The environment children grow up in may have a much smaller impact than their genes on how long they stay in education. This is the conclusion of a study on twins, which suggests that almost all previous such research has overestimated the effects of children’s upbringing by failing to take into account two key sources of bias.
 |
| Your upbringing may have little influence on whether you graduate from university Brittany Murray/MediaNews Group/Long Beach Press-Telegram via Getty Images |
It has long been debated whether various aspects of our personalities and abilities are influenced more by our genes or our early environment.
Scientists sometimes investigate the question by comparing different kinds of twins. Identical twins share 100 per cent of their DNA, while non-identical twins are assumed to share 50 per cent, like ordinary siblings.
The extent to which identical twins are more similar in any particular trait than non-identical twins reveals how much variation in that trait is down to genetics. The rest is usually assumed to be due to the environment.
Such studies have shown that many aspects of our abilities and personalities have little environmental contribution. But a measure called educational attainment – defined as how many years people spend in full-time education – has often been calculated to have a relatively large environmental component, of around 35 per cent.
These studies usually overlook two factors, however, leading to falsely high estimates of environmental effects, says Damien Morris at King’s College London. The biggest overlooked factor is that people tend to have children with someone with a similar educational background as themselves – for instance, someone with a university degree is more likely to partner with someone else with a degree, an idea known as assortative mating.
“It’s a source of substantial bias,” says Morris. “Parents match very closely on this characteristic, but it has not been taken into account in the traditional twin method.”
If assortative mating with respect to any trait does happen, then any full siblings, including non-identical twins, would share more than 50 per cent of the genes that influence that trait, because of the genetic overlap between their parents. This means that the usual twin study methods for calculating the genetic contribution would give a falsely low figure and therefore the environmental contribution would appear falsely high.
Morris and Tobias Wolfram at Bielefeld University in Germany investigated whether assortative mating for education really does happen by using a recent German twin study in which data had been collected, not just for the twins, but also for their parents and another sibling of the twins.
The pair examined data on years of education for nearly 1000 families including identical or non-identical twins who were born in the 1990s.
Analysing this data using the traditional twin study design, environmental differences between families were estimated to account for 43 per cent of the variation in years of schooling.
Then Morris and Wolfram analysed the same data using a different method that also included the parents’ years of education. Sure enough, there was more similarity between the educational years of each mother and father pair than would be expected by chance. Taking this into account, the environmental contribution was estimated to be just 26 per cent.
The pair did a further kind of analysis too, to investigate the impact of a second factor usually overlooked in twin studies. This is the fact that the early environments of twins are more similar than those of non-twin siblings because they are born at the same time. As well as having shared a uterus, this means that they are more likely to have grown up in the same financial circumstances, gone to the same school and so on.
The new study measured how much this effect normally skews the results of twin studies, by comparing the number of years that pairs of non-identical twins and pairs of one twin and their non-twin sibling spent in education. When this was taken into account as well, the estimated environmental influence fell to 10 per cent.
The revised low estimate of the influence of upbringing has surprised some other researchers in the field. “This is a very intriguing finding,” says Jeremy Freese at Stanford University in California. “The technical prowess and the thoughtfulness in this paper is pretty evident. But it is unimaginable to me, thinking about how college decisions are made, to think that everything that goes on with parents is only a bit player in the story.”
Other kinds of studies have given higher estimates for the environmental contribution to years of schooling, says Alexander Young at the University of California, Los Angeles. “Ultimately, we need to triangulate different sources of evidence,” he says.
Journal reference:
0 Comments