How do they define failure, then?

On Monday the figures for last year's GCSE results were announced. There was no great fanfare, from which you would be right to infer that there was little progress, if any, to report.

Concentrating on differences achieved between pupils on free school meals, whom I shall call "poor" as a shorthand, and others, a number of things stand out.

Among these poor children 21% obtained the Government benchmark of five passes, grade A to C and including maths and English. This compares with 49% for other children. This is a very large difference, and only a slightly smaller difference than in 2002, when the vast sums of money were beginning to arrive in education.

At age 11, that is when leaving junior school, the two groups had different achievement levels in reaching the required standards in English and maths. In 2002 the poorer group results in the two subjects were 26 percentage points and 16 percentage lower than in the other group. When that cohort took their national curriculum tests at 14 years, the gap was 27 points in each subject. They poor actually deteriorated in achievement.

Some other statistics may explain this. In secondary schools in areas of recognised social deprivation 12% of pupils are likely to be classified as persistent truants, while in the least deprived the figure is 2%. ("Persistent truancy" means a pupil who has missed at least one fifth of the school sessions in a year, or 63 sessions.)

A similar result arises in calculating suspensions for physical assaults on adults. In the 10% of schools with most poor pupils in 2006 there 1,700 such suspensions, while in the 10% with fewest school meals there were 210 assaults.

The picture arising is of schools with higher concentrations of poor pupils producing proportionately many more unsatisfactory outcomes than other schools in the two core subjects, and that in secondary education high concentrations make the gap even wider. The possible reasons are the incidence of truancy and the atmosphere of assault, both of which could also be a result of under-achievement.

Clearly so far large sums of money and frequent policy initiatives have had little effect. It is surely time to apply the two recent principles of Conservative thinking.

1) Discipline at school must be improved. This could be done by making exclusion trigger expensive attention to offenders, rather than by reinstating them after appeal, either in the same school or another.

2) Societal breakdown comes largely from family breakdown, - a finding associated with Ian Duncan Smith. This is much more difficult to deal with, and will take longer to achieve, but unless family life styles, attitudes and behaviour change the problem will be largely intractable.