealing effectively and fairly with failing teachers and headteachers.

8 Failing schools differ from struggling schools chiefly in that their capacity for self-improvement is strictly limited. A failing school is therefore, by definition, one that is providing an inadequate education and unable, without external intervention, to improve.

Closure and the dispersal of pupils is one option to consider in these circumstances. David Blunkett created a storm at the Association of Teachers and Lecturers’ conference when he suggested closing failing schools. In fact what he said about closure (as opposed to Fresh Start) is nothing new. It is happening now. Six of the first 14 schools found to be failing are in the process of being closed. If there are sufficient alternative places available in good schools within the area, it may be the most sensible option.

9 A second option is to develop an effective action plan as a result of discussion between OFSTED, the school and the LEA, to replace staff in cases where there is clear evidence of serious under-performance and, possibly, to invest some limited additional resources in specific aspects of the improvement strategy.

Progress, in these circumstances, would need to be monitored carefully. Schools need success criteria related not only to academic performance, which often takes some time to improve, but also pupil behaviour, truancy, pupil and parent attitudes and staff development which may be more amenable to short-term change. There should be evidence of a renewed sense of direction within months, of progression against some of these targets within six to 12 months, and of improvements in terms of pupil outcomes within two years. Each failing school’s action plan should be tailored to its specific circumstances.

In the case of some failing schools this kind of approach has worked over the last year and several more are likely to “come off the list” within the next few months, following the example of Brookside special school in Derby which was removed just before Easter. Each is a credit to the Department for Education and its School Effectiveness Division, OFSTED, the LEA concerned and, above all, to the staff of the school.

10 There is, however, a small number of schools where the improvement strategy has not worked. In these circumstances a school has been found to be failing and is still failing a year or more later. If the needs of the pupils are to be the prime concern, this is a situation in which urgent action is needed.

Closure and dispersal might again be appropriate. But what if that would mean pupils making long, inconvenient journeys? Would they be likely to do so after their experience of education? Or would many of them join the ranks of “the disappeared”, those teenagers, between 5 and 10 per cent of their peers in some urban areas, who cease to attend school and turn up next, if they turn up at all, in the criminal justice statistics?

In these specific circumstances, David Blunkett’s Fresh Start idea makes sense. The LEA would close the school and re-open a new school on the same site. The change would have to be done quickly, perhaps over an extended summer break. It would also have to be more than cosmetic.

What would this new school need? It would need new governors; professional leadership of the highest calibre; resources to attract new staff to what would inevitably be a risky venture. It would also need to draw on the latest research and use innovative approaches to teaching and learning. It would need to be perceived by its staff and, above all, by pupils and the community, as an ambitious and exciting attempt to create success and to make a clean break from the troubles of the past.

Some critics have already condemned this idea. It won’t work, they say, but their case is fundamentally flawed. First, it has been done successfully in America and, in effect, here too. In 1988 the Inner London Education Authority intervened dramatically, controversially and, in the long run, successfully at Highbury Quadrant Primary School. Hammersmith and Fulham LEA has acted boldly in the case of the Hammersmith School within the past few weeks. Neither of these was a fully-fledged Fresh Start (not least because the law in its present form makes Fresh Start difficult), but they show it could be done.

Second, the critics of Fresh Start offer no alternative. By implication, therefore, they are prepared to tolerate continuing failure in the circumstances I have described. On the brink of the 21st Century, this surely is unacceptable.

The traditional response to those who raise the question of failure is to suggest that we should discuss the success of the many, not the failure of the few. This is a classic false dichotomy which ought to be unmasked. A serious debate about failure is, in fact, a precondition of success. If we really want success for everyone, we should stop pretending that failure does not exist.

Michael Barber is professor of education at Keele University. Copies of the full text are available from The TESGreenwich Lecture, TESPromotions, Admiral House, 66-68 East Smithfield, London E1 9XY. Please enclose a cheque for Pounds 3.50 made payable to The TES.