In the United States, Blacks score about 15 points lower on IQ tests than Whites do (Roth et al., 2001). There is a debate about why. On the one hand, there are “hereditarians” who argue that this gap is due to a mixture of genetic and environmental causes and that both factors contribute a significant amount to the gap. On the other hand, there are “environmentalists” or “egalitarians” who argue that the gap is 100% caused by the environment.
By looking at Blacks raised in White homes, either due to adoption or because they are biracial, we can test whether or not home variables account for why Blacks differ from Whites in terms of their mean IQ. To the extent that they do, Blacks should have the same IQs when raised in the same homes.
These studies cannot test whether or not non-home environmental factors account for racial IQ gaps. For instance, if racism causes Blacks to score lowly on IQ tests and living in a White home does not free them from facing said racism, then their IQ scores will still be depressed even when living in White homes. Similarly, if pre natal differences explain racial IQ differences then equalizing their home environment after they are born will not eliminate them.
Before we look at these studies, it is important to note that basic behavioral genetic research on intelligence makes it unlikely that home environmental factors can explain racial intelligence differences. Research has shown that, in adulthood, adopted siblings have IQs which are no more similar than average even though they were raised in the same home (Plomin et al.,1997). If having a far more similar than average home experience causes virtually no change in IQ then, outside of rare and extreme home environments, differences in home environments account for little to no variation in IQ in the general population.
It is also worth mentioning research which has been done on how well a child’s IQ is predicted by their HOME inventory score. The HOME inventory measures a wide range of home environmental variables ranging from how parents discipline children, to how a child’s day is organized, to how parents communicate with children, to how safe the physical environment of a child is, to how strongly the children are encouraged to learn, to how common learning material is in the home, to how often the TV is watched, to how often parents talk to children. The correlation between HOME inventory score and IQ is only around 0.2, a weak effect (Herrnstein and Murray, 1994).
It is also worth looking at data from IQ standardization samples. These are highly representative national samples that IQ companies administer their tests to prior to releasing them to the public. The 2003 standardization sample for the WISC, a popular IQ test for children, shows a raw Black-White IQ gap of 11.5 points. This gap was only narrowed to 8.8 points after matching children in terms of parental education, how many parents were present in their home, region, sex, and age (Prifitera et al., 2003, table 1.4).
The relevance of the home environment is also cast into doubt by early intervention studies. In these studies, a group of disadvantaged kids is exposed to an extremely enriching and stimulating cognitive environment. Home life, nutrition, education, etc,. are all improved. These programs produce IQ gains while they are ongoing, but meta-analyses have shown that these gains completely disappear by the time the kids grow up (Protzko, 2015). The fact that massively improving home life doesn’t produce lasting effects on IQ clearly supports the notion that the home environment cannot explain why Blacks and Whites differ in IQ scores in adulthood.
Given all this, we must look at the following adoption studies with a bit of skepticism. Of course, we should keep an open mind if the evidence meets high standards of scientific rigor, but we should be especially careful in making sure they do given that there is substantial background knowledge which makes it unlikely that the home environment can explain much about racial IQ differences.
The Minnesota Study
The most well known study in this literature is often referred to as the Minnesota Trans-racial adoption study. In this study, 265 children had their IQs tested at age seven. At age 17, 196 of these kids returned to have their IQs tested again.
When tested at age seven, the Black and Biracial (Black and White) children in this sample had significantly higher IQs than their groups do on average in America. However, by age 17 their IQ scores had fallen to meet those averages almost exactly.
It is also noteworthy that 12 of the biracial children’s adopting parents falsely believed that they had two Black parents and their IQs did not significantly differ from the other 56 biracial children.
This study clearly supports the contention that differences in home environments cannot explain the black-white IQ gap. It also suggests that whatever the cause is may become stronger with time.
This could be either genes or the non-home environment. Because the heritability of IQ increases with age, the so-called “Wilson effect”, a hereditarian would probably predict that the black-white IQ gap would increase with age, and/or that genes would become a more important factor in it with age (Bouchard 2013). On the other hand, an environmentalist could argue that the effect of the non-home environment accumulates over time. Either interpretation is plausible.
A second important study in this literature is Moore (1986). In this study, Moore compared the mean IQ of 23 Black children adopted into Black homes to that of 23 Black children adopted into White homes. These children were between the ages of 7 and 10. On average, the children adopted into White homes had an IQ score of 117 compared to an average score of 103 for children adopted into Black homes.
Two things stand out about this study. First, its sample size was tiny. Second, even the Blacks raised in Black homes scored higher on IQ tests than Whites typically do in the general population. Thus, the sample was not only small but also unrepresentative. Moore was studying a sample of Blacks in which there was no Black-White IQ gap to begin with. Given this, it isn’t clear that Moore’s study has any scientific worth.
Setting these (frankly fatal) flaws aside, Moore’s data does support an environmentalist view. Being adopted into a White home, rather than a Black home, was associated with a 13.5 point increase in IQ. This is larger than the entire Black-White IQ gap at the age of the children.
Tizard (1974) reported on two other studies in this literature. In the first study, Tizard and colleagues compared the scores of 85 nursery children on measures of cognitive ability in the first years of life and found that the scores of Black, biracial, and White children either did not significantly differ or differed in a way which favored non-Whites. Though these children were not adopted into the same homes, they were being raised in the same nurseries and so had equalized, if unusual, home environments.
In the second study, Tizard et al., looked at 65 children around 4.5 years old who were either institutionalized, adopted, or given back to their biological mother. In every condition, there was no significant difference between Whites and non-Whites on IQ scores.
On its face, this data seems to offer clear evidence in favor of egalitarianism. However, a previous analysis of these studies pointed out that environmentalists would need to suppose an environmental advantage of .60SD and .37SD in favor of Blacks in order to account for them outscoring Whites in this dataset. Given the low value of the heritability of IQ at this age, hereditarians would only need to suppose an environmental advantage of .77SD and .46SD for their model to predict the observed results. The difference between these two suppositions, .17SD and .13SD, is a very small difference considering the small sample sizes present in Tizard’s studies and could easily occur by chance.
Moreover, the Minnesota study should give us pause about interpreting any transracial adoption study on young children which does not include a follow-up in adulthood. After all, if we had just looked at the age 7 data from that study we might have come away thinking that it supported the environmentalist view too.
Overall, Tizard’s study does offer evidence for the environmentalist view, but only weak evidence.
Willerman et al. (1974)
Another study in this literature is Willerman et al. (1974). Willerman looked at the IQ scores of 129 four year old biracial children and found that they varied depending on whether the child had a Black father and White mother or a Black mother and a White father.
Children with Black mothers were found to score 8 points lower on IQ tests than did children of White mothers. Willerman argued that this was due to the higher quality parenting of White mothers.
A few things are worth noting about this study. First, the sample of Blacks is very small (28). Second, no effect was found among females born to married women. This might be attributed to the superior mothering abilities of married women, but WISC standardization sample data shows that controlling for sex and the number of parents in the home leaves most of the Black-White IQ gap unaccounted for. A third thing to note is that the IQ scores of children born to Black women were depressed by a small number of males who were, literally, retarded (IQ<70.01). Fourthly, these biracial children had a mean IQ score of 100, making them an unrepresentative and elite sample of biracial children. Finally, these children were quite young. With all this taken into account, Willerman’s study offers only weak evidence in favor of environmentalism.
The final and earliest study in this literature is Eyferth (1961). Eyferth collected the IQ scores of 181 children (mean age = 10) who were either biracial (black/white) or White. All the children were the result of American military mating with German women during the occupation of Germany following the second world war.
Eyferth found that White children had a mean IQ score of 97.2 compared to 96.5 for mixed children, a gap of only 0.7 IQ points.
The first step in interpreting this study is to understand what we would expect to be true if the hereditarian hypothesis is correct. Because we are dealing with mixed race children, the expected Black-White IQ gap can immediately be cut in half to 7.5 points, giving Blacks a predicted IQ of 92.5. Next, we should account for the fact that the military rejected the bottom 30% of Black applicants compared to only 3% for Whites. This is therefore an elite sample of Blacks. This would lead us to predict a Black-White IQ gap several points lesser, perhaps 5 or 4 points, and a mean Black IQ of something more like 94.5 or 95.5. Then, we have to take into account the fact that the White mothers in question were of low socio-economic status. The Black-White IQ gap is generally much smaller among low SES families, and this alone could lead us to predict a gap several points smaller, maybe 2.5 or 3.5 points, giving us a predicted biracial IQ of something like 96.5 or 97.5. This, we might note, is exactly what the biracial IQ score was.
The problem, then, comes from the Whites, and in particular the White women. Standardization data for the IQ test administered in this study showed that it exhibited no significant difference between samples of men and women. Among the biracial children, this holds true. However, there is an 8 point gap between White males and White females. Because of the previously referenced standardization data, we can therefore conclude that this sample of Whites is unrepresentative and, due to random sampling error, has unusually dumb White females. Among males, the IQ gap between biracial and White children was 4 points. As we’ve seen, this is fully consistent with the hereditarian hypothesis.
Eyferth’s study demonstrably suffered from sampling error and contained an elite sample of Blacks. After taking these flaws into account, Eyferth’s findings are totally consistent with both hereditarianism and environmentalism.
As we’ve seen, different evidence points to different conclusions. However, there is a definite trend to these differences. The studies that support environmentalism tend to be of extremely low quality. The data which supports the hereditarian model tends to have larger and more representative samples, and older participants. Because of this, the totality of evidence on home environments supports the view that it does not play an important role in the Black-White IQ gap.