Wednesday, 28 November 2012

Social class and university entrance

How many children from each social class will enter university?

In Britain today, social class no longer determines our chances in life. A parent’s social class accounts for only 3% of the social class mobility of their children.  The ability of the individual child accounts for 13%. For all we know, the rest of the difference may be due to personality or perhaps even physical attractiveness, but it is not social class.  Without quite realising it, we have achieved considerable social mobility between generations, with far more of that change being due to ability than to social class itself.

One surprising effect of this meritocracy is that social classes still differ in intelligence, simply because the most able have been given a chance to rise into more demanding jobs, and the less able have been left in less prestigious occupations.  Opportunity has allowed people to spread out more, and has brought the best brains to bear on the hardest problems, regardless of their social background.  Separately, there has also been “social class inflation”, with more people doing managerial work, and far fewer in unskilled manual jobs. These manual classes have been stripped of many of their brighter people, who have moved upwards as opportunities opened up.  

If entry to university were based solely on intelligence, how many children from each social class would enter university? Making that calculation depends on some assumptions. First, that people marry partners of roughly the same intelligence. This seems to be true, in that married couples are even more concordant for intelligence than they are for height.  Secondly, that some parental intelligence is passed on to children by genes, and recent heritability estimates of 66% have been established, on samples of 11,000 children (Haworth et al. 2009).

Calculating the estimated intelligence of university applicants according to the social class of their parents is pretty straightforward. Using data analysed by Daniel Nettle on the 1958 generation, the average intelligence score of each social class is multiplied by 66% to get an estimate of the average intelligence of their children. Since the other 34% of intelligence differences are not due to parental intelligence, the class averages will all converge on the population average, a phenomenon known as regression to the mean. The children of professional parents will fall back somewhat towards the population average, though they will remain above it. The children of unskilled manual workers will rise back somewhat towards the average, though they will remain below it. In this way ability is gradually reshuffled each generation, though not totally. At the end of this generational process there will be some differences in the average IQs for each social class. However, these small average differences translate into substantial differences at the upper reaches of the intelligence distribution. This is simply because most people’s abilities pile up in the centre of the intelligence range and there are fewer people at the edges. A small average group difference leads to into big differences in the numbers of individuals at rarified levels of IQ.

As regards university entrance, society can set any cutoff point it likes.  The Table shows what could be expected at various levels of participation. The top 50% was the stated national aspiration, and incidentally corresponds currently to the percentage of the school population who get 5 or more A to C grades at GCSE. The top 40% is close to our current participation level. The top 15% corresponds very roughly to the old universities, and the top 2% to the most intellectually demanding courses at the most highly ranked universities.  The Table shows that social class differences are greatest when the cutoff point is set very high, simply as a consequence of the normal distribution of intelligence.

The point of this exercise is not to say that entry to university should be based on IQ tests. Universities base entry on scholastic tests, particularly those that identify the very brightest candidates. Nor do these calculations lead to setting any particular cutoff point for university entrance as a whole.  That is a social decision.

The real point is to explain that different rates of entry to university according to social class are a direct consequence of a meritocratic society. If people are allowed to rise to the jobs which they merit, (true of Britain from 1958 to 2000 and most probably beyond), then there will be a slight but significant difference in the average intelligence of their children. These differences become quite marked at the outer reaches of the intelligence distribution, leading to actual university entrance figures being legitimately different from simple expectations.  One should not expect every social class to have university entrance rates directly proportional to their numbers in the population, because people are selected into jobs by ability.

Oddly enough, when we hear that proportionately more middle class children are going to university we should reply “So they ought to be, if their parents were correctly selected for their jobs”.

Percentage of each social class who will be admitted if the university takes the top 50, 40, 15 or 2 % of the student population

Student IQ
Top 50%

Top 40%
Top 15%
Russell Group
Top 2%


Nettle, D. (2003) Intelligence and class mobility in the British population. British Journal of Psychology, 94, 551-561.

CMA Haworth, MJ Wright, M Luciano, NG Martin, EJC de Geus, CEM van Beijsterveldt,
M Bartels, D Posthuma, DI Boomsma, OSP Davis, Y Kovas, RP Corley, JC DeFries, JK Hewitt, RK Olson, S-A Rhea, SJ Wadsworth, WG Iacono, M McGue, LA Thompson, SA Hart, SA Petrill, D Lubinski and R Plomin (2009)  The heritability of general cognitive ability increases linearly from childhood to young adulthood. Molecular Psychiatry 1–9.

No comments:

Post a Comment