Hacker News Reader

Comment by lern_too_spel | Hacker News Reader

Without a randomized experiment on babies of different races born to surrogates, it is difficult to say that observed differences in intelligence between races is due to genetics or to prenatal and postnatal environment. 50 years ago, it was assumed that Chinese and Indians were stupid according to the correlational data available at the time, and Chinese Americans whose ancestors arrived in the US in the 19th century still perform poorly in tests of general intelligence, but it is now abundantly clear that this isn't due to a genetic racial disadvantage.

The left would say that environment can and should be fixed.

View on HN

There is an iconic study that is exactly what you're looking for, and may be the reason such studies are also no longer carried out. It is the Minnesota Transracial Adoption Study [1]. They had an array of children broken into transracial, black, and white all be adopted by wealthy white families in Minnesota. Everything was controlled so much as possible including adoption age (less than one year in all cases), family selection, and so on. They even controlled for geography - every single child within the sample living within a 150 mile radius of a fixed point. It then measured the IQ of the children at age 7 and at age 17.

At age 7 the difference in IQ between a adopted white child and adopted black child was 20.1 points. At age 17 this changed to 17.8 points. The difference for a half black child and a white child was 6.1 at age 7, and 8.3 at age 17. The goal of the study was to show the opposite of what it ended up showing. Consequently many individuals, including the authors of the study themselves, have still tried to argue that this is not evidence of a genetic link to IQ. Arguments against it have run the gamut from environmental factors from before the children were adopted (which again was less than 1 years old in all cases) to skin color causing environmental effects. These certainly could be contributing factors, however they would undermine any and all results of the study - which is indicative of a 'nuclear' after-the-fact effort to undermine their own results, rather than a reasoned concern.

That concern is also mitigated by an an unbelievably fortuitous 'accident' that helped create a further control sample. A number of the parents of the interracial children believed that the biological parents of these children were both black. And consequently these children certainly would have believed the same, which would lead one to believe that any environmental affects that would have affected black children more negatively than transracial children would also have affected these children similarly. In reality, they ended up scoring comparably to the interracial children whose parents knew they were interracial.

[1] - https://en.wikipedia.org/wiki/Minnesota_Transracial_Adoption...

I am aware of that study, and it is not what I am looking for. It does not control for prenatal environment or even fully control for postnatal environment prior to adoption. It does not even control for the cause of putting children up for adoption across races.

I know a a few climate change deniers who would be convinced if we could do a study in which we compared the temperature on a planet that was the exact same distance from a star that was the same age and type as our sun that had exactly the same environment as Earth but had not undergone the Industrial Revolution with the increase in CO2.

"Race" derives from, and interacts with, precisely the same institutional social structures which are principle determinants to capability and opportunity, ranging from chemical and biological hygiene factors (including contaminants and pollutants such as lead, asbestos, particulate inhalants, prenatal and early childhood nutrition), to psychological factors (stress and trauma, general socialisation), educational opportunities and environments, and far more.

The fact that the definitions and consequences, no matter how ultimately arbitrary the first should be so deeply intertwined and pervasive, is hardly surprising.

Your argument is not a scientific one.

There is no alternative hypothesis for climate change that fits the data. As the sibling commenter noted, there are many credible alternative hypotheses for the observed differences in general intelligence across races that do fit the data (and moreover, predate GP's study and were the preferred explanation of that study's authors), and from what we have seen with Asians as a prior, I am more inclined to believe them until an experiment can separate the competing hypotheses.

If you want to challenge a hypothesis it's insufficient to simply try to spread uncertainty, you need to create and provide data for an alternative hypothesis. That people, and groups of people, are radically different is supported by everything we know. Genetic factors influence every aspect of our beings, and so to argue that intelligence is somehow immune to this is something that requires evidence. That was the entire point of this study - to show that once environmental factors are carefully controlled for, IQ differences disappear. However, it ended up showing the exact opposite.

When there is a hypothesis that is heavily supported by data, something that might be right, but lacks data, is not a credible alternative. This is the same fallacy that 'climate deniers' make. They think that by arguing that our understanding of climatic systems is limited (which is indeed true) that they can then argue that therefore we can claim alternatives that lack the same scale of evidence (such as that climate change is natural and the CO2 link is a correlation, not causation). That's not how this works...

---

And there's one meta aspect, I want to consider here for a minute. First consider the case of the numerous parents and children who were led to believe that their child was biologically black. These children, and their adoptive parents, would have behaved in a fashion identical to all other parents of biologically black children. Yet, they performed academically and on the IQ exams at the same level as their genetic peers, other transracial children.

This works as a phenomenal control for environmental factors after the period of adoption. And the extremely early period of adoption means if there is indeed an environmental effect happening here then it's either prenatal or in the earliest stages of infancy. Do you understand how enormous the incentive to discover this would be, if people genuinely believed it? You're talking about raising the IQ of millions of people by up to 20 points. That is something that would radically reshape society and would be, by a very wide margin, one of the most important discoveries of the social sciences ever.

Yet oddly enough, nobody seems to be running experiments to try to control for environmental factors to see what exactly it might be that all black parents seem to be doing to their children that's lowering their IQ so much. And this isn't something that would be hard. You'd simply need to offer enough to a sample of families seeking to have children, of differing raises, to have their behavior monitored immediately following conception. What would that take? $10k and free medical checkups? I'm sure that'd be plenty, and there would certainly be grants available to fund it. Yet... nobody seems to be going this route. Think about that for a minute. Actions - or the lack thereof, as always, speak much more loudly than words.

> If you want to challenge a hypothesis it's insufficient to simply try to spread uncertainty, you need to create and provide data for an alternative hypothesis.

You've simply repeated my point. All of the existing data supports GP's sibling's competing hypotheses.

> When there is a hypothesis that is heavily supported by data, something that might be right, but lacks data, is not a credible alternative.

The competing hypotheses are supported by exactly the same data and are further supported by the data from the Asian diaspora and by the twin study data. You must find an experiment that supports one hypothesis and rejects the others, which is what climate science has plenty of.

> That people, and groups of people, are radically different is supported by everything we know. Genetic factors influence every aspect of our beings, and so to argue that intelligence is somehow immune to this is something that requires evidence.

You make a giant leap. Intelligence is partly inherited -- that is true according to twin studies (which don't control for the fact that adopters have to meet a certain bar and therefore can't show the actual effect of post-natal environment and also don't control for prenatal environment). However, claiming that the difference in intelligence between racial groups is due to genetic differences vs. due to environmental differences is not supported by evidence. In 1947, the literacy rate in India was 12%. After seeing the rise in literacy rate over the past 70 years, you would be hard-pressed to claim that this deficiency was genetic.

It seems you can parrot the concepts of scientific reasoning that I explained in the GP post but are simply unable to apply them.

If you're actually interested in seeing what experiments have been done and what the results are, this [1] lengthy but also plain english study provides an excellent overview of some 30 years of research results and what has been discovered. It goes over common misconceptions and arguments in quite extensive detail.

IQ is dominantly heritable. The consensus on heritability (that is the percent of deviation that's explained by genetics alone) has been increasing over time. It used to be believed to be 'only' around ~60% whereas more recent results have shown it to be upwards of 80%. There are lots of strange things about IQ that can lead to poor measurement, like the fact that environmental roles can play some in young children, yet play a negligible role for people as early as adolescence. Anyhow, that paper again provides a great overview of all the issues here. If you're not up for it, Wiki has a page on it [2]. But I'm already cringing there. On the first page, I noticed some editor apparently and literally thinks regression to the mean means two people with high IQ are more likely to produce children of average IQ... Wiki has gotten really quite awful, but I digress!

I'm unsure what you think you're referencing with the "Asian diaspora" or twin study data. Twin study data is conclusive on this topic. I could spam you with sources here, but seriously just searching (google scholar is a useful tool) for iq twin study will lead to a very one-sided picture. If you can provide some papers on the "Asian diaspora" that'd be appreciated, though.

[1] - https://www1.udel.edu/educ/gottfredson/30years/Rushton-Jense...

[2] - https://en.wikipedia.org/wiki/Heritability_of_IQ

The OP did go to the trouble of specifying prenatal and postnatal environment. That study does nothing to normalise prenatal environment, which would be extremely difficult to create a natural experiment for, and it takes an unusually extreme degree of faith in the construction of the study to argue that there is no "reasoned concern" that skin colour might have significantly affected the environments of a cohort of children brought up during the forced busing era.

This is why people do twin studies. Orphaned twins that were separated at birth are as close to a natural instrument as one could ever hope to get.

Unless you can control for the surrounding culture and allow for ~150 years of development of same, even twin studies will be of limited power.

Twin studies have shown that at least 30% of intelligence is determined by post adoption environment. (This is a lower bound because adopters must meet a minimum bar, limiting the differences in adoption environments.) Add that with prenatal environment, and you could have differences in intelligence distributions across races entirely explained, but again, twin studies aren't enough to show that.

Reply on news.ycombinator.com