Monday, May 11, 2009
The Intelligence of Between-Group Differences
James Watson, co-discoverer of the double helix and winner of the Nobel Prize, may be the most important figure in the history of science alive today. However, a series of statements attributed to him in a 2007 issue of The Sunday Times left the pedestal on which he stands irreversibly tarnished. At the heart of these quotations is the implication that people of African descent are, on average, less intelligent than people of other ancestral origins. Watson did not explicitly deny having made the statements in a written response to the Times article, though he did note that from his “point of view, there is no scientific basis for such a belief.”
Below these alleged comments, the article’s author, Charlotte Hunt-Grubbe, quotes from the following passage in Watson’s recently published autobiography: “A priori,” Watson writes, “there is no firm reason to anticipate that the intellectual capacities of peoples geographically separated in their evolution should prove to have evolved identically. Our wanting to reserve equal powers of reason as some universal heritage of humanity will not be enough to make it so.” Hunt-Grubbe does not draw a distinction between Watson’s alleged verbal statements, which unambiguously imply that people of African descent are intellectually deficient, and this written excerpt. However, a distinction does exist, and it is an important one to emphasize – for on the latter point, Watson is correct: the claim that racial or ethnic groups may differ in mean intelligence cannot be dismissed a priori, and to insist that it can lends credence to its proponents. Instead, it is a debate that must be resolved empirically, if it is to be resolved at all.
I cannot think of any practical benefit that would arise from the knowledge, if it were true, that groups differ in average intelligence as a result of genes. Few would argue that such differences justify differential treatment or unequal allotment of resources. Still, given many persist in believing that a genetic basis for group differences in intelligence exists, it is worth noting why it is probably false.
A flaw in the methodology
The controversy over between-group differences in genetic correlates of intelligence stems from the finding that IQ is highly heritable. One way that this has been determined is by comparing the likeness of monozygotic twins raised apart to that of monozygotic twins raised together. If intelligence is mediated entirely by environmental factors, then twins raised together should score the same or similarly on IQ tests, while twins raised apart should score no more similarly than two people chosen at random. On the other hand, if intelligence were mediated entirely by genes, then we would expect siblings from both sets of twins to score exactly the same on IQ tests. So, by comparing the average likeness of IQ for twins from each group, we can derive an estimate of the heritability of intelligence. Using this methodology, the Minnesota Study of Twins Reared Apart estimates that approximately 70% of the variance in IQ can be attributed to genes. Or, in other words, the similarity in IQ for twins reared apart is almost as high as for twins reared together, thus discounting the influence of environmental factors on general intelligence.
However, while this evidence for the heritability of intelligence may explain IQ differences among individuals, it may not account for differences in mean IQ among racial or ethnic groups (such as those reported in Hernstein and Murray’s The Bell Curve, published in 1994). One reason to doubt that these between-group differences are the result of genetic influences comes from the work of Eric Turkheimer, a psychologist at the University of Virginia.
Turkheimer’s research addresses a major assumption underlying twin methodologies – that the differences in environment for twins raised apart are representative of real-world environmental disparities. While experiments like the Minnesota Study attempt to procure subjects from significantly different socio-economic backgrounds, this difference is often between adequate environments and enriched ones, since impoverished caretakers are generally not invited to adopt and the destitute rarely spend their free time volunteering to benefit psychology.
So, Turkheimer set out to investigate what happens when children raised in impoverished environments are incorporated into the data. As hypothesized, the heritability of IQ did not vary linearly with respect to socioeconomic status. While for subjects being raised in adequate and enriched environments, IQ did appear to be more strongly influenced by genes than by environment, the effect was flipped for subjects being raised in impoverished circumstances. In fact, in 2003, Turkheimer and colleagues wrote that for “impoverished families, 60% of the variance in IQ is accounted for by the shared environment, and the contribution of genes is almost zero.” This finding presents a serious challenge to the claim that between-group differences in intelligence are mediated by genetic factors since the groups being compared often differ sharply in the percentage of people who are impoverished.
The Flynn Effect
Another reason to doubt that between-group differences in average intelligence reflect genetic disparities concerns the discovery of surprisingly large IQ gains within nations across generations – a phenomenon known as the “Flynn Effect”. In a watershed 1987 study, James Flynn presents evidence for massive IQ gains (ranging from 5 to 25 points) over a single generation in 14 countries. Interestingly, Flynn notes, “The magnitude of these score differences matches the size of all major between-group differences in the literature, whether these refer to races, classes or nations.” Since Flynn is comparing groups with (roughly) the same genes, it must be possible for distinct but cotemporaneous groups to differ markedly in IQ without there being a corresponding genetic disparity.
Still, Flynn admits, the size of these effects is perplexing. How can the environment account for such enormous differences in IQ across groups, if it’s responsible for only 30% of the variance across individuals? Poverty alone cannot explain it. So, Flynn and the economist William T. Dickens concluded, intelligence must not be simply the sum of a person’s genetic and environmental inputs. Instead, these components must interact and multiply.
Consider the analogy to basketball skills. As Flynn and Dickens point out, there are genetic factors that might make you better at basketball – being tall, slim, and well-coordinated, to name a few. There are also environmental factors that affect this ability – like growing up with a basketball hoop, owning comfortable sneakers or having a good relationship with your high school coach. But it would be absurd to suggest that these factors are unrelated. After all, your parents might purchase a basketball hoop upon noticing that you are above average in height, you may spring for nicer sneakers if you play a lot of sports, and your coach is almost sure to like you if you’re an asset to the team. In other words, what began as a slight genetic advantage can multiply through interactions with the environment.
This presents a confound for the twin methodology, since, Flynn and Dickens write, “To the extent that environmental quality is matched with genetic endowment, there will be a tendency for identical twins to resemble one another…whether they are raised together or separated at birth.” Thus, certain traits may be under less genetic control than the twin studies suggest.
But also consider how this kind of multiplier effect might occur on the group level. Suppose you live in a society where no one has a basketball hoop in their front yard, comfortable sneakers are hard to come by, and there is no high school basketball team. Now, none of your genetic advantages get multiplied, and everyone is exponentially worse at basketball. If you doubt that these differences would have a significant impact on basketball skills, just consider the state of play in the United States prior to 1945 or the performance of the Angolan national team at the 2008 Olympics. In this way, even small environmental differences can have enormous effects.
The Harlem Miracle
Last week, in a column titled, “The Harlem Miracle,” David Brooks detailed the outstanding success of Promise Academy, a charter middle school in New York City. “In Math,” Brooks reports, the school “eliminated the achievement gap between its black students and the city average for white students.” I do not mention this to demonstrate that it is possible to close the achievement gap between groups by altering the environment. After all, achievement is not the same as general intelligence (which IQ tests purport to measure), and it might be possible to close the achievement gap even if between-group differences were mediated by genes.
Instead, I mention it to emphasize the importance of environmental influences on intelligence more generally. The concept of IQ and the fact of its genetic underpinnings seem to suggest a deep further fact about intelligence – an intimate and indelible relation between your g-factor and who you are, actually. But, as I have begun to illustrate here, this conception is flawed. At most, IQ is a useful predictor of one’s cognitive potential, and this predictor becomes meaningless in the absence of necessary environmental input. As the Flynn-Dickens model suggests, even small environmental deficiencies may have massive effects. Fortunately, the consequences of positive alterations may be just as great.
Posted by Olivia Scheck at 12:25 AM | Permalink