Initial teacher training is in a mess, in part because politicians abolished quality control. Michael Barber proposes a fresh route to rigour and relevance.

Most of the 10,000 new entrants to the teaching profession each year arrive there having taken part in a course at a university and then a one-year PGCE course, two-thirds of which is spent in schools. A substantial minority, most of whom enter primary education, has done a four-year BEd or BA(Ed), generally at one of the new universities. They too, will have spent a substantial amount of time in schools. Finally, there is a small minority of entrants who have joined the profession through a training course based solely in schools. Amazingly, the Government, when it created this route into the profession, christened it SCITT, which stands for school-centred initial teacher training.

Until a few years ago, trainees had to complete a probationary year. Depending on the school and the local authority involved, this was a more or less rigorous procedure. During that dreadful time when Kenneth Clarke and John Patten were responsible for education which I call the era of free-market Stalinism, this control on quality was abolished. If a school wanted to keep a teacher it would decide to do so, the thinking ran. If not, it would let them go. It was a simple matter of allowing the market to operate. This missed the point. It is one thing to decide whether a particular teacher is right for a particular job or whether the school can afford them: it is another to decide whether that teacher is fit to be a member of a noble profession, regardless of the short-term pressures on the school concerned.

Once the probationary year had been abolished, there were no further controls on professional competence from the profession or anywhere else. This seems to border on the foolhardy. Successful completion of a one-year course of training or a four-year degree course including training can hardly be considered a sufficient guarantee of a teacher's ability to succeed throughout a career. I would like to see a complete rethink of entry into the teaching profession which would be simultaneously looser about course design and content, but tighter about outcome.

ÖйúAƬ

Modules in education and in teaching and learning (what German universities call Pedagogik) should be available to the vast majority of undergraduates whatever their course of study and should be known as "Standard Credits".

These might be attractive to students whether or not they were sure they wanted to become teachers. An understanding of education and teaching and learning will, after all, be central to many walks of life - certainly business, the professions and the civil service - in a learning society. Thus, in a three-year undergraduate degree in, say, English literature, a student might do two education-related modules out of a total of 18. Through this approach, our ÖйúAƬ system would be preparing a pool of potential educators. Some of these would actually become school teachers, teachers in other sectors or managers with responsibility for promoting learning across a company. Others would, hopefully, apply their knowledge of education and learning in whatever career they chose. As the importance of learning in every workplace grows, this would become increasingly beneficial to society as a whole.

ÖйúAƬ

Students who have successfully completed their degree, including the standard credits, and want a career in teaching, should join a three-year entry programme or apprenticeship, which could start or finish at any time of year. A person who wanted to add the standard credits to their qualification could do so at any stage.

The apprenticeship would involve a combination of practical experience in more than one educational institution and the completion of a range of courses in educational and learning theory which would be taught by universities. The precise order in which those were undertaken could be left - within a broad framework - for the individual apprentice and the institution which employed them to decide.

At the end of the three years, the performance of the apprentice would be assessed in three different ways: there would be a report from a mentor in the institution where the apprentice worked; there would be a report from an independent assessor (for example, a teacher from another school who had been trained for the role), who had watched the apprentice teach; and there would be both an extended dissertation and a written examination, which focused on educational and learning theory and was set by the university with which the apprentice was registered.

Those apprentices who successfully completed each of these three assessments would become fully fledged teachers. The salary would rise modestly on successful completion of the programme. In this way, the best elements of a school based approach could be linked flexibly to the crucial aspects of initial training that only universities can provide. The dissertation and examination would enhance the status among aspirant teachers of a knowledge of the research.

ÖйúAƬ

After two or more years in the profession, new entrants would, like every other teacher, face a five-yearly assessment of their professional competence by their peers under a process laid down by a General Teaching Council, which we might call an MOT. The first MOT, however, should be considered a major step forward. The teacher would by then have completed not only a rigorous three-year apprenticeship, but also two years as a fully functioning teacher. Once the first MOT was completed, the beginning teacher would step on to the fully professional salary grade.

These proposals would have numerous advantages over the present arrangements. They would provide rigour and strengthen quality control. They would link theory and practice integrally from the outset. They would enable mature students to earn a living during their training and, crucially, they would provide the salary progression, after three and five years, that is so obviously missing at present.

Michael Barber is professor of education at the University of London Institute of Education. His book, The Learning Game: Arguments for an Education Revolution, is published by Victor Gollancz, Pounds 25 and Pounds 12.99.