New Thinking on Children, Poverty & IQ

Jan Richter
November 10, 2003

Research published this month is shedding new light on the old nature/nurture argument. In a dramatic shift from previous findings, this analysis finds that for families at the very bottom of the socio-economic scale, environmental factors have a much greater impact on the variations in children's IQ (Intelligence Quotient) than genes.

For a piece of dense research couched in the language of advanced statistics, Eric Turkheimer's paper in the November, 2003 issue of Psychological Science is making quite a splash. That's because the main finding has such powerful policy implications.

Understanding the Research Findings
But reading Turkheimer's report of his research is not easy—it relies on new statistical methods and computations that are difficult for the layman to understand. So I decided to ask Turkheimer, a professor of psychology at the University of Virginia, for a tutorial. He and I met on a beautiful afternoon in his office in mid-October, and talked for about an hour.

I came away with a better understanding of why these findings are different from those of many previous studies. And why he's confident his findings have serious policy implications for impoverished families.

Until now, "hard science" has been unable to demonstrate that home and community environments have much impact on variations in children's IQ. Instead, genes have been found to account for most of the differences in IQ.

But what Turkheimer and his colleagues found was that although the environmental impact on IQ is very small when you look at families with incomes in the moderate range, the reverse is the case when you look at very low-income families.

ACCORDING TO TURKHEIMER'S CALCULATIONS: At the bottom of the socioeconomic scale genes account for very litle of the differences in IQ. Moving up the scale, genes play more and more of a role. Most of the differences in IQ at the botom of the socioeconomic scale are attibutable to the shared family environment. Moving up the scale, the impact of environment declines. NOTE: The socioeconomic status variable is calculated using information on parental education, occupation and income.


What's So Special
About IQ?

A person's IQ, or Intelligence Quotient, is simply defined as a measure of a person's intelligence as indicated by an intelligence test.

IQ tests generally include vocabulary; verbal analogies; number sequencing,
and some math.

IQ tests have always been controversial but continue to be widely administered and used in research.

One reason is that IQ scores correlate strongly with other measures of scholastic
knowledge, such as SAT tests and grades in school.


How a Different Sample and Different Methods Reveal Different Findings
One thing I wanted to know was why Turkheimer's results were so different from those of earlier studies. On this, he made two points: first, that some earlier studies had found similar results, but the findings hadn't been as clear; second, that the most impoverished families were underrepresented in most earlier studies.

"Our study found the phenomenon more definitively than anyone had ever found it before, using a better sample and better methods, but it wasn't a bolt from the blue," said Turkheimer.

Turkheimer explained that he had always accepted the view that genes play an important role in determining IQ. "What seemed troubling was the other side of that coin: that you couldn't find an effect of family environment on IQ, or at best, at very best, you could find a small effect under ideal circumstances. And that just always seemed wrong," he told me.

So when the opportunity arose to more closely investigate the relationship between IQ and environment among impoverished families, Turkheimer seized it.

The National Collaborative Perinatal Project database:
The project ran from 1959 to 1972.

25 percent of household heads were not educated past 9th grade; 25 percent had incomes below the federal poverty line; 14 percent were unemployed.

Researchers followed 35,000 mothers from early in pregnancy until the children were seven years old.

The children included 320 pairs of twins: 205 fraternal, and 114 identical.

The twins were classified as 43 percent White, 54 percent Black, and 3 percent "other."

Turkheimer's research used IQ scores from tests given at age 7.


Two things made the project possible: the public release of a large, rich database including an unusually large number of children from families of very low socio-economic status; and advances in statistical methods that use computers to carry out huge numbers of computations very quickly.

Twin studies are the basis for much of the research into the relative roles played by genes and the environment in IQ. Research like Turkheimer's relies on the difference between fraternal twins, who are as genetically similar as any siblings, and identical twins, who share exactly the same genes.

For traits that are primarily determined by genes, identical twins will show no variation, but fraternal twins will. For traits that are determined by environment, identical twins and fraternal twins will show similar patterns of variation in the trait. For traits that reflect an interaction between genes and environment, identical twins will show somewhat less variation than fraternal twins. (Confused about twin studies?)

What Turkheimer found was dramatic: for the families in the study at the very bottom of the socioeconomic scale, shared family environment accounted for 60 percent of the variance in IQ; and the contribution of genes was close to zero. (A third variable, non-shared environment, which includes factors such as gender, accounted for the remainder.)

As he looked at families further up the socio-economic ladder, the effects of environment rapidly declined, and genes took on an increasingly important role.

"The more poor families that you leave out (of the sample), the more important the genes get. If you leave out the bottom third of the sample, which is what a typical twin study would do, you would completely misrepresent what is going on," said Turkheimer.

What the New Findings Mean for Parents and Policies
Turkheimer believes his findings are helpful for clarifying both public policy and decisions within individual families.

For middle class families, he notes, "In the range where a lot of people spend their time—you know, 'Should I hang the black and white mobile over my kids' crib?' kind of thing, it probably does not matter."

Hear Turkheimer's view on how this research fits in with findings on Head Start. audio (:18)

For very low-income families, on the other hand, programs like Head Start and others that aim to provide low-income children with better nutrition and health care and richer experiences are "a good thing." But he cautions that "In the long run, what's going to help the most is getting kids out of that state called poverty."

What his research cannot do is identify the specific environmental or family factors that contribute to the variations in IQ among impoverished children. Lead poisoning, poor nutrition, asthma, limited exposure to varied experiences, stress, low-quality schooling —there are plenty of potential contributors. "My honest expectation is that it is not the case that there is some major secret ingredient to bad environments that is accounting for this," said Turkheimer. "What poverty is, is an accumulation of a great many very small deprivations."

"I would be very surprised if anyone ever did research that showed, well, if your kids are raised in poverty, if they get a good hot lunch every day, then their IQ's are going to be OK. Or if somebody reads to them every day, or anything else like that," he added.

Turkheimer hopes that people will understand that his findings do not mean that genes aren't important in IQ. "Too often, I think, people are attracted to a polarized view. Genes ARE playing a role even in the poorest of families, and across a broad range of families they play a very important role, and it would be foolish to close our eyes to that," said Turkheimer.

But understanding that there is a segment of our population for whom family environment is much more important than genes in determining IQ variation suggests that changing environmental factors for these families can change their children's ability to learn.

In this study, as in IQ studies generally, higher socioeconomic status is associated with higher IQ scores.

Hear Turkheimer's view on the politics of IQ. audio (:49)

Turkheimer's research has been drawing praise and vitriol on internet blogs from the left and right ever since the day a description of his research appeared in the Washington Post on September 2, 2003.

Turkheimer says he's confident that his results are sound, and will be replicated by other researchers. "The most important thing to say is that it is an extremely robust finding—I've analyzed the data in a number of ways that are not repeated in the paper—and you get the exact same result no matter how you do it."

If you ask him what he'd like people to understand about his research, he'd say it gives us a different picture than we had before about the relative role of environment and genes on IQ differences among children. It's not that genes are not important, but, for children living in deepest poverty, environment accounts for more of the IQ differences than genetic inheritance.

But if you pin him down "to one thing that would make a big difference" he will say that helping impoverished families improve their circumstances—getting them out of poverty and the conditions that typically accompany poverty—is critically important.


The research referred to in this interview has been published in the November issue of Psychological Science, Vo,. 14, No. 6, under the title Socioeconomic Status Modifies Heritability of IQ in Young Children, Eric Turkheimer, Andreana Haley, Mary Waldron, Brian D'Onofrio, Irving I Gottesman.

The graphs were adapted from that article by Connect for Kids.

Jan Richter is the former advocacy director for Connect for Kids.



Bitly Shortener

Share this SparkAction Link: click here to shorten


What about the role of the number of children per family in the IQ ?

This number is generaly higher in poor families...

Thank you for your forum post.Really looking forward to read more. Fantastic.