The genetic lottery: Why DNA matters for social equality

[Gloss: each person, each member of our study population is assigned to a genotype and to a life outcome for our study life topic. A genotype enumerates the permitted values for each of the relevant genes, these last being some probably small proportion of the total. Genotypes are grouped into clusters, defined in some way by the values of the genotype’s constituent genes. We then compute risk factors for each cluster, for each outcome. Risk factors, like probabilities are between zero and one. For this to work, study populations need to be large. More detail follows below].

Reference 1 is a reasonably hostile review of the book at reference 2. The authors of the review article are professors at Stanford, one of science, in particular computational genomics, the other of the history of science, while the author of the book is a professor of psychology at University of Texas at Austin. A lady, according to Wikipedia, with a conservative, Pentecostal background on which connection, I suppose I should declare a prejudice against Pentecostalists. As far as I am concerned, a very rum bunch.

A book which is possibly controversial in the same way as that at reference 4, a book which argued firstly that intelligence, as measured by intelligence tests was an important predictor of life outcomes and secondly that there was a link between this intelligence and race. My recollection being that the link was quite weak: the variation within race was much larger than the variation between races. Again, as far as I can recall, my position was that, given that the link, such as it was, was weak, it might have been better, given all the history, particularly in the US, not to have stirred things up in this way.

While this book argues that it is possible to compute functions which get from any particular set of genes, the genes of some particular person, to a measure of risk that this person will have this or that bad life outcome. That it might be appropriate to give special attention to young people with high risks in this sense.

The latest of a number of attempts over the past two hundred years or so to reduce complicated problems to routine, more or less mechanical measurements of people. With one of the first such attempts being that of the phrenologists with their callipers.

Part of the motivation for this attempt is that while if you take a random couple of people, nearly all of their genes are going to be identical, there are still a lot of genes which exhibit variation, which come in two or more reasonably common varieties, sometimes called alleles. While looking at variation of just one gene at a time has not been terribly successful, maybe we will do better if we can look at variation of lots of genes at the same time. This sort of variation might well explain stuff.

In slightly more detail, we call the genes – not necessarily all the genes, but the certainly the same genes for each person – for a particular person their genotype. You then get the genotype of a large number of people, the study population, for whom you know something about life outcomes, nothing like as expensive these days as it would have been a few years ago. So if the outcome of interest is becoming an unmarried mother, you might have just two outcomes: having one or more children before marriage or not. Supposing here that other children born out of wedlock later on don’t count. You then look at the map from genotype to outcome. You try to cluster the genotypes in some sensible way, using relevant alleles of relevant genes in some way, and come up with a risk factor for each cluster for each outcome – perhaps arranging things so that the risk factors for a cluster add up to one, in the way of probabilities. You might, going even further, speculate about the distribution of outcomes, for example, distribution of outcomes for the set of people for whom the risk factor for unmarried, usually juvenile, maternity is greater than 0.66. All of which might go under the fashionable name of a Genome Wide Association Study or GWAS. In any event, given some new person, you can get their genotype and out pops the risk factor for any particular outcome. Straightforward sausage machine sort of business. And if you were a policy person, you might think about policy options for those people.

The authors of the article find this obnoxious, I think in large part because it facilitates the more or less mechanical identification of inferior groups of people for special treatment, a proceeding  which is open to abuse and which is apt to be traumatic for those so identified. I associate to the 11-plus examination of old, here in the UK, that is to say the consigning of four fifths of our children to bog-standard secondary modern schools at the age of 11. 

But also because they argue that real world problems and disorders are much more the product of environmental factors than genetic factors. Intelligence is much more a function, for example, of a satisfactory home environment than of genes. And I dare say this is even more true of youth offending and single parenthood. So addressing the environmental factors is much more likely to produce the goods than dubious appeals to genome science

I don’t think that I am going to dig any deeper.

PS 1: along the way we are told of the Bucharest Early Intervention Project, described at reference 3. For various reasons, it seems that when the communist regime in Romania collapsed in 1989, there were a large number of unwanted children in care in institutions, this at a time when most countries in the west had replaced, to the extent that this was possible, institutional care for children by foster care. This project took a sample of 136 of these children and placed a random half in foster care. There was a control group of 72 children who were with their natural parents. Researchers were not terribly surprised to find that, in the round, children with their natural parents did best, then children with foster carers, then children in institutions. A project perhaps more interesting for the moral issues involved than for the results.

PS 2: the article is topped and tailed with the story about the frog heated up in a pan of water so slowly that it does not bother to jump out – and so expires. This by way of comparison with the book which, the article tells us, starts with various uncontroversial not to say banal propositions, cruises through a gentle chain of argument to arrive at very controversial conclusions. We are invited to jump out, which is, we are told, what real frogs actually do.

PS 3: there was also talk of bandwagons in connection with GWASs and I wondered about the origin of the word. Bing not helpful. OED not helpful, despite spending some columns to list all the possible uses of the word ‘band’, together with various compounds. Not this one. But Webster did list the word, without a hyphen, and talks of early use as the fancy wagon used for the band in circus parades. By extension a party, faction or tendency which catches the attention and adherence of lots of people. Perhaps also their parades, introductory or victorious. With just a hint of the possibility of personal gain therein.

References

Reference 1: Why biology is not destiny – N. Feldman, Jessica Riskin, NYRB – 2022. 21st April number.

Reference 2: The genetic lottery: Why DNA matters for social equality – Kathryn Page Harden – 2021.

Reference 3: Case Study in Ethics of Research: The Bucharest Early Intervention Project – Charles H. Zeanah, Nathan A. Fox, Charles A. Nelson – 2012. Open access at the address given below.

Reference 4: The Bell Curve: Intelligence and Class Structure in American Life - Richard J. Herrnstein, Charles Murray - 1994.