While Harden wants to distance herself from the reactionary politics of the hereditarians, she agrees with them about how social scientists are afraid of genetics research: “like an ostrich sticking its head in the sand, blinding ourselves to genetic data doesn’t make genetic differences go away” (p. 170). The problem, of course, is that it is a myth that ostriches try to avoid danger by sticking their heads in the sand. It is similarly a myth that social scientists do not discuss genetics. Indeed the prestigious Annual Review of Sociology published a lengthy and respectful review of “sociogenomics” last year. Such reviews seem thick on the ground (see here, here, or here). Harden, herself, published such a review in the Annual Review of Psychology and has written about her research in the New York Times, Aeon, and other well-known publications. Most recently, she’s been featured in a long profile in the New Yorker. Harden’s book has already been the subject of significant discussion regarding the strength of its claims and its evidence (see here, here, here, here, here, here, and here). These critiques are sharp and engage directly with the central claims of her book. They give lie to Harden’s claim about a “tacit collusion among many social scientists to ignore genetics…motivated by well-intentioned by ultimately misguided fears…that even considering the possibility of genetic influence implies a biodeterminism or genetic reductionism they would find abhorrent” (p. 186). That behavior genetics, and her specific claims, have been criticized on scientific and policy grounds cannot be evidence that those things have been ignored, but that she and her colleagues have failed to make a persuasive case.
Harden, unlike hereditarians peddling this myth, admits that human genetics has a horrific history of racism and eugenics and that social scientists have a right to be somewhat wary of behavior genetics. She has taken it upon herself to argue that contemporary behavior genetics does not need to be wedded to that past. But her effort does little to assuage those justified fears.
Jam Yesterday and Jam Tomorrow for Behavior Genetics
One reason to be suspicious of Harden’s claim that behavior genetics can inform the social sciences is that such promises have been made before and not panned out. For a long time now, behavior geneticists were telling us that, any day now, molecular genetics, most recently, Gene Wide Association Studies (GWAS), were going to give us major breakthroughs in the study of human behavior and psychological disorders. In 1997, Michael Rutter and Robert Plomin promised that “it is obvious that these are likely to be forthcoming very soon” (p. 212). In 2008, Matt McGue all but guaranteed that “The anticipated yield from genome-wide association studies gives much reason to be optimistic about the future vitality of behavior genetics” (p. 1073). And now, in 2021, Harden writes of the “upcoming avalanche of genetic data” that will inform us of differences among human genetic populations (p. 93). Behavior geneticists can only promise us “any day now” before we begin doubting they will make any such discoveries.
Aaron Panofsky’s comprehensive study of behavior genetics offers a compelling reason why it is always jam-tomorrow-never-jam-today in the field. Behavior genetics has always been a field in disarray. There is no consensus, even among behavior geneticists themselves, about methods, objects of study, or goals of the science. New technological developments, such as GWAS, are adopted eagerly in the firm belief that this will provide the proper tools for showing how genes influence behavior, so far with disapointing results.
It has always been thus for those trying to understand the connections between genes and human behavior. I hope we need not delve into the atrocious record of early-20th century eugenicists. In the 1970s, it was sociobiology claiming it would rework all of the social sciences in its image, a claim that proved to be utterly false. In 1978, Sherwood Washburn, the anthropologist who merged population genetics into anthropology and thus could not possibly be considered to be anit-genetic, pointed to one reason sociobiology was doomed to failure:
This is the fundamental weakness…in the majority of sociobiological thinking as applied to human behavior. Writers are so confident in the power of the theory (selection, adaptation, inclusive fitness) that a minimum effort is made to learn the facts of human behavior and of human history. (p. 36)
These days, It is hard to find any scientist proclaiming sociobiology can inform the social sciences in any significant way. Evolutionary psychologists (EP) threw a fresh coat of paint on it and are now making similar claims about the necessity of EP for all of the social sciences, a claim David Depew and I criticized heavily in our book from both the perspective of the social sciences and of population genetics.
Almost three decades after Washburn’s critique of sociobiology, in 2006, Michael Rutter, leveled a similar charge against behavior genetics:
It is quite striking that behavioral genetics reviews usually totally ignore the findings on environmental influences. It is almost as if research by non-geneticists is irrelevant. The underlying problem is that many behavioral geneticists have been reluctant to pay attention to evidence that does not derive directly from the use of genetic designs. The end product has been a rather one-sided approach to research findings. (p. 11-12)
Harden doesn’t exactly ignore the social sciences, rather she caricatures them with the result that her claim that behavior genetics is necessary for the social sciences collapses.
The Phantom of “All People are Identical”
Harden argues that, without genetic information, social scientists assume humans are as identical as pawns on a chessboard. She offers no evidence for this claim, save a quotation from Bill Clinton (p. 19). I would think, if social scientists embrace the belief that human individuals are completely identical, Harden would be able to muster up at least one social scientist saying it, but she does not.
Nonetheless, the idea that one needs genetics to recognize that people have different capacities and abilities is one of the central claims of the book. In chapter 12, she claims the present state of social science and educational administration are “genome-blind” and believes things such as: “Ignore genetic differences even if it wastes resources and slows down science” (p. 234). “Pretend that all people have an equal likelihood of achieving all social roles or positions after taking into account their environment” (p. 235). There’s more but you get the idea. She never cites anyone who actually holds those positions.
Anyone who attended an American school knows that teachers, and the American educational system, does not believe the positions Harden attributes to them. Our schools are filled with special programs tailored to different kinds of students with different abilities. Evidence for this is offered by sources Harden herself cites. For example, Harden cites this OECD report (p. 161). The report acknowledges behavior genetic research on educational attainment (p. 5), providing evidence that social scientists do not ignore such research. More importantly, the central focus of the report is to deal with learning differences among students (e.g. p. 41). It certainly does not recommend treating all students as if there were identical to each other.
Consider eyeglasses, an example Harden uses in her book (P, 154), and is a standard example to argue against genetic determinism. Most people (estimates are around 60% of the world’s population) use some sort of corrective lenses. Most poor vision owes to genetics. Yet, a simple environmental fix, corrective lenses, makes the genetic differences among us largely irrelevant. Hence, there is no need to fear genetic differences in humans as we can adjust the environment to minimize their impact if we desire to do so.
But there is another lesson to be learned from eyeglasses. Humans were using corrective lenses long before any genetic knowledge about poor vision. Your optometrist does not take a family history or ask for a DNA swab before diagnosing your myopia. Similarly, we can address the needs of poor-performing students without their genetic information.
Another example is Harden’s discussion of her work, specifically this study:
Abstract: Maximizing the flow of students through the science, technology, engineering, and math (STEM) pipeline is important to promoting human capital development and reducing economic inequality. A critical juncture in the STEM pipeline is the highly cumulative sequence of secondary school math courses. Students from disadvantaged schools are less likely to complete advanced math courses. Here, we conduct an analysis of how the math pipeline differs across schools using student polygenic scores, which are DNA-based indicators of propensity to succeed in education. We integrated genetic and official school transcript data from over 3000 European-ancestry students from U.S. high schools. We used polygenic scores as a molecular tracer to understand how the flow of students through the high school math pipeline differs in socioeconomically advantaged versus disadvantaged schools. Students with higher education polygenic scores were tracked to more advanced math already at the beginning of high school and persisted in math for more years. Analyses using genetics as a molecular tracer revealed that the dynamics of the math pipeline differed by school advantage. Compared to disadvantaged schools, advantaged schools buffered students with low polygenic scores from dropping out of math. Across all schools, even students with exceptional polygenic scores (top 2%) were unlikely to take the most advanced math classes, suggesting substantial room for improvement in the development of potential STEM talent. These results link new molecular genetic discoveries to a common target of educational-policy reforms.
Harden argues that her study shows that polygenetic scores can be used to identify successful math education programs. One can just as easily draw the opposite conclusion: these math programs were successful for traditionally low-achieving students even though the teachers and schools did not know anything about the students’ polygenetic scores. In other words, we already know what works and what doesn’t for educating low-achieving students and genetics is completely superfluous.
That genetics is irrelevant to building a just society is obvious from the five principles Harden endorses:
1. Stop wasting time, money, talent, and tools that could be used to improve people’s lives.
2. Use genetic information to improve opportunity, not classify people.
3. Use genetic information for equity, not exclusion.
4. Don’t mistake being lucky for being good.
5. Consider what you would do if you didn’t know who you would be. (p. 233)
1, 4, 5 have nothing to do with genetics at all but are simply principles we all could endorse. The word “genetic” in 2 and 3 adds nothing to those principles and if you deleted it, or, perhaps, replaced it with the word “all,” the principle would be much stronger.
Could we do better in building our educational system and society? Of course. But Harden’s assumption, that we we need more genetic information to do so is simply false.
Harden’s Shifting Standards for Evidence
Despite what her own study seems to prove, Harden denies we know how to improve student success. Ignoring genetics is a “methodological flaw [that] would be excusable if these fields had a track record of rapid progress in the development of successful intervention programs to improve children’s lives. But they don’t” (p. 234). Nothing seems to irritate Harden more than the claim that we know what works in education and she has a entire section entitled, “We Don’t Already Know What to Do.” Examining her argument from this section and then comparing it to her claims about behavior genetics shows her shifting standards of evidence from field to field.
Harden (p. 175) cites this article’s claim that genetic “tools will not reduce, much less eliminate, the health disparities that are produced by the unjust social conditions that are so excruciatingly obvious in our current [Covid] crisis.” There is good reason to think that is correct, but Harden does not engage with that evidence, and, since it is about delivery of healthcare during Covid, it is really not relevant to educational interventions, her supposed topic.
She also cites Ruha Benjamin’s book, which she correctly claims is an extension of John Warner’s argument in this article. She cites these sentences: “I cannot imagine a subject on which we know more about than the environments under which children learn best….We know what to do for students…t’s not mysterious.” Strong claims indeed! But, what is behind those ellipses? Quite a lot, actually. Here are those sentences in their full context:
I cannot imagine a subject on which we know more about than the environments under which children learn best. It has been the subject of study and discussion for well more than a century. Are we suddenly unsure that poverty has a negative effect on educational attainment? We know what to do for students: Feed them, make them safe, respect their autonomy, challenge and support them, give them space to fail, desegregate their schools and neighborhoods, support their teachers…need I go on. It’s not mysterious, and the remedies to our problems aren’t found in our DNA.
Warner is discussing racial equity, respect for students, and especially poverty. Apart from the last sentence, Harden undoubtedly agrees with everything Warner wrote, indeed, her article discusses the differences between “advantaged schools” and “disadvantaged schools.” On the next page, where she lays out why “we don’t already know what to do” she suddenly and, I believe, misleadingly changes the topic entirely.
On the next page (p. 176) Harden points to “paucity of successful [educational] intervention research” But the research she discusses is this survey of randomized controlled trials (RCT) which found that only 12 of 90 such studies found a positive effect of educational interventions. The report also notes, in its first footnote, that those results are not out-of-line with RCTs from other fields of research. More importantly, the interventions the study refers to are not the kind of broad social policies aimed at the eradication of poverty that Warner discussed. In other words, Harden’s discussion of RCTs is irrelevant to Warner’s point.
There is another issue. Harden’s claim that “we don’t know” is based on these criticisms of RCTs, but RCTs are an extremely rigorous method of testing. Harden cannot be possibly arguing that they are the only source of knowledge available to educational researchers which would have to be the case to support her claim that we “don’t know” what works in education.
The issue of scientific rigor has a further dimension worth discussing. Another survey of educational RCTs mentioned by Harden is this study which lays out the rigorous methods of RCTs:
In this article, we focus on RCTs commissioned by the EEF and NCEE. Both organizations commission trials that involve large numbers of participants, often more than a thousand per trial. Moreover, to ensure the quality of their trials, both organizations follow strict methodological guidelines that include comparing the intervention to an active control group, using reliable and valid outcome measures that are not excessively aligned with the intervention, preregistering measures and analyses, commissioning independent evaluators to randomize the participants and analyze the data, and publishing the findings regardless of outcome (EEF, 2017; NCEE, 2017). The EEF and NCEE are not the only funders who commission rigorous large-scale RCTs (e.g., the National Center for Education Research [NCER], another U.S.-based funder, also commissions similar trials). However, they are the only funders we know of who explicitly require all their trials to be published in a standard format that prevents publication bias. This is vital as publication bias can substantially inflate effects in published results.
Does Harden employ equivalent high standards for the studies and methods of behavior genetics? It would seem she does not and it would seem impossible for behavior genetics to meet rigorous standards of evidence demanded by mainstream genetics. For example, a key idea in behavior genetics, is that of heritability: how much of the differences in a trait in a population can be accounted for by genetics? The concept was developed in agricultural genetics, indeed, at my alma mater: Iowa State University. The geneticists at good ol’ ISU never thought human behavior genetics met the standards they expected of themselves (see here, here, and here for examples).
Take twin studies, a centerpiece of those hoping to link genetics to behavior since 1875. Harden cites the University of Minnesota’s (alas! my other alma mater!) Thomas Bouchard as the author of a “classic study” (p. 203) in behavior genetics. But, this study would never meet rigorous standards of evidence Harden demands of other fields. The study’s authors have not shared their raw data and broken traditional standards of providing biographical backgrounds for the twins under study. If Harden would demand equivalent standards of evidence for her own field as she does for others she would would have written a very different book.
Harden, argues that “many researchers are excited by twin studies, adoption, studies, GWAS, and polygenetic indexes” because “Researchers want tools to make genetics recede into the background, to get it out of the way (pp. 187-8). But, this is to get geneticists’ methods completely backwards. Geneticist Richard Lewontin wrote, “The papers on the subject that appear in the journals of behavior genetics would never pass review for publication in the journals of agronomy and animal breeding” (p. 152) because those studying humans cannot control the environment as those in the agricultural sciences can. Every bit of research conducted by Harden and her colleagues has been done in a world beset with social, economic, and racial inequities. If we are interested in finding how genes influence outcomes, we must first enact significant changes to address those inequities to bring genetic differences (f any) fully into view.
Slaying the Racist Demon
Harden goes to great lengths to criticize scientific racists. She acknowledges this is because genetics has been used to justify racist, eugenic practices in the past. Well and good. However, she then claims that “Ignoring genetic differences between people also leaves an interpretive vacuum that political extremists are all to happy to fill” (p. 21). Excuse me? How can there be an “interpretive vacuum” if behavior genetics is the strong, thriving, anti-racist field Harden claims it to be? And the ongoing existence of people like Charles Murray, Jared Taylor, and the rest somehow owes to social scientists not studying genetics? Long-time readers will recognize this is the same argument David Reich offered a few years ago. Like Reich, Harden offers not a scrap of evidence, or even a coherent line of reasoning, to support this allegation. Allow me to offer just a few counter arguments about what disciplines and whose work is providing aid and comfort to scientific racism. If Harden meant her book to fill the “interpretive void” caused by social scientists, she might ask herself why her book, and behavior genetics more generally, have been so well-received by those she opposes.
- Harden opens her book by noting it was unjust to compare behavior genetics to Holocaust denial (p. 10). “Bravo for Prof. Harden. Let her words ring out throughout the social sciences and in the halls of government.” This is a quotation from a review of The Genetic Lottery that appeared on the Unz Review, a Holocaust denial website.
- Harden describes Jared Taylor as an “extremist” (p. 21). But Taylor writes: “First, let us praise Prof. Harden. For an academic, it is risky to say such things as, ‘If people are born with different genes, if the genetic Powerball lands on a different polygenic combination, then they differ not just in their height but also in their wealth.’…all snideness aside, I’m very glad Prof. Harden wrote a book that could prompt more accusations that she is lending comfort to the alt-right.”
- One reason Harden thinks Taylor is an extremist is that he “was a recent recipient of Pioneer Fund money” (p. 15). But so was Thomas Bouchard, the leading behavior geneticist whom Harden cites as an authority. Bouchard’s acceptance of that money lent his credibility, that of the University of Minnesota, and that of behavior genetics to the leading funder of scientific racism in the the post-World War II world.
- Steve Sailer, a regular columnist for the Unz Review, wrote in the equally antisemitic Taki’s Magazine, that, while Harden’s progressive policies are wrong-headed, she has the science of inequality basically correct.
- Harden’s claims about the “tacit collusion” among social scientists to ignore genetics mirrors rather closely scientific racists’ claims about the “equalitarian dogma” (see here, here, here and the paper I published with Andrew Winston). Sometimes, behavior geneticists even cite as an authority neo-Nazi Roger Pearson (a Pioneer Fund grantee), as John Loehlin did in the Handbook of Behavior Genetics.
I could multiply these examples of scientific racists’ admiration for behavior genetics as well as behavior geneticists citing people like Pearson. Rather than training her guns on non-geneticists, I suggest Harden do more within her own discipline to manage how it deals with scientific racism.
Must We Play the Lottery?
Finally, I don’t think Harden understands how lotteries or luck work. She offers a homey illustration of a Rawls-inspired system. She sometimes buys a large cookie to split between her two children. One child splits the cookie, and Harden then hides the pieces in her hands behind her back and the children chooses a hand. Since the cookie splitter does not know which half they will receive the cookie will be evenly split (p. 251-2).
I’m sorry, but this is just Wrong! As a former child myself, and the parent of two (now grown) daughters, I can easily point out the flaw in her system and it is one that, ironically, relies on her assumption that each child is equally prone to risk-taking (children aren’t identical to each other!). What if the cookie splitter is someone willing to take a gamble and divides the cookie unevenly? They have a 50/50 chance of getting the bigger piece and would be willing to take a chance. Or, what if one child knows the parent’s “tell” and can accurately “read” which hand has the bigger piece? And, what the hell, if the guessing child guesses wrong, there is always next time.1
It is this last point that shows why a lottery is such a poor analogy for Harden. She wants to emphasize that one’s genes is simply a matter of “luck” just as winning the lottery is. Of course, “winning” the lottery is just as likely to make you unhappy as it is to make you happy. But, more importantly, if you don’t win the lottery this time, you always have another shot at it. And you don’t have to play the lottery at all (I’ve never played, I think the lottery is a tax on people bad at math). None of these things are true of Harden’s genetic lottery.
As it happens, there is a lottery that is a much better analogy to Harden’s genetic lottery. It is found in Shirley Jackson’s chilling short story, “The Lottery.” This was required reading way back when I was in high school. If you’ve never read it, go read it. Like the genetic lottery, tradition demands every villager must play Jackson’s lottery. Like the genetic lottery, tradition demands a serious sacrifice from the lottery’s losers. Harden thinks we should keep the tradition of respecting the lottery but somehow change the consequences for losing it. I think we should choose not to play at all. Throw the box away completely. Mrs. Hutchinson’s words at the end of the story are correct, “”It isn’t fair, it isn’t right.”
1 The One True Way to split a cookie between two siblings is to have one child divide the cookie and the other child choose which piece to take. This practically guarantees each child gets the same amount of cookie. I’m surprised to find this One True Way is not practiced by everyone.
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.