Book review: Jay Joseph, The Trouble with Twin Studies. A Reassessment of Twin Research in the Social and Behavioral Sciences, New York, Routledge, 2015.
Let me first explain this review’s unusual length: I think strong claims require substantial evidence. The reader will soon see why I felt the need to back my criticisms with quite lengthy quotes.
Jay Joseph has previously written two books dedicated to the criticism of twin studies: The Gene Illusion (2003/4) and The Missing Gene (2006). However, there is an essential difference between these books and The Trouble with Twin Studies; the latter clearly is a more difficult read due to the author’s willingness to enter technical considerations, although Joseph explains he tried to combine precision with accessibility:
“My original intention had been to write a book that could be understood by a larger audience, but as the work progressed, I realized that this would be a more difficult task than I first imagined. At the same time, I have tried to cover the main topics and controversies as simply as possible.” (p. xi)
The book contains three parts that are dedicated, respectively, to studies of twins reared apart, studies of twins raised together and the future of twin studies.
Twins Reared Apart
For twin researchers, the existence of genetic influences on human behavior is no longer under dispute. Joseph disagrees and argues that the question is still a matter of controversy:
“the controversy had centered on whether genetic or environmental factors have a predominant influence on human behavior, or more properly, on human behavioral differences.” (p. 4)
He first provides a long analysis (p. 18-48) of the four studies of reared-apart twins conducted before the Minnesota Study of Twins Reared Apart (MISTRA). These four studies are Popenoe’s 1922 study (followed by a 1925 report by Muller), Newman and his colleagues’ 1937 study, Shields’s 1962 study, and Juel-Nielsen’s 1965 study. Joseph concludes that these studies “have provided no scientifically acceptable evidence in support of genetic influences on human behavioral differences” (p. 48). While many of Joseph’s points are worth reading (e.g., some twins were only partially reared apart), one can seriously ask whether such a strong conclusion is warranted. First, the author’s objections are numerous (e.g., p. 22-24), but he does not show that they are so critical as to make the results of these studies utterly unreliable. Secondly, it is one thing to assess the biases of individual studies, but it is another one to say that the four studies together do not provide any acceptable scientific evidence.
Science always proceeds by aggregating imperfect data (no design can be perfect), and if different studies with different flaws and different methodologies all suggest the existence of genetic influences, one can reasonably consider that these influences may be real. In addition, along with studies of twins reared apart, the twentieth century progressively saw the accumulation of studies pointing out genetic influences on human behavior. If we take intelligence as an example, Bouchard and McGue’s 1981 systematic review is based on 111 familial studies of intelligence, excluding Cyril Burt’s studies. It includes studies of MZ twins reared apart, MZ twins reared together, and DZ twins reared together, but also adoption studies, comparisons between siblings reared together, and siblings reared apart, between half-cousins, between cousins, between non-biological siblings reared together, and even studies of assortative mating.
Bouchard and McGue note: “the pattern of averaged correlations is remarkably consistent with polygenic theory” (p. 1058 of their review). Again, consistency is an important factor: if researchers obtain results that are incompatible with those of mainstream science, they are much more likely to look for potential flaws in their study. By contrast, if they find out that their results are consistent with those of dozens of studies of different kinds and conducted with many methodologies, and since it is statistically unlikely that different methodological flaws in dozens of studies of different types consistently lead to a similar big picture, they can reasonably think that their studies are reasonably well designed or, at least, that their flaws do not make the results utterly unacceptable.
Thereafter, Joseph discusses the Cyril Burt scandal. This is a relevant topic since Burt was accused of fabricating data related to reared-apart twins. But then, Joseph starts a relatively lengthy discussion on scientific fraud in general and fraud in psychological research in particular (p. 48-50). One may ask whether this extensive discussion was needed. Psychology is by no means a uniform research area, and although fraud is always a theoretical possibility, behavioral geneticists should be criticized on the basis of their publications, not those of psychologists as a whole. Many behavioral geneticists would add that they have been able to obtain results that now replicate, partly because a number of public controversies compelled them to meet always higher standards of evidence.
Joseph also discusses Arthur Jensen’s 1969 famous article, due to the impact it had on the public perception of intelligence research and behavioral genetics. This is where the author’s arguments start to be really objectionable. In this article, Jensen suggests (among others) that genetic factors may be part of the reason why Whites score higher, on average than Blacks on IQ tests. Joseph describes this view in a way which makes it sound much more offensive than the original article: he writes that the authors of The Bell Curve “cautiously endorsed Jensen’s position that black people are genetically inferior in measured intelligence” (p. 51). But Jensen is a much more careful man than what Joseph’s rewording would have the reader believe. The word “inferior” does not appear a single time in his 1969 article, and he wisely warns against potential misinterpretations and misuses of his research:
“it is unjust to allow the mere fact of an individual’s racial or social background to affect the treatment accorded to him. All persons rightfully must be regarded on the basis of their individual qualities and merits, and all social, educational, and economic institutions must have built into them the mechanisms for insuring and maximizing the treatment of persons according to their individual behavior.” (A. Jensen, “How Much Can We Boost IQ and Scholastic Achievement,” Harvard Educational Review 39, fasc. 1, p. 78)
Even more problematic is what Joseph writes on Jensen’s concerns about the link between fertility and intelligence among Blacks and Whites. Jensen is worried about the finding that this correlation is negative and that this negative correlation is much stronger for Blacks than for Whites, which he fears may imperil future achievement of this minority:
“Certain census statistics suggest that there might be forces at work which could create and widen the genetic aspect of the average difference in ability between the Negro and white populations in the United States, with the possible consequence that the improvement of educational facilities and increasing equality of opportunity will have a decreasing probability of producing equal achievement or continuing gains in the Negro population’s ability to compete on equal terms.” (Jensen’s article, p. 95; Jensen’s emphasis)
“Negro middle- and upper-class families have fewer children than their white counterparts, while Negro lower-class families have more.”(Jensen’s article, p. 95)
Joseph quotes a passage which relates to this concern, in which Jensen asks whether there is “a danger that current welfare policies, unaided by eugenic foresight, could lead to the genetic enslavement of a substantial segment of our population? The possible consequences of our failure seriously to study these questions may well be viewed by future generations as our society’s greatest injustice to Negro Americans” (Jensen’s article, p. 95, quoted by Joseph, p. 52). Thus, for Jensen, this negative correlation is the problem, and future generations – especially Blacks – would have every reason to blame 20th-century society’s inability to revert it. But for Joseph, what Jensen actually believes is that Blacks themselves are the problem:
"In other words, Jensen believed that eugenic measures were needed to reduce black people’s rate of reproduction, which would reduce their numbers in the U.S. population and therefore raise the national intelligence level.” (p. 52)
This is an extremely grave accusation, and what I have quoted above of Jensen’s article clearly shows that it does not withstand scrutiny. I encourage all readers of this review to read Jensen’s original article and to compare it with Joseph’s accusation, and also to ask what kind of mindset is needed to interpret the article the way Joseph does.
At the end of chapter 2, Joseph gives a presentation of the MISTRA (p. 53-58), but it is only in the following chapter that he starts his critique. In this chapter, “Studies of reared-apart twins: the critics respond” (p. 61-74), Joseph provides a long list of criticisms that were made against twin studies from the 1970s on, including those of psychologist Leon Kamin (1974), sociologist Howard Taylor (1980), Susan Farber (1981), Kamin’s arguments in Eysenck vs Kamin (1981), Richard Rose (1982), Richard Lewontin, Steven Rose & Leon Kamin (1984), Ken Richardson (1998), Leon Kamin & Arthur Goldberger (2002), and Joseph’s own critiques. There are two problems with the way Joseph proceeds.
Firstly, he resorts to a rhetorical tactic known under the name of Gish gallop; the ‘Gish gallop’ consists in overwhelming an opponent with as many arguments as possible, irrespective of their validity or relevance, to prevent this opponent from refuting them all. Joseph mentions many alleged problems with twin studies, but he does not demonstrate the validity of the critics’ objections, nor does he tell the reader which arguments have already been answered to. This reminds me of what Einstein is said to have said about the book 100 Authors against Einstein, a memorable example of Gish gallop:
“Why one hundred? If I were wrong, one would have been enough.” (quoted by Michael Shermer, see here)
Secondly, and even more problematically, Joseph does not quote or even provide references to the twin researchers’ replies to all these objections. The reader is presented with a one-sided view of the debate: while the critics’ arguments are extensively quoted or described, twin researchers are simply not given a fair hearing. He quotes a few words of Bouchard’s 1983 reply to Taylor, but only to denounce Bouchard’s use of the word “pseudo-analysis” and tone towards critics (p. 64). Joseph does not address the substance of Bouchard’s critique, and what he quotes does not even give a vague idea of Bouchard’s detailed analysis of Taylor’s arguments and objections. In this 1983 article, Bouchard also addresses the criticisms made against the four studies of twins reared apart that were conducted before the MISTRA – also criticized by Joseph himself (see above), but while Joseph quotes Taylor as one of his sources (p. 23 & p. 40), he completely ignores Bouchard’s reply to Taylor about these four studies.
In his reply to Eric Turkheimer’s review of The Trouble with Twin Studies, Joseph wrote: “I argue that criticism of behavioral genetic research strengthens science since good science is greatly served by rejecting and casting out bad science” (see here). This is a valid point, but one does not make science better by ignoring so much of what the researchers already replied to their critics.
In chapter 4, “Studies of Reared-Apart Twins: Basic Assumptions and Potential Fallacies” (p. 75-101), the Gish gallop goes again, but in another form: instead of a chronological list of arguments against twin studies, Joseph presents categories of potential fallacies, including (but not limited to) “Psychometrics” (p. 76-77), “Heritability” (p. 77-82), Model Fitting (p. 82-87), “Random Assignment” (p. 87-88), and “IQ tests” (p. 88-90). Here again, Joseph presents numerous arguments without distinguishing the valid ones from those that were refuted long ago. Perhaps one of the most striking examples of this attitude is what he writes about IQ tests and their usage by twin researchers.
“Over the years, critics have highlighted many problems in IQ testing, which include: (a) that general intelligence is merely a statistical artifact, and therefore has no physical reality; (b) that there is no consensus definition of “intelligence”; (c) that IQ tests measure school-related learning more than general cognitive ability; (d) that there are different types of intelligence, as opposed to a single general factor; (e) that IQ tests measure only narrow abilities and ignore “real-world” intelligence; (f) that IQ tests serve the interests of the upper classes to “scientifically” legitimatize their position in society based on the alleged biological reality of human inequality; and (g) that IQ tests are biased on the basis of race, culture, and economic class” (p. 89)
Once again, if Joseph believes that IQ tests are questionable, one good argument is enough! In fact, all the arguments listed above have already been refuted. Also, Joseph’s usage of arguments that are devoid of scientific validity (e.g., that IQ tests serve the interests of the upper classes) actually weakens his argumentation.
The same problem appears with Joseph’s eight objections to the concept of ‘heritability’ and to the production of heritability estimates (p. 77-78). His seventh objection is:
“The finding of high heritability within populations says nothing about whether genetic differences exist between populations.” (p. 78)
This is absolutely true, as Bouchard himself recognized in his 1995 review of The Bell Curve... and it is also entirely irrelevant! This warning against a speculative usage of heritability coefficient for understanding group behavioral differences do not in any way invalidate the usage of these coefficients for assessing the heritability of individual differences.
In chapter 5, “The Minnesota Study of Twins Reared Apart I: Biases, Assumptions, and Other Problem Areas” (p. 102-127), Joseph tackles the Minnesota Study of Twins Reared Apart, and he keeps accumulating arguments without referring to the replies made by Bouchard or by other behavioral geneticists. The reader cannot know, from Joseph’s book, how the twin researchers reacted (or would react) to allegations such as the following:
“Perhaps many MZA pairs were truthful, but there were incentives for them to exaggerate their degree of separation (or even to invent their separation), or to concoct “eerie” similarities between themselves.” (p. 118)
Joseph’s objections include categories such as “Sample Size” (p. 120), “Confirmation Bias” (p. 120-122), and even “Use of the MISTRA Findings by Disreputable Groups” (p. 125-126). What does the latter do in a scientific monograph about the fallacies of twin studies?
Chapter 6, “The Minnesota Study of Twins Reared Apart II: IQ and Personality Studies” (p. 128-149), is more focused: it consists of a long list of correlations for IQ and personality that were not reported by the Minnesotan twin researchers. Generally speaking, it is entirely legitimate to ask researchers to make their data available to the public and to their fellow researchers. Nonetheless, one can ask why Joseph feels the need to provide such a circumstantiate list to the reader if he accepts neither the validity of IQ testing nor the validity of heritability coefficients. In this case, his critique is focused on the attitude of behavioral geneticists, not the intrinsic validity of the research they conduct. Once again, one feels that Joseph would to resort to any argument to criticize Bouchard and other behavioral geneticists.
Twins reared together: Not-so-circular arguments
The second part of Joseph’s book (p. 151-203) is dedicated to the studies of twins raised together, much more numerous than the reviews of twins reared apart. It starts with a discussion of what is usually called the Equal Environment Assumption (EEA), that is, the assumption that monozygotic twins and dizygotic twins have similar parental environments and that, as a result, the greater similarity of monozygotic twins can be explained by their genetic similarity and not by the attitude of parents who would treat the first ones more similarly than the second ones, for instance, by giving them identical clothes. Joseph’s central question is: Does this assumption hold?
He presents two arguments made by twin researchers to support the EEA: Argument A is that twins create their own environment and that, as a result, greater genetic similarity between them results in a greater similarity of their environment (p. 158-160). Argument B is that EEA does not posit the absence of environmental differences but does posit an equal environment for the trait under study (p. 160-163). For instance, parents may be more likely to give the same clothes to identical twins, but such an attitude would not explain these twins’ greater similarity, compared to fraternal twins, for traits such as alcoholism or intelligence.
Let’s start with Argument A. In Joseph’s view, this argument is problematic because it is a circular argument, which could be described as follows:
- Premise: Since identical twins have greater genetic similarity than fraternal twins, they will create more similar environments; behavioral similarity will be caused by genes.
- Finding: Identical twins are more similar in behavior than fraternal twins.
- Conclusion: Identical twins are more similar in behavior because of their greater genetic similarity (restates the premise).
One of the examples of Argument A given by Joseph is the following quote, from an article by Nicholas Martin, Dorret Boomsma, and Geoffrey Machin:
“A series of ingenious studies ... have all pointed to the conclusion that, for the most part, the more similar treatment of MZs is not the cause of their greater phenotypic similarity, but, rather, a consequence of their genetic identity and the more similar responses this elicits from the environment.” (quoted by Joseph, p. 158)
Indeed, this would be a problematic argument if it really were what twin researchers are arguing. However, in his quote, Joseph has suppressed a reference to a crucial piece of evidence used by researchers, the studies on mistaken zygosity. Here is the original text:
“As good as the twin design is, it is not without its problems and detractors. The most constant and potentially the most damaging criticism is that MZ twins share more similar post-natal environments than DZ twins, so their excess similarity cannot necessarily be attributed to their genetic identity. The objection has face validity, as it has been repeatedly shown that MZs do indeed experience more similar environments both as children and later in life. However, a series of ingenious studies, involving ethological observations and making use of mistaken zygosity diagnosis and differing degrees of separation, have all pointed to the conclusion that, for the most part, the more similar treatment of MZs is not the cause of their greater phenotypic similarity but, rather, a consequence of their genetic identity and the more similar responses this elicits from the environment.” (N. Martin, D. Boomsma & G. Machin, “A twin-pronged attack on complex traits,” Nature Genetics 17, December 1997, p. 390; I emphasize the crucial part omitted by Joseph).
Studies on ‘mistaken zygosity’ are focused on fraternal twins who are perceived and treated by their parents as identical twins, or vice versa. Their results indicate that, as far as behavioral characteristics are concerned, fraternal twins who are mistaken for identical twins are actually as similar as fraternal twins, while identical twins that are mistaken for fraternal twins are as similar as identical twins. This suggests that it is their genetic makeup, not the way they are treated by their parents, that explains their degree of genetic similarity. Once this is taken into account, it is no longer reasonable to accuse twin researchers of circular reasoning.
Argument B is presented by Joseph as an argument ad ignorantiam: he writes that twin researchers make the mistake of arguing that it is incumbent on their critics to demonstrate that the environment can affect the trait under study, genetics being the default hypothesis. Again, Joseph’s reasoning is no longer tenable when one takes into account the results of studies on mistaken zygosity: they indicate that fraternal twins who are treated as identical twins do not become as similar as identical twins, and vice versa. This contradicts the idea of a decisive influence of parental environment. As a result, twin researchers can legitimately ask for more evidence for strong evidence of the familial environment, and it needs to be strong enough to overwhelm the results of mistaken zygosity studies. Thus, with the results of studies on mistaken zygosity, it is reasonable to reject Joseph’s conclusion that twin studies would be “unable to disentangle” (p. 177) genetic and non-genetic influences.
To be precise, Joseph does discuss the results of mistaken zygosity studies in his two previous books, The Gene Illusion (p. 92-94) and The Missing Gene (p. 191-195), but his arguments about the EEA are entirely different in these books. In The Gene Illusion, he criticizes two such studies, by Scarr (1968) and Scarr & Carter-Saltzmann (1979), on the basis of various points, none of which is extensively developed – again, a Gish gallop; Joseph does not refer to similar studies that were already available when we wrote his book, by Munsinger & Douglass (1976), Matheny (1979), Goodman & Stevenson (1989), and Kendler et al. (1993). In The Missing Gene, Joseph mentions all six studies, but he uses a surprising argument. In response to the idea that greater genetic similarity leads to more similar environments, he writes:
“Their critical error and that of twin researchers ever since is that the reason identical twins experience more similar environments than fraternal twins has no bearing on the validity of the twin method. Contrary to the views of most contemporary twin researchers, the only relevant question is whether, as opposed to why, identical and fraternal environments differ.” (The Missing Gene, p. 193)
In other words, Joseph states that the chain of causation between behavioral similarity and environmental similarity is irrelevant. Perhaps he later realized that this argument was not really convincing and that, fundamentally, the best tactic was instead to hide the smoking gun.
Chapter 8 (p. 179-203) is focused on the usage of the results of twin research in psychiatry; it includes reflections on the EEA in psychiatry (p. 184-189), but also on problems with psychiatric diagnoses as such. Joseph denounces the “failure to demonstrate the reliability, the validity, biological basis, and genetic basis of its diagnoses, in addition to its cozy relationship with the makers of highly profitable psychotropic drugs (dubbed “Big Pharma” by the critics)” (p. 190-191). On the alleged failure to demonstrate the biological basis of diagnoses, I invite the reader to check Daniel Hanson’s review of Joseph’s earlier book The Gene Illusion. As for “Big Pharma,” financial interests can be a matter of concern but should never constitute as such a reason for dismissing the results of scientific studies. As in earlier chapters, Joseph uses arguments that are not linked to the methodology, the assumptions, or the results of twin studies. For instance, he writes that “psychiatric labels often have a harmful and stigmatizing effect and can dehumanize people and obscure the meaning of their behavior” (p. 191). It is true that these labels may have their limitations, but if Joseph wants to get rid of them, what alternative does he offer to identify, and potentially cure psychiatric problems? More importantly, how do these considerations improve the reader’s understanding of twin studies? Is it supposed to be a book about twin studies, or about psychiatry?
No future for twin studies?
In part 3, “Approaching a Post-Behavioral-Genetics Era?” (p. 205-252), Joseph starts by giving numerous quotes from researchers who confess their inability to find genes associated with behavior (p. 209). One of these quotes actually includes the main reason given by molecular geneticists for the failure to find such genes. This reason is tiny effect sizes:
“QTL [quantitative trait loci] associations have been reported for several candidate genes and personality traits. However, similar to research on psychopathology, replication of associations has been difficult in part because effect sizes are much smaller than originally anticipated. Genomewide association studies have also not yet yielded consistent results.” (R. Plomin, J. C. DeFries, V. S. Knopik & J. M. Neiderhiser, Behavioral genetics 6th ed., New York, Worth Publishers, 2013, p. 296, quoted by Joseph, p. 209; my emphasis)
But Joseph does not seriously address this possibility. Rather, he commits the fallacy he previously accused others of (p. 160-163), an argument ad ignorantiam:
“Ironically, the only positive contribution that the field of psychiatric genetics has ever made to the human condition is its finding that genes for the major psychiatric disorders do not appear to exist.” (p. 202)
At this point, the alert reader may ask what Joseph thinks about the genetic variants that are known to have significant effects on health conditions, e.g., BRCA1 and BRCA2 for breast cancer and the E4 variant of ApoE for Alzheimer’s disease; these variants were discovered long before the GWAS era. In fact, Joseph does not refer to them at all in his book, even though the existence of these genes is mentioned in an article quoted by Joseph more than ten times (and thus, probably entirely read) by Joseph himself, p. 212-213 (Latham & Wilson, 2010). This is understandable: if Joseph were to recognize that such genes can have an impact on health and also, for the E4 variant of ApoE, on human behavior (in this case, Alzheimer), it would become harder for him to argue against the possibility that there may also be behavior-related genes with tiny effects.
After a lengthy section in which Joseph highlights excessively optimistic predictions made by behavioral geneticists (p. 223-234), he offers a “Human Genetics Parable” (p. 236-246), about a village where arsonists are burning houses, and the mayor focuses first and foremost on the houses’ resistance to fire. In his review of Joseph’s book, Eric Turkheimer has already noticed the problematic nature of this metaphor, which actually implies that the material used to make houses may have some importance. What I find very interesting is Joseph’s reaction to Turkheimer’s remarks about the metaphor:
“Although in retrospect I should have explained this better, the main purpose of the parable was to show that even if there are genetic differences among people that contribute to different behavioral outcomes, an emphasis on genetics is still a wrong and even harmful approach. Behavioral genetics, despite its’ leaders’ frequent statements that environmental statements that environmental factors are important, promotes an approach that emphasizes genetics, largely absolving society and political leaders from the need to improve people’s social, familial, physical, and political environments.” (bold in the original; see here)
I sympathize with Joseph’s concern, and I agree that there is a nontrivial (but I would not say inevitable) risk of people emphasizing genetic factors to downplay the importance of environmental factors. For instance, it is not without reason that the tobacco industry provided financial support to Hans Eysenck, who investigated possible genetic factors for lung cancer and coronary heart disease. However, I would also argue that the benefits of twin research largely outweigh the possible risks. Firstly, it is dubious whether a majority of those who attribute a significant role to genes would also renounce to improve the social, cultural and political environment. In fact, a recent study – published after Joseph’s book – actually indicates that liberals are more, not less likely than conservatives, to attribute individual differences to genetic factors (see here). One can also notice the attitude of Robert Plomin, who wrote in his book Blueprint that “economic inequality could be dealt with directly through a redistributive tax system that reduces the gap between rich and poor” (R. Plomin, Blueprint, London, Allen Lane, 2018, p. 104). This is the vision of a man who also thinks that “[i]nherited DNA differences are by far the most important systematic force in making us who we are” (p. 31).
Secondly, the discovery of genetic influences may yield significant benefits by helping us understand why people are the way they are, and in particular why some people have psychiatric diseases or anomalies. More specifically, since genetic disorders are not the result of anyone’s conscious choice, it may be helpful for the parents of a child with such a disease that their behavior did not cause this child’s condition. Naturally, when someone is guilty (e.g., the tobacco industry), it needs to be called out, but when genes are the reason for a disease, it is better to know it instead of unsuccessfully looking for potential scapegoats.
Thirdly, the findings of behavioral genetics can benefit not only specific individuals, e.g., the families of those having genetic diseases but also society as a whole. To make the world a better place, as both Joseph and I want to, we first need to understand it, and in particular, we need to know what policies are likely to result in the outcomes that we regard as desirable. In this regard, perhaps one of the most sorrowful examples of unrealistic and avoidable policies is the No Child Left Behind Act. Without any regard for humanity’s intellectual variability – which, the findings of behavioral genetics tell us, is partly the result of genetic factors –, its stated aim was to make all school children reach a level defined as “proficient” within 12 years. As Charles Murray insightfully puts it:
“The United States Congress, acting with large bipartisan majorities, at the urging of the President, enacted as the law of the land that all children are to be above average” (quoted from here).
Once again, when environmental factors are known to harm people, they must be acted upon. But we should not accept without admissible scientific evidence the idea that equal opportunities combined with the absence of social and economic handicaps would automatically result in equal outcomes. If we insist on believing this and find out that some inequalities seem to be impossible to suppress, it can only result in resentment and hostility. This, we should all know, would definitely not make the world a better place.
As I have explained before, Joseph’s book contains some arguments that are irrelevant for assessing the scientific validity of twin studies. This is exemplified by the fairly numerous references to eugenics, to destructive policies, and to powerful interest groups (p. 18; p. 63; p. 64; p. 125; p. 157; p. 201; p. 236; p. 237). The same can be said about Appendix A (p. 253-258), which is entirely dedicated to the funding of the MISTRA by the Pioneer Fund.
Appendix B, “A Little-Known Behavioral Genetic Adoption Study Whose Results Contrast with the MISTRA” (p. 259-266), is about a 1998 study of the Colorado Adoption Project which, in Joseph’s view, would reveal biases in the attitudes of behavioral geneticists:
“This story is not as much about a specific adoption study as it is an example of how the genetic biases of behavioral genetic researchers influence their investigations, including most importantly the assumptions they accept and the conclusions they reach.” (p. 259)
More precisely, Joseph writes about their reaction to the results of this study:
“Clearly, Plomin and colleagues were in the camp of those seeing adoption studies as providing a better method of disentangling genetic and environmental factors than that provided by twin method MZT-DZT comparisons. It follows that they should have concluded that the results of the 1998 CAP study indicated that there is something wrong with twin studies of personality and that previous interpretations of twin data in favor of genetics should be reevaluated. Instead, they wrote that “tests of the equal environment assumption of the twin method generally support the reasonableness of the assumption” (Plomin et al., 1998, p. 216). The arguments put forward in support of this position were examined in Chapters 7 and 8, and were rejected.” (p. 264)
Joseph presents a false alternative, in which one should reject either MZT-DZT comparisons, or the results of the CAP study. In fact, as Joseph himself acknowledge, the researchers have found a possible solution to this apparent paradox:
“The researchers argued that studies based on first-degree relatives, such as their 1998 CAP study, “will not detect nonadditive genetic variance (p. 217), and chose to conclude that their results can be explained by “nonadditive genetic influence, which can be detected by twin studies but not by adoption studies” (p. 211; for a critical review of the CAP, see Richardson & Nordgate, 2006).” (p. 264).
Just after this paragraph, Joseph starts a section on “Alternative Explanations of the Results” (p. 261-263). But if he disagrees with the hypothesis offered by the researchers – namely, that twin studies detect additive plus nonadditive genetic influences and adoption studies only additive genetic influences –, he should have at least tried to address it, rather than directly starting a new section. In addition, it is misleading to present this 1998 publication as a study whose embarrassing results would have been downplayed by twin researchers, whereas the difference between the results of twin studies and adoption studies are in fact largely recognized within the field of behavioral genetics (see M. Trzaskowki, P. S. Dale & R. Plomin, “No genetic influence for childhood behaviour problems from DNA analysis,” Journal of the American Academy of Child and Adolescent Psychiatry 52, 2013, p. 1048-1056). The differences between the results of twin studies and adoption studies need not be seen as contradicting each other. In fact, they are a good example of the benefits of combining different methods while being aware of the limitations of each of them, to see what the “converging evidence” suggests.
Finally, Appendix C (p. 267-272) includes a list of quotations using what Joseph described as Argument A (see above) to support the Equal Environment Assumption. The quote from Martin, Boosma & Machin’s article appears again in its incomplete form, p. 270. Joseph also quotes the following passage of a 2010 article by Patrick Sturgis and his colleagues:
“If MZ twins are treated more similarly because they are biologically more alike, this can hardly be considered a violation of the EEA. For the reason that MZ environments are more similar than DZ environments (if indeed they are) is because of the initial difference in genetic predispositions.” (quoted by Joseph, p. 272; italics in original).
Without the context, one may believe that the authors of this article are indeed guilty of circular reasoning. But in fact, these authors too refer to studies on mistaken zygosity:
“Perhaps the key and most controversial aspect of the twin study design is the equal environments assumption (EEA). If the imposed environments of MZ twins are more similar than those of DZ twins, we might wrongly attribute differences in the magnitude of covariance of the phenotype as being due to genes rather than the environment. This might happen, for instance, if parents intentionally or unintentionally provide more similar environments for MZ pairs due to a belief that this is the most appropriate way to rear identical twins. Empirical investigations, however, have shown this to be an unlikely contingency. Because large numbers of parents are mistaken about the zygosity of their twins, it is possible to compare concordances on phenotypes according to perceived and actual zygosity. It transpires that perceived zygosity makes no difference―DZ twins believed by their parents to be MZ are no more similar than those correctly classified as DZ, and vice versa” (P. Sturgis et al., “A Genetic Basis for Social Trust?,” Political Behavior 32, fasc. 2, 2010, p. 221)
Once again, Joseph describes the point of view of behavioral geneticists in a way which does not accurately convey how they reached their conclusion.
Did Joseph write the wrong book?
To some readers of this review, it may seem incredible that an academic monograph, whose stated topic is “the fallacies of twin studies and the failure to discover genes for human behavioral differences, and the relationship between the two” (p. xi), contains not only numerous considerations that are not about these fallacies, e.g. the behaviour of pre-1945 twin researchers, but also some arguments that are falsely presented as relevant, e.g. about between-group heritability, and even outright misrepresentations, e.g. the quote from Martin, Boomsma & Machin (1997) – I am aware that this is a serious accusation, but I do think I have provided sufficient evidence for it. If Joseph cares about a lack of intellectual rigor of twin studies, why does his book contain these inaccuracies and these straw men?
The answer probably lies in what Joseph wrote in his reply to Turkheimer: “even if there are genetic differences among people that contribute to different behavioral outcomes, an emphasis on genetics is still a wrong and even harmful approach” (see above). With this in mind, Joseph’s attitude is understandable: to serve society, it makes sense to take the necessary steps to undermine the credibility of a research area which is perceived as socially dangerous, even if it requires using arguments that many would regard as unethical.
Joseph did not need to write the way he did. Instead of insisting, at the price of accuracy, that there is no acceptable scientific evidence for significant genetic influences on human behavior, it would have been much wiser to write a book about the dangers of emphasizing genetic factors regardless of whether genes influence human behavior or not. There may be good reasons to consider these dangers to be real, as I explained with the case of Hans Eysenck and the tobacco industry (see above). Personally, I would not necessarily agree with the point of view expressed in such a book, but it would have been an honest book and a book which I would have found interesting and probably even enjoyable. It would also have been a very bold and original approach compared to the works of Kamin and other critics.
In the long term, since the internet makes it possible to check quickly who really wrote what, I think Joseph’s approach in The Trouble with Twin Studies will turn out to be a bad decision: once it is established that you resorted to dubious arguments to make your case, it becomes much harder to convince anybody else.
Since The Trouble with Twin Studies has not received much press coverage and, to the best of my knowledge, has been reviewed only once in the scientific literature (Turkheimer’s review), some twin researchers may believe that I am providing attention to a book which should have been utterly ignored. On the contrary, after reading this book, I can confidently state that the researchers who dismiss or ignore The Trouble with Twin Studies will do so at their peril. Joseph is a highly intelligent and determined writer, who has read a lot to be able to make his case and who knew precisely what he had to quote and what he had to leave out. Those who are deeply familiar with twin research are unlikely to be convinced, but many non-alert readers may end up believing that twin studies are indeed an ocean of fallacies and failures, and some might even agree with Joseph’s conclusions on the future of twin studies:
“The human race will survive without twin method behavioral research just as it has survived, or will survive, without alchemy, astrology, channeling, cold fusion, craniometry, creation science, crystal healing, dowsing, faith healing, levitation, mesmerism, numerology, palmistry, phrenology, psychic surgery, and the Ptolemaic astronomy system. There is no doubt that the twin method will join this list. The only question is when.” (p. 252)
Those who regard this view as dubious or dreamful had better keep in mind what Robert Plomin explained in his recent book Blueprint:
“It might seem unbelievable today, but thirty years ago it was dangerous professionally to study the genetic origins of differences in people’s behavior and to write about it in scientific journals” (R. Plomin, Blueprint, Allen Lane, 2018, p. xi).
Twin research is now seen as a respectable research area, but it was not the case thirty or forty years ago, and one should not take for granted that it will be the case in the next decades. Twin researchers need to pay attention to the criticisms of their work and point out the specious arguments that they are able to identify. They also need to close the knowledge gap between the perception of the general public and the accepted results of behavioral genetics, to ensure that the public is able to resist fallacies and to benefit from the enlightening findings of the field (this mission is currently pursued by The Accessible Genetics Consortium). Most importantly, I hope enough twin researchers will take the time to answer misconceptions (e.g., Nathan Comfort’s infamous question: if environment or experiences don’t matter, “what’s the use of living?”), and to warn against the potential misuses of their research findings. If we could live in a world in which most of us accept the influence our genes have on us and deal with what we know factually and responsibly, this would be the ultimate refutation of Joseph’s worldview.
Note: The views expressed in this review do not necessarily reflect the views of the organizations or of the individuals with whom the reviewer is associated.