A social divide based on merit: There is a demonstrable link between class and intelligence, argues Peter Saunders

Is there a relationship between intelligence and social class? This is the delicate question addressed by the American right-wing policy analyst Charles Murray in his new book, The Bell Curve. Predictably, reaction has been fierce, and mostly hostile.

For the past 10 years, Mr Murray has been warning of the emergence of an ill-educated, dependent and disaffected 'underclass' in America and Britain.

Now he has gone one step further, suggesting IQ tests show that the American underclass is on average less intelligent than other people, and that this difference is reproduced from generation to generation.

In one sense there is nothing new about Mr Murray's findings, for we have known for a long time that IQ scores vary on average between social classes.

When Britain still had the 11-plus examination, children of professional and managerial parents recorded average IQ scores of 113, compared with an average of around 96 for the children of unskilled manual workers. Similar differences have been recorded in the US and elsewhere.

Many people, however, do not accept IQ scores as a valid measure of intelligence. They claim that lower-class children are not less intelligent; they simply perform worse in tests which are biased towards those from middle-class backgrounds.

Since the argument has never been resolved between people who accept IQ tests as valid and people who do not, there seems little point in Mr Murray using such tests to investigate differences in intelligence; the validity of the results will always be disputed.

There is, however, another way of analysing the link between social class and intelligence which does not rely on measuring IQs. Instead, we begin by comparing the proportion of working- class children who make it into the middle class with the proportion of middle-class children who manage to stay there.

A survey of 10,000 men carried out by researchers at Nuffield College, Oxford in the 1970s found substantial movement between the classes in Britain, but also revealed a clear link between social background and occupational success. Nearly six out of every 10 boys born to middle-class fathers made it into the middle class themselves, but this was true for fewer than two in 10 boys from the working class. Middle-class children are thus three to four times more likely to succeed.

The Nuffield team believed that such a large disparity must be due to the influence of social background. This was, however, never demonstrated, for the study did not attempt to investigate the intelligence of the people it surveyed. We must therefore consider whether a difference in success rates as great as 4:1 in favour of middle-class children could be explained by middle-class parents producing more intelligent offspring.

Unless we are willing to assume that we are all born with exactly the same intellectual capacity (in which case any of us could have written the plays of Shakespeare or discovered the theory of relativity), we have to accept that innate differences of intelligence exist, and that such differences will tend to be passed on from parents to their children. Of course, just as tall parents can sometimes produce short children, bright parents can sometimes produce dull ones. The point, however, is that on average, intelligent parents are more likely to produce bright offspring.

This has a crucial implication. Imagine a society where competition for jobs was genuinely open, and where the brightest children were recruited to the top jobs irrespective of their social origins. In such a society, not only would the middle class be more intelligent than the working class, but it would also tend to produce more than its fair share of bright children who would themselves be eligible for middle-class positions. In a truly meritocratic society, we would thus expect to find some association between class origins and occupational success.

How strong would this link be? Could it produce differences in the order of three or four to one?

Fifty years ago, the professional middle class in Britain was quite small - about one in seven of the workforce. Had entry to the middle class depended solely on intelligence, it would have required an IQ of around 116, for about one-seventh of the population has an IQ this high.

One generation later, the middle class has expanded to around one-quarter of the employed population. This next generation would therefore have required an IQ of 109 or thereabouts to secure a middle-class position based on intelligence.

What is the likelihood of parents with an IQ of at least 116 producing children with an IQ as high as 109? The answer is that about six out of every 10 of their children would score this high. So if class recruitment in these two generations had been based entirely on intelligence, about six out of every 10 middle-class children would have been bright enough to qualify for middle- class entry.

How many working-class children would qualify? Based on the same calculations, working-class parents would have an IQ no higher than 102, and fewer than two out of 10 of their children could be expected to have an IQ as high as the 109 required for middle- class entry.

Remarkably, these figures correspond almost exactly to the findings of the Nuffield team. The survey found that middle-class children were three or four times more likely to succeed than working-class children, but this is exactly what would happen if intelligence alone determined class membership.

Two important conclusions follow from this. First, Britain looks surprisingly like a society divided into classes on the basis of talent. The pattern of social mobility is broadly consistent with what should happen in a perfectly open society with recruitment based solely on intelligence.

The second conclusion is that we do not need to do IQ tests to find evidence supporting the link between social class and intelligence. The close approximation between what would happen under open competition and what does happen in Britain indicates that ability probably does coincide to a large extent with class positions. This lends strong support to Mr Murray's claim of a link between low average intelligence and low class position.

What it does not do is support Mr Murray's additional claim that intelligence is also linked to race. Social classes are recruited through competition and therefore change their membership over time, but racial groups are fixed at birth. The class system can sift each generation by intelligence, but there is no comparable process through which a link between race and intelligence could be sustained.

The writer is Professor of Sociology at the University of Sussex.