Post by Admin on Feb 10, 2021 19:24:40 GMT
Jay Joseph: Left-Wing Behavioral Genetics? A Closer Look at the Genetic Evidence in “The Cult of Smart”
Behavioral genetic “discoveries” are a mirage, a house of cards that ignores contradictory evidence from countless real-world examples and research findings from other fields, and collapses under serious critical analysis.
Left-Wing Behavioral Genetics? A Closer Look at the Genetic Evidence in “The Cult of Smart”
www.madinamerica.com/2021/02/left-wing-behavioral-genetics/
In his 2020 book The Cult of Smart: How Our Broken Education System Perpetuates Social Injustice, writer/academic Fredrik deBoer attempted to reconcile his own and others’ socialist/left political perspectives with what he saw as discoveries produced by twin and adoption “kinship” studies coming from the field of human behavioral genetics, plus recent claims of associations between some genetic variants and IQ.1 A major theme of the book was that “academic potential varies between individuals, and cannot be dramatically improved.” This led deBoer to propose changes in the U.S. education system:
“Since cognitive talent varies from person to person, our education system can never create equal opportunity for all. Instead, it teaches our children that hierarchy and competition are natural, and that human value should be based on intelligence. These ideas are counter to everything that the left believes, but until they acknowledge the existence of individual cognitive differences, progressives remain complicit in keeping the status quo in place…We [should] embrace a new goal for education: equality of outcomes. We must create a world that has a place for everyone, not just the academically talented.”2
deBoer’s recommendations flowed from his belief that behavioral genetic research has shown that “human behavioral traits, such as IQ, are profoundly shaped by genetic parentage, and this genetic influence plays a larger role in determining human outcomes than the family and home environment” (The Cult of Smart, p. 23). Simply put, he wrote that “children are shaped predominantly by nature, not nurture” (p. 128).
deBoer endorsed the behavioral genetic position that environmental influences can be divided into “shared and unshared” components, a position that “emerged most directly from twin studies” (p. 126). Also based on behavioral genetic claims, he argued that “shared” influences such as “parenting, the family, and home life simply don’t seem to matter much on a number of definable behavioral traits, including intelligence” (p. 128). Rather than reviewing the book as a whole, I will focus on how deBoer discussed the behavioral genetic research methods and studies he put forward as supporting a “predominant” role for genetic (nature) influences on human behavioral differences.
I am also on the left politically, which means that my analysis of deBoer’s claims comes from within that broad camp. Since 1998, with important help from others, I have developed a critique of behavioral genetic and psychiatric genetic research and theories in four books, and in many peer-reviewed and online articles. A significant portion of my focus since 1998 has been on major problems in behavioral twin and adoption research, and how false claims based on these research methods have been used to lend “scientific” support to policies maintaining the oppressive and economically very unequal political status quo throughout the world.
Like Harvard psychologist/author Steven Pinker before him, deBoer repeatedly invoked the “blank state” straw man caricature of environmentalists (in the causes of behavioral differences sense), who allegedly believe that “we… are all equally equipped at birth to succeed in academics and all that comes with them” (p. 28). I say caricature because few critics or environmentalists make such a claim. What they do argue is that while there may be inborn differences among individuals for certain abilities, it is crucial to understand that family, social, cultural, educational, geographical, and political environments together play a powerful role in shaping human behavior. They conclude that social, political, and academic attention should be focused on improving these environments, not on people’s brains and genes. Environmentalists point to the many characteristics and potentials we all share, and advocate for social policies or even radical social change to create societies providing healthy environments for everyone to grow up in and thrive.
As Current Affairs Editor-in-Chief Nathan Robinson pointed out in his 2020 review of The Cult of Smart, the book reads like a left-wing version of Richard Herrnstein and Charles Murray’s controversial 1994 book The Bell Curve. Unlike those authors, however, deBoer “unequivocally and explicitly” rejected the “position that some races are inherently more intelligent than others” (p. 122). Although in this article I focus on deBoer’s claims about genetics, I note Robinson’s negative assessment of the book:
“The central argument of the book is not just wrong, but wrong in the strongest possible sense of that term. It is based on fallacious reasoning. It is a mistake. An error. The whole argument falls apart completely the moment you touch it. Now, you can usually make an argument for almost anything, because so many arguments are based on subjective value judgments that are hard to prove or disprove. But unless you abandon the basic principles of logical inference, you can’t successfully make the argument of this book.”
The Cult of Smart was positively reviewed in the conservative National Review and in The American Conservative. Noting the irony of a leftist thinker attempting to take down environmentalist positions on behavior and cognitive ability, the conservative author of the National Review piece wrote that for him, reading the book brings forth the pleasure of “watching a pitcher on a [baseball] team you don’t cheer for throw a no-hitter against your crosstown rivals.”
Twin Studies
Although deBoer’s arguments were based on supposed findings from the behavioral genetics field, he admitted that he had no “expertise” in this area, and that he relied “on the work of scholars in that field to understand the dynamics I’m about to discuss” (p. 123). As he wrote in a post-publication blog post dated 8/10/2020, “I am a (well read) total amateur attempting to engage with scientific concepts.” He relied mainly on influential books popularizing behavioral genetic positions, such as Pinker’s The Blank Slate and Judith Harris’s The Nurture Assumption, and on articles such as behavioral geneticist Eric Turkheimer’s 2000 “Three Laws of Behavior Genetics and What They Mean.” deBoer did not cite a single original-data twin, adoption, or molecular genetic study in support of genetics.
Twins Reared Together “Twin Method” Studies
Studies of reared-together twins, deBoer recognized, have an “obvious confound” because “twins are usually reared together, in the same homes, by the same parent or parents.” For “cognitive traits like intelligence,” he reasoned, “it would seem that there’s much more opportunity for parental influence. In order to study purely genetic effects, we’d have to study twins reared by different people and in different homes, and that’s precisely what researchers do” (p. 124).
In the twin research area, therefore, deBoer didn’t seem to place much faith in studies using “confounded parentally influenced” reared-together MZ twin pairs, believing that studies using reared-apart twins provide a much better way of identifying “purely genetic effects”—a point I will soon return to. (MZ twins are also known as monozygotic or identical; DZ twins are also known as dizygotic or fraternal). Nevertheless, deBoer cited a 2015 twin study meta-analysis in favor of 40% to 50% heritability “for most behavioral traits” (p. 127). This meta-analysis was based on pooled studies using reared-together MZ versus DZ “twin method” comparisons, in other words, what deBoer saw elsewhere as “confounded” twin data.
As critics have argued since the 1930s, genetic interpretations of MZ-DZ comparisons are based on the false assumption that both types of twin pairs grow up experiencing roughly “equal environments.” The assumption is false because in addition to MZs being treated more similarly and spending more time together, levels of identity confusion and attachment are much higher among MZ pairs than among DZ pairs (evidence here). Although the academic consensus says otherwise, reared-together MZ-DZ twin method comparisons do not prove anything about genetic influences on behavior. (I explain why they don’t here.)
The “Heritability Fallacy”
Critics also argue that the heritability concept itself, which deBoer endorsed throughout the book, is both misunderstood and highly misleading, and does not indicate “how much” genes influence behaviors and abilities, or how changeable they are. As David Moore and David Shenk wrote in “The Heritability Fallacy,” the “term ‘heritability,’ as it is used today in human behavioral genetics, is one of the most misleading in the history of science.” Often based on twin study data, researchers calculate “heritability estimates” ranging from 0% to 100%, and these estimates are frequently found in textbooks and other authoritative works. The heritability concept was originally developed to predict the results of selective breeding programs of economically desirable traits in farm animals, such as milk production in cows. Since the 1950s it has been extended, controversially, as a measure of the “strength” or “magnitude” of genetic influences on various psychiatric disorders and human behavioral characteristics, including especially IQ.3
I agree with biologist Steven Rose, who concluded that the “heritability measure…except in the very specific context for which it was originally devised (agricultural breeding experiments) [is] rarely applicable, widely misunderstood and in most cases meaningless.”4 (More on problems with the heritability concept here.)
Twins Reared Apart (TRA) studies
We have seen that, due to “obvious confounds” in reared-together twin studies, deBoer based much of his argument in favor of high IQ heritability on studies of “twins reared by different people and in different homes.” These are known as “twins reared apart,” or “TRA” studies. Only six such studies have been performed, with the first appearing in 1937. Again, deBoer relied on influential behavioral genetic researchers and popularizers, not on his own reading of these original TRA study publications. He appeared to know little about these studies, whose massive flaws and biases have been highlighted by critics for decades (summarized here).
The “Minnesota Study of Twins Reared Apart.” The only TRA study deBoer mentioned was Thomas J. Bouchard, Jr. and colleagues’ widely cited Minnesota Study of Twins Reared Apart, or “MISTRA.” He did not cite any original MISTRA study publications or review articles, and none of the behavioral genetic authors he relied on described in any detail how the study was performed.
deBoer implied that Bouchard and his MISTRA colleagues obtained their reared-apart twins from the “Minnesota Twin Registry” (MTR), “which seeks to gather data from all of the twins born in that state over the course of several decades” (p. 125). Referring to me as a “dogged critic of twin studies,” deBoer claimed that I “argued that the [MTR] data set is flawed because many of the included twins were in fact separated only after years of cohabitation, confounding [the] analysis” (p. 133). He cited my book The Trouble with Twin Studies, yet mentioned only one small aspect of my larger critique of TRA research.5
To set the record straight, I did not write that the Minnesota Twin Registry “data set is flawed,” or anything else about the MTR in the TRA study context. This is because the MISTRA did not obtain its reared-apart twins from the MTR, which was based on reared-together pairs. Instead, Bouchard had to recruit twins, most of whom volunteered to participate after hearing about the study in the media. In her 2012 book about the study, Born Together—Reared Apart: The Landmark Minnesota Twin Study, MISTRA researcher Nancy Segal described the study’s twin sample as “a collection of cases”:
“Lacking a national registry, the MISTRA studied a collection of cases because our reared-apart twins surfaced in many ways and at any time….The majority of pairs were identified by self-referral, but most of the twins learned about the study through media reports.”6
Of the total MISTRA reared-apart twin sample, only 45% of the pairs were recruited from the United States, with the rest coming from overseas. Around one-third came from England.7
deBoer wrote that “many” TRA studies draw from data sets such as the MTR, with the “drawback” being that “any problems with that data would be replicated across different studies” (p. 125). That may be true for the numerous studies of reared-together twins, but here deBoer confused reared-together and reared-apart twin samples. Reared-apart twins are extremely rare, are exceedingly difficult to locate and to study, and there has never existed a registry of such pairs. deBoer’s unfamiliarity with twin research came through at every step.
Most twin pairs were only partially reared apart. In TRA studies, deBoer wrote, “it’s imperative that twins reared apart are truly reared apart. If the twins are raised partially together, it would jeopardize the inferences made based on the data set” (p. 133). Yet he provided no evidence that twins actually were “truly reared apart.” In fact, most studied reared-apart MZ twin pairs, also known as “MZA” pairs, were only partially reared-apart (evidence here).
In the larger context of kinship studies, deBoer wrote that “identical twins are closely aligned in behavioral and cognitive metrics even when raised in profoundly dissimilar environments” (p. 126). Here he seemed to be referring to single-case anecdotal reports, such as the “Jim Twins” and the now more-famous “Three Identical Strangers” triplets, whose cherry-picked “spooky similarities” have been reported in the media, and have been misleading the public, for decades. It is difficult to determine how similar or dissimilar the 81 MISTRA MZA pairs’ environments actually were, because Bouchard and colleagues denied critics and independent observers access to their raw data.8
Does the MZA correlation “directly estimate heritability”? The standard behavioral genetic position is that the MZA correlation “directly estimates heritability.” For example, a .50 MZA IQ-score correlation = 50% IQ heritability. Following this script, deBoer wrote,
“If identical twins raised apart are exactly like each other on a given cognitive trait, it would imply that the trait is purely genetic; if identical twins raised apart are entirely unalike each other on a given cognitive trait, it would imply that the trait is purely environmental. And any values in between could be used to assess what proportion of the (population-level!) trait is genetic or environmental.” (p. 124)
According to this view a significantly above-zero MZA correlation, for example IQ test scores, is like a dimmer switch of genetic importance—the brighter the light (the higher the correlation), the stronger the genetic influence. deBoer and the behavioral geneticists are mistaken, however, because in addition to problems with the heritability concept itself, even perfectly separated MZA pairs—simply because they are born at the same time—grow up sharing many non-familial “cohort” influences in common (over 30 listed here). MZAs also share a prenatal environment, a very similar physical appearance, and are the same sex. All these factors will contribute to MZA behavioral similarity for non-genetic reasons, and thereby confound genetic interpretations of MZA behavioral correlations.
Bouchard, Segal, and their MISTRA colleagues got around these problems by keeping their raw data secret, and by counting the environment as genetic (yes, you read that correctly). As seen in their widely cited 1990 Science article, their conclusion that IQ is “strongly affected by genetic factors” was based on their decision to count environmental influences on MZAs’ behavioral resemblance as genetic influences—thereby creating a “heads I win, tails you lose” argument they could not lose, because they counted both the “nature” and “nurture” sides of the behavioral coin as nature.9
As I showed in The Trouble with Twin Studies, the Minnesota Study of Twins Reared Apart was massively flawed, heavily genetically biased, and based on numerous false or questionable assumptions. The researchers had to suppress their reared-apart DZ control group IQ correlations to reach the desired genetic conclusions. Had they decided to publish these correlations in their main IQ paper (the 1990 Science article), they would have been forced to conclude that their study found no evidence that genes influence IQ scores and “general intelligence” (explained here).
Adoption Studies
The other main kinship research method DeBoer cited was the study of non-twin adoptees and their biological and adoptive families. Researchers conducting these studies typically conclude that adopted children correlate higher with their biological as opposed to their adoptive relatives, as genetic theories predict. Behavioral geneticists assume that in these adoption studies, birthparents share genes but not environments with their adopted-away offspring. Based on this assumption, in the IQ studies they estimate heritability by doubling the birthparent/adopted-away offspring IQ-score correlation.10 However, even children adopted away at birth share several environmental factors in common with their birthmothers. This always includes the prenatal environment, usually includes skin color and “race” (frequently leading to oppression or privilege), and often includes similar physical appearance, social class, culture, religion, and so on.
Moreover, couples wishing to adopt children are carefully screened by adoption agencies for emotional and financial stability, and children are often intentionally placed into adoptive homes matched on some of the characteristics—including the assumed genetic status—of their biological family background. “Fitting the home to the child,” a psychology researcher acknowledged in 1975, “has been the standard practice in most adoption agencies, and this selective placement can confound genetic endowment with environmental influence to invalidate the basic logic of an adoptive study.”11 This is known as “selective placement bias.” Other biases in adoption research include attachment rupture and its impact on a child’s developing brain, late separation from the birthparent, late placement after separation, and the environmental “range restriction” found in the adoptive parents group.12 interpretations and conclusions” found in IQ adoption study publications.] For more on these and other IQ adoption study problem areas, see the 2006 analysis by Ken Richardson and Sarah Norgate.
Genetically oriented writers usually focus on the correlational adoption studies, whereas environmentally oriented commentators often highlight studies showing large IQ gains made by children born to poor parents, who are then adopted by professional or upper-class parents. In one such study, the biological children of poor or working-class parents showed a substantial 14-point IQ-score increase when they were raised in the adoptive homes of families in the upper ranges of the socioeconomic scale. The behavioral genetics popularizers that deBoer relied on tend to downplay or ignore such findings.
A more fitting name for this area of research is the study of abandoned and rejected children. Abandonment and rejection are among the most psychologically harmful events a child can endure, and for this reason alone it is questionable whether the results of “adoption” studies apply to people who were not abandoned or rejected as children.
Molecular Genetic Research
For decades, molecular genetic researchers have attempted, and for the most part have failed, to identify specific genes shown to cause (not just be “associated with”) differences in IQ, the major psychiatric disorders, and behavioral differences in general. Citing only a 2017 New York Times article by science writer Carl Zimmer, deBoer claimed that researchers using Genome-Wide Association Studies (GWAS) have discovered seemingly “clear” associations between intelligence and some genetic variants (p. 136). The Times and other corporate media outlets regularly publish sensationalized false-alarm gene discovery claims for intelligence, psychiatric disorders, violent crime, loneliness, gang membership and other areas of behavior. This helps lead their readers to the conclusion that human behavioral differences are largely “in the genes.”
Many people on the left understand that corporate media outlets at times provide misleading and biased coverage in politically related areas of science, just as they do in relation to war, foreign policy, and political issues in general. This doesn’t mean that leftists and progressives shouldn’t cite or discuss politically relevant science articles appearing in the corporate media, but they should approach such articles with a critical eye when their authors’ claims align with those of the outlet’s corporate ownership.
deBoer seemed unaware of the fact that countless false-alarm behavioral gene-association publications have appeared since the 1960s. Decades of “genes for behavior” false alarms indicate that there are major methodological flaws and biases underlying these studies, and lend support to the critics’ contention that molecular genetic researchers have been misled by twin studies, adoption studies, and heritability estimates. This is also true for more recent studies based on “polygenic risk scores” (PRS).
In his 2014 book Misbehaving Science, sociologist Aaron Panofsky described the behavioral genetic gene-discovery failure “coping strategy” of “technological optimism.” By this he meant the “optimism that the next level of technology will overcome past disappointments.”13 Most likely, current excitement about PRS studies in behavioral genetics will fade when this method ends up becoming yet another “next level” method that failed.
Given the dreadful track record, by default we should assume that current and future behavioral gene-association claims, and polygenic risk scores, are false-positive and/or non-causal findings until proven otherwise.
Conclusions
In The Cult of Smart, Fredrik deBoer attempted the difficult task of reconciling his socialist beliefs with what he saw as “the fact that our genetic heritage deeply influences our behavioral selves, including our academic selves…” (p. 23). His admirable goal was to recommend ways to improve the “broken” U.S. education system, but he knew little about the twin, adoption, and molecular genetic studies upon which his recommendations were based. It’s fine not to know much about these things, but it’s not fine to publish a book whose arguments are based on research one knows little about, and which may even lend support to “strange bedfellow” political forces one opposes.
A week after the book came out, deBoer published a blog post with the puzzling title, “If Genes Contribute Nothing, My Conclusions Are All the Same.” In it, deBoer wrote,
“some serious people with serious evidence believe that our genetic endowment shapes our behavioral outcomes like how well we perform in school. Whether they’re right or not, I can’t be the one to say.” (italics added)
A major theme of The Cult of Smart, however, was not that “serious people’s” claims that “genetic endowment shapes our behavioral outcomes” might be right, but that they are right.
In the book, deBoer discussed the characteristics of “ardent hereditarians” and the “well-meaning” people who oppose them:
“On the one hand, you have the ardent hereditarians who reduce all of human life to the outcome of genes, and who are often incorrigibly racist. On the other hand, you have well-meaning but misguided people who dismiss the importance of genetics out of hand, who associate any discussion of the heritability of behavioral traits with the worst elements of eugenics and colonialism, and who make grandiose claims about a scientific literature they have never read. This leaves us unable to argue honestly and persuasively when we need to most.” (p. 139)
As we have seen, deBoer himself made claims about a scientific literature that for the most part he has never read. But I do agree with him that anti-hereditarians should not reflexively answer hereditarian claims with charges of “eugenics and colonialism,” even though it is important to know and write about this history. And it isn’t even necessary since, as seen in the 1996 book Inequality by Design, and in Ken Richardson’s 2017 Genes, Brains, and Human Potential and elsewhere, hereditarian and racial-differences arguments fall apart upon close examination.
deBoer concluded that his fellow progressives and socialists should know that “genes matter,” and advised them not to leave the discussion to “pseudoscientific racists” and “the worst elements”:
“Genes matter. They matter for us as individuals and they matter for us as a society. And they matter whether progressive people believe them to matter or not. What the political left must understand is that sticking our heads in the sand and ignoring the rapidly developing world of human genetic research does no one any favors. To leave the field of argument to the worst elements, the pseudoscientific racists and gender essentialists, is to abdicate our sacred responsibility to defend our values.” (p. 237)
If behavioral genetic “discoveries” were real, then deBoer’s advice would make some sense. But they are a mirage, a house of cards that ignores contradictory evidence from countless real-world examples and research findings from other fields, that collapses under serious critical analysis.14 In the process of exposing between-group racist pseudoscience, then, progressives and socialists must understand that claims relating to the “high heritability” of within-group “individual differences” in IQ and other behavioral areas are also based on flawed research and concepts, including the IQ and “general intelligence” concepts themselves. Learning more about this through the works of critics will help arm them in the debates and struggles ahead.15
Behavioral genetic “discoveries” are a mirage, a house of cards that ignores contradictory evidence from countless real-world examples and research findings from other fields, and collapses under serious critical analysis.
Left-Wing Behavioral Genetics? A Closer Look at the Genetic Evidence in “The Cult of Smart”
www.madinamerica.com/2021/02/left-wing-behavioral-genetics/
In his 2020 book The Cult of Smart: How Our Broken Education System Perpetuates Social Injustice, writer/academic Fredrik deBoer attempted to reconcile his own and others’ socialist/left political perspectives with what he saw as discoveries produced by twin and adoption “kinship” studies coming from the field of human behavioral genetics, plus recent claims of associations between some genetic variants and IQ.1 A major theme of the book was that “academic potential varies between individuals, and cannot be dramatically improved.” This led deBoer to propose changes in the U.S. education system:
“Since cognitive talent varies from person to person, our education system can never create equal opportunity for all. Instead, it teaches our children that hierarchy and competition are natural, and that human value should be based on intelligence. These ideas are counter to everything that the left believes, but until they acknowledge the existence of individual cognitive differences, progressives remain complicit in keeping the status quo in place…We [should] embrace a new goal for education: equality of outcomes. We must create a world that has a place for everyone, not just the academically talented.”2
deBoer’s recommendations flowed from his belief that behavioral genetic research has shown that “human behavioral traits, such as IQ, are profoundly shaped by genetic parentage, and this genetic influence plays a larger role in determining human outcomes than the family and home environment” (The Cult of Smart, p. 23). Simply put, he wrote that “children are shaped predominantly by nature, not nurture” (p. 128).
deBoer endorsed the behavioral genetic position that environmental influences can be divided into “shared and unshared” components, a position that “emerged most directly from twin studies” (p. 126). Also based on behavioral genetic claims, he argued that “shared” influences such as “parenting, the family, and home life simply don’t seem to matter much on a number of definable behavioral traits, including intelligence” (p. 128). Rather than reviewing the book as a whole, I will focus on how deBoer discussed the behavioral genetic research methods and studies he put forward as supporting a “predominant” role for genetic (nature) influences on human behavioral differences.
I am also on the left politically, which means that my analysis of deBoer’s claims comes from within that broad camp. Since 1998, with important help from others, I have developed a critique of behavioral genetic and psychiatric genetic research and theories in four books, and in many peer-reviewed and online articles. A significant portion of my focus since 1998 has been on major problems in behavioral twin and adoption research, and how false claims based on these research methods have been used to lend “scientific” support to policies maintaining the oppressive and economically very unequal political status quo throughout the world.
Like Harvard psychologist/author Steven Pinker before him, deBoer repeatedly invoked the “blank state” straw man caricature of environmentalists (in the causes of behavioral differences sense), who allegedly believe that “we… are all equally equipped at birth to succeed in academics and all that comes with them” (p. 28). I say caricature because few critics or environmentalists make such a claim. What they do argue is that while there may be inborn differences among individuals for certain abilities, it is crucial to understand that family, social, cultural, educational, geographical, and political environments together play a powerful role in shaping human behavior. They conclude that social, political, and academic attention should be focused on improving these environments, not on people’s brains and genes. Environmentalists point to the many characteristics and potentials we all share, and advocate for social policies or even radical social change to create societies providing healthy environments for everyone to grow up in and thrive.
As Current Affairs Editor-in-Chief Nathan Robinson pointed out in his 2020 review of The Cult of Smart, the book reads like a left-wing version of Richard Herrnstein and Charles Murray’s controversial 1994 book The Bell Curve. Unlike those authors, however, deBoer “unequivocally and explicitly” rejected the “position that some races are inherently more intelligent than others” (p. 122). Although in this article I focus on deBoer’s claims about genetics, I note Robinson’s negative assessment of the book:
“The central argument of the book is not just wrong, but wrong in the strongest possible sense of that term. It is based on fallacious reasoning. It is a mistake. An error. The whole argument falls apart completely the moment you touch it. Now, you can usually make an argument for almost anything, because so many arguments are based on subjective value judgments that are hard to prove or disprove. But unless you abandon the basic principles of logical inference, you can’t successfully make the argument of this book.”
The Cult of Smart was positively reviewed in the conservative National Review and in The American Conservative. Noting the irony of a leftist thinker attempting to take down environmentalist positions on behavior and cognitive ability, the conservative author of the National Review piece wrote that for him, reading the book brings forth the pleasure of “watching a pitcher on a [baseball] team you don’t cheer for throw a no-hitter against your crosstown rivals.”
Twin Studies
Although deBoer’s arguments were based on supposed findings from the behavioral genetics field, he admitted that he had no “expertise” in this area, and that he relied “on the work of scholars in that field to understand the dynamics I’m about to discuss” (p. 123). As he wrote in a post-publication blog post dated 8/10/2020, “I am a (well read) total amateur attempting to engage with scientific concepts.” He relied mainly on influential books popularizing behavioral genetic positions, such as Pinker’s The Blank Slate and Judith Harris’s The Nurture Assumption, and on articles such as behavioral geneticist Eric Turkheimer’s 2000 “Three Laws of Behavior Genetics and What They Mean.” deBoer did not cite a single original-data twin, adoption, or molecular genetic study in support of genetics.
Twins Reared Together “Twin Method” Studies
Studies of reared-together twins, deBoer recognized, have an “obvious confound” because “twins are usually reared together, in the same homes, by the same parent or parents.” For “cognitive traits like intelligence,” he reasoned, “it would seem that there’s much more opportunity for parental influence. In order to study purely genetic effects, we’d have to study twins reared by different people and in different homes, and that’s precisely what researchers do” (p. 124).
In the twin research area, therefore, deBoer didn’t seem to place much faith in studies using “confounded parentally influenced” reared-together MZ twin pairs, believing that studies using reared-apart twins provide a much better way of identifying “purely genetic effects”—a point I will soon return to. (MZ twins are also known as monozygotic or identical; DZ twins are also known as dizygotic or fraternal). Nevertheless, deBoer cited a 2015 twin study meta-analysis in favor of 40% to 50% heritability “for most behavioral traits” (p. 127). This meta-analysis was based on pooled studies using reared-together MZ versus DZ “twin method” comparisons, in other words, what deBoer saw elsewhere as “confounded” twin data.
As critics have argued since the 1930s, genetic interpretations of MZ-DZ comparisons are based on the false assumption that both types of twin pairs grow up experiencing roughly “equal environments.” The assumption is false because in addition to MZs being treated more similarly and spending more time together, levels of identity confusion and attachment are much higher among MZ pairs than among DZ pairs (evidence here). Although the academic consensus says otherwise, reared-together MZ-DZ twin method comparisons do not prove anything about genetic influences on behavior. (I explain why they don’t here.)
The “Heritability Fallacy”
Critics also argue that the heritability concept itself, which deBoer endorsed throughout the book, is both misunderstood and highly misleading, and does not indicate “how much” genes influence behaviors and abilities, or how changeable they are. As David Moore and David Shenk wrote in “The Heritability Fallacy,” the “term ‘heritability,’ as it is used today in human behavioral genetics, is one of the most misleading in the history of science.” Often based on twin study data, researchers calculate “heritability estimates” ranging from 0% to 100%, and these estimates are frequently found in textbooks and other authoritative works. The heritability concept was originally developed to predict the results of selective breeding programs of economically desirable traits in farm animals, such as milk production in cows. Since the 1950s it has been extended, controversially, as a measure of the “strength” or “magnitude” of genetic influences on various psychiatric disorders and human behavioral characteristics, including especially IQ.3
I agree with biologist Steven Rose, who concluded that the “heritability measure…except in the very specific context for which it was originally devised (agricultural breeding experiments) [is] rarely applicable, widely misunderstood and in most cases meaningless.”4 (More on problems with the heritability concept here.)
Twins Reared Apart (TRA) studies
We have seen that, due to “obvious confounds” in reared-together twin studies, deBoer based much of his argument in favor of high IQ heritability on studies of “twins reared by different people and in different homes.” These are known as “twins reared apart,” or “TRA” studies. Only six such studies have been performed, with the first appearing in 1937. Again, deBoer relied on influential behavioral genetic researchers and popularizers, not on his own reading of these original TRA study publications. He appeared to know little about these studies, whose massive flaws and biases have been highlighted by critics for decades (summarized here).
The “Minnesota Study of Twins Reared Apart.” The only TRA study deBoer mentioned was Thomas J. Bouchard, Jr. and colleagues’ widely cited Minnesota Study of Twins Reared Apart, or “MISTRA.” He did not cite any original MISTRA study publications or review articles, and none of the behavioral genetic authors he relied on described in any detail how the study was performed.
deBoer implied that Bouchard and his MISTRA colleagues obtained their reared-apart twins from the “Minnesota Twin Registry” (MTR), “which seeks to gather data from all of the twins born in that state over the course of several decades” (p. 125). Referring to me as a “dogged critic of twin studies,” deBoer claimed that I “argued that the [MTR] data set is flawed because many of the included twins were in fact separated only after years of cohabitation, confounding [the] analysis” (p. 133). He cited my book The Trouble with Twin Studies, yet mentioned only one small aspect of my larger critique of TRA research.5
To set the record straight, I did not write that the Minnesota Twin Registry “data set is flawed,” or anything else about the MTR in the TRA study context. This is because the MISTRA did not obtain its reared-apart twins from the MTR, which was based on reared-together pairs. Instead, Bouchard had to recruit twins, most of whom volunteered to participate after hearing about the study in the media. In her 2012 book about the study, Born Together—Reared Apart: The Landmark Minnesota Twin Study, MISTRA researcher Nancy Segal described the study’s twin sample as “a collection of cases”:
“Lacking a national registry, the MISTRA studied a collection of cases because our reared-apart twins surfaced in many ways and at any time….The majority of pairs were identified by self-referral, but most of the twins learned about the study through media reports.”6
Of the total MISTRA reared-apart twin sample, only 45% of the pairs were recruited from the United States, with the rest coming from overseas. Around one-third came from England.7
deBoer wrote that “many” TRA studies draw from data sets such as the MTR, with the “drawback” being that “any problems with that data would be replicated across different studies” (p. 125). That may be true for the numerous studies of reared-together twins, but here deBoer confused reared-together and reared-apart twin samples. Reared-apart twins are extremely rare, are exceedingly difficult to locate and to study, and there has never existed a registry of such pairs. deBoer’s unfamiliarity with twin research came through at every step.
Most twin pairs were only partially reared apart. In TRA studies, deBoer wrote, “it’s imperative that twins reared apart are truly reared apart. If the twins are raised partially together, it would jeopardize the inferences made based on the data set” (p. 133). Yet he provided no evidence that twins actually were “truly reared apart.” In fact, most studied reared-apart MZ twin pairs, also known as “MZA” pairs, were only partially reared-apart (evidence here).
In the larger context of kinship studies, deBoer wrote that “identical twins are closely aligned in behavioral and cognitive metrics even when raised in profoundly dissimilar environments” (p. 126). Here he seemed to be referring to single-case anecdotal reports, such as the “Jim Twins” and the now more-famous “Three Identical Strangers” triplets, whose cherry-picked “spooky similarities” have been reported in the media, and have been misleading the public, for decades. It is difficult to determine how similar or dissimilar the 81 MISTRA MZA pairs’ environments actually were, because Bouchard and colleagues denied critics and independent observers access to their raw data.8
Does the MZA correlation “directly estimate heritability”? The standard behavioral genetic position is that the MZA correlation “directly estimates heritability.” For example, a .50 MZA IQ-score correlation = 50% IQ heritability. Following this script, deBoer wrote,
“If identical twins raised apart are exactly like each other on a given cognitive trait, it would imply that the trait is purely genetic; if identical twins raised apart are entirely unalike each other on a given cognitive trait, it would imply that the trait is purely environmental. And any values in between could be used to assess what proportion of the (population-level!) trait is genetic or environmental.” (p. 124)
According to this view a significantly above-zero MZA correlation, for example IQ test scores, is like a dimmer switch of genetic importance—the brighter the light (the higher the correlation), the stronger the genetic influence. deBoer and the behavioral geneticists are mistaken, however, because in addition to problems with the heritability concept itself, even perfectly separated MZA pairs—simply because they are born at the same time—grow up sharing many non-familial “cohort” influences in common (over 30 listed here). MZAs also share a prenatal environment, a very similar physical appearance, and are the same sex. All these factors will contribute to MZA behavioral similarity for non-genetic reasons, and thereby confound genetic interpretations of MZA behavioral correlations.
Bouchard, Segal, and their MISTRA colleagues got around these problems by keeping their raw data secret, and by counting the environment as genetic (yes, you read that correctly). As seen in their widely cited 1990 Science article, their conclusion that IQ is “strongly affected by genetic factors” was based on their decision to count environmental influences on MZAs’ behavioral resemblance as genetic influences—thereby creating a “heads I win, tails you lose” argument they could not lose, because they counted both the “nature” and “nurture” sides of the behavioral coin as nature.9
As I showed in The Trouble with Twin Studies, the Minnesota Study of Twins Reared Apart was massively flawed, heavily genetically biased, and based on numerous false or questionable assumptions. The researchers had to suppress their reared-apart DZ control group IQ correlations to reach the desired genetic conclusions. Had they decided to publish these correlations in their main IQ paper (the 1990 Science article), they would have been forced to conclude that their study found no evidence that genes influence IQ scores and “general intelligence” (explained here).
Adoption Studies
The other main kinship research method DeBoer cited was the study of non-twin adoptees and their biological and adoptive families. Researchers conducting these studies typically conclude that adopted children correlate higher with their biological as opposed to their adoptive relatives, as genetic theories predict. Behavioral geneticists assume that in these adoption studies, birthparents share genes but not environments with their adopted-away offspring. Based on this assumption, in the IQ studies they estimate heritability by doubling the birthparent/adopted-away offspring IQ-score correlation.10 However, even children adopted away at birth share several environmental factors in common with their birthmothers. This always includes the prenatal environment, usually includes skin color and “race” (frequently leading to oppression or privilege), and often includes similar physical appearance, social class, culture, religion, and so on.
Moreover, couples wishing to adopt children are carefully screened by adoption agencies for emotional and financial stability, and children are often intentionally placed into adoptive homes matched on some of the characteristics—including the assumed genetic status—of their biological family background. “Fitting the home to the child,” a psychology researcher acknowledged in 1975, “has been the standard practice in most adoption agencies, and this selective placement can confound genetic endowment with environmental influence to invalidate the basic logic of an adoptive study.”11 This is known as “selective placement bias.” Other biases in adoption research include attachment rupture and its impact on a child’s developing brain, late separation from the birthparent, late placement after separation, and the environmental “range restriction” found in the adoptive parents group.12 interpretations and conclusions” found in IQ adoption study publications.] For more on these and other IQ adoption study problem areas, see the 2006 analysis by Ken Richardson and Sarah Norgate.
Genetically oriented writers usually focus on the correlational adoption studies, whereas environmentally oriented commentators often highlight studies showing large IQ gains made by children born to poor parents, who are then adopted by professional or upper-class parents. In one such study, the biological children of poor or working-class parents showed a substantial 14-point IQ-score increase when they were raised in the adoptive homes of families in the upper ranges of the socioeconomic scale. The behavioral genetics popularizers that deBoer relied on tend to downplay or ignore such findings.
A more fitting name for this area of research is the study of abandoned and rejected children. Abandonment and rejection are among the most psychologically harmful events a child can endure, and for this reason alone it is questionable whether the results of “adoption” studies apply to people who were not abandoned or rejected as children.
Molecular Genetic Research
For decades, molecular genetic researchers have attempted, and for the most part have failed, to identify specific genes shown to cause (not just be “associated with”) differences in IQ, the major psychiatric disorders, and behavioral differences in general. Citing only a 2017 New York Times article by science writer Carl Zimmer, deBoer claimed that researchers using Genome-Wide Association Studies (GWAS) have discovered seemingly “clear” associations between intelligence and some genetic variants (p. 136). The Times and other corporate media outlets regularly publish sensationalized false-alarm gene discovery claims for intelligence, psychiatric disorders, violent crime, loneliness, gang membership and other areas of behavior. This helps lead their readers to the conclusion that human behavioral differences are largely “in the genes.”
Many people on the left understand that corporate media outlets at times provide misleading and biased coverage in politically related areas of science, just as they do in relation to war, foreign policy, and political issues in general. This doesn’t mean that leftists and progressives shouldn’t cite or discuss politically relevant science articles appearing in the corporate media, but they should approach such articles with a critical eye when their authors’ claims align with those of the outlet’s corporate ownership.
deBoer seemed unaware of the fact that countless false-alarm behavioral gene-association publications have appeared since the 1960s. Decades of “genes for behavior” false alarms indicate that there are major methodological flaws and biases underlying these studies, and lend support to the critics’ contention that molecular genetic researchers have been misled by twin studies, adoption studies, and heritability estimates. This is also true for more recent studies based on “polygenic risk scores” (PRS).
In his 2014 book Misbehaving Science, sociologist Aaron Panofsky described the behavioral genetic gene-discovery failure “coping strategy” of “technological optimism.” By this he meant the “optimism that the next level of technology will overcome past disappointments.”13 Most likely, current excitement about PRS studies in behavioral genetics will fade when this method ends up becoming yet another “next level” method that failed.
Given the dreadful track record, by default we should assume that current and future behavioral gene-association claims, and polygenic risk scores, are false-positive and/or non-causal findings until proven otherwise.
Conclusions
In The Cult of Smart, Fredrik deBoer attempted the difficult task of reconciling his socialist beliefs with what he saw as “the fact that our genetic heritage deeply influences our behavioral selves, including our academic selves…” (p. 23). His admirable goal was to recommend ways to improve the “broken” U.S. education system, but he knew little about the twin, adoption, and molecular genetic studies upon which his recommendations were based. It’s fine not to know much about these things, but it’s not fine to publish a book whose arguments are based on research one knows little about, and which may even lend support to “strange bedfellow” political forces one opposes.
A week after the book came out, deBoer published a blog post with the puzzling title, “If Genes Contribute Nothing, My Conclusions Are All the Same.” In it, deBoer wrote,
“some serious people with serious evidence believe that our genetic endowment shapes our behavioral outcomes like how well we perform in school. Whether they’re right or not, I can’t be the one to say.” (italics added)
A major theme of The Cult of Smart, however, was not that “serious people’s” claims that “genetic endowment shapes our behavioral outcomes” might be right, but that they are right.
In the book, deBoer discussed the characteristics of “ardent hereditarians” and the “well-meaning” people who oppose them:
“On the one hand, you have the ardent hereditarians who reduce all of human life to the outcome of genes, and who are often incorrigibly racist. On the other hand, you have well-meaning but misguided people who dismiss the importance of genetics out of hand, who associate any discussion of the heritability of behavioral traits with the worst elements of eugenics and colonialism, and who make grandiose claims about a scientific literature they have never read. This leaves us unable to argue honestly and persuasively when we need to most.” (p. 139)
As we have seen, deBoer himself made claims about a scientific literature that for the most part he has never read. But I do agree with him that anti-hereditarians should not reflexively answer hereditarian claims with charges of “eugenics and colonialism,” even though it is important to know and write about this history. And it isn’t even necessary since, as seen in the 1996 book Inequality by Design, and in Ken Richardson’s 2017 Genes, Brains, and Human Potential and elsewhere, hereditarian and racial-differences arguments fall apart upon close examination.
deBoer concluded that his fellow progressives and socialists should know that “genes matter,” and advised them not to leave the discussion to “pseudoscientific racists” and “the worst elements”:
“Genes matter. They matter for us as individuals and they matter for us as a society. And they matter whether progressive people believe them to matter or not. What the political left must understand is that sticking our heads in the sand and ignoring the rapidly developing world of human genetic research does no one any favors. To leave the field of argument to the worst elements, the pseudoscientific racists and gender essentialists, is to abdicate our sacred responsibility to defend our values.” (p. 237)
If behavioral genetic “discoveries” were real, then deBoer’s advice would make some sense. But they are a mirage, a house of cards that ignores contradictory evidence from countless real-world examples and research findings from other fields, that collapses under serious critical analysis.14 In the process of exposing between-group racist pseudoscience, then, progressives and socialists must understand that claims relating to the “high heritability” of within-group “individual differences” in IQ and other behavioral areas are also based on flawed research and concepts, including the IQ and “general intelligence” concepts themselves. Learning more about this through the works of critics will help arm them in the debates and struggles ahead.15