In Part One, we saw that the “Minnesota Study of Twins Reared Apart” (MISTRA) of Thomas J. Bouchard, Jr. and colleagues was a behavioral genetic twin study conducted between 1979 and 2000, with media reports and academic papers based on it continuing to appear into the current period. In my recent book The Trouble with Twin Studies: A Reassessment of Twin Research in the Social and Behavioral Sciences and elsewhere, I showed that the MISTRA and other “TRA” (twins reared apart) studies (also known as “separated twin studies”) are enormously flawed on several critical dimensions, and that accompanying media stories of individual reared-apart pairs provide no evidence that genetic factors influence behavioral similarities and differences in the general (non-twin) population. Reared-apart MZ (identical) pairs, who are said to share a 100% genetic similarity, are also known as “MZA” pairs, or “monozygotic twins reared apart.” Reared-apart DZ (fraternal) pairs, who are said to share an average 50% genetic similarity, are also known as “DZA” pairs, or “dizygotic twins reared apart.” (A summary of the TRA study critique can be found here. A brief critique of reared-together twin study assumptions can be found here.)
The public’s knowledge of the MISTRA is based largely on television reports, numerous academic texts and popular books, and countless articles appearing since 1979. A major theme of these “mistraphile” reports and publications has been that the MISTRA researchers/scientists discovered, often to their astonishment, that genetic factors play a predominant role in most aspects of human behavior and abilities (including IQ and personality).
The latest endorsement of MISTRA claims comes from cancer physician and researcher Siddhartha Mukherjee, who published an article in the May 2, 2016 edition of The New Yorker entitled “Same but Different: How Epigenetics can Blur the Line between Nature and Nurture.”1 In Part One I showed that this article’s MISTRA-related paragraphs were wrong on several accounts, and helped perpetuate myths about behavioral genetic and psychiatric genetic research in general. Here I focus on Mukherjee’s very favorable description of the MISTRA in his 2016 book, The Gene: An Intimate History, currently a “#1 Best Seller” on Amazon.com. Mukherjee is an influential author due to his medical credentials, accessible writing style, promotion by the mainstream media, and a previous Pulitzer Prize winning book. And yet, he appears to be only nominally familiar with the MISTRA, based mainly on one of its many publications. In his book he arrived at the following conclusions about behavioral genetic research, based in good part on the MISTRA claims:
“Gender. Sexual preference. Temperament. Personality. Impulsivity. Anxiety. Choice. One by one, the most mystical realms of human experience have become progressively encircled by genes. Aspects of behavior relegated largely or even exclusively to cultures, choices, and environments or to the unique constructions of self and identity, have turned out to be surprisingly influenced by genes.”2
I will not review The Gene: An Intimate History as a whole here, but instead I will highlight areas where Mukherjee was wrong about the basic facts of the MISTRA. Some overlap with points made in Part One is unavoidable.
1. Mukherjee implied that Bouchard was the first to use reared-apart twins to “find a way out of the impasse” of environmental confounds in twin method reared-together MZ-DZ comparisons. In fact, MZA pairs have been studied since the 1920s, and the first systematic study was published in 1937.3 Behavioral genetic and psychiatric genetic researchers have also claimed for decades that (non-twin) adoption studies offer a way out of this “impasse,” yet Mukherjee paid little attention to this major area of research.
2. Referring specifically to the frequently cited MISTRA 1990 Science magazine publication,4 Mukherjee wrote that the MISTRA pairs were “separated at birth” and grew up in “often radically different” environments, with “frequent [SES] discordances between two individual twins (one reared in a poor family, another adopted by a wealthy family).” However, Bouchard and colleagues made no such claims in their 1990 Science article, which was the only original MISTRA publication cited by Mukherjee. There is no published evidence that most MISTRA MZA pairs grew up in environments that differed greatly, but there is much published evidence from earlier TRA studies suggesting that most did not.5 Most pairs in these earlier TRA studies were only partially reared apart, and for a number of reasons it is unlikely that the Minnesota MZA pairs were any more “separated” than these pairs were.
3. MZA behavioral similarities, according to Mukherjee, “could have nothing to do with nurture; they could only reflect hereditary influences—nature.” This statement is completely false. MZA pairs share a common prenatal environment, are the same age, are the same sex, and share a striking physical resemblance. The last three certainly increase behavioral similarity, and the MISTRA researchers themselves recognized that “for most psychological, physiological, and medical variables there are substantial age and sex effects” (italics added).7 MZAs share many other environmental similarities and non-genetic cohort effects. The “cohort effect” concept refers to similarities in age-matched people’s behavior, preferences, beliefs, physical condition, and other characteristics that are caused by experiencing stages of life at the same time in the same historical period and cultural milieu. In addition to sharing a common prenatal environment and experiencing similar postnatal healthcare, most reunited MZA pairs grew up sharing at least ten different behavior-molding cultural/environmental influences in common: national, regional, language, ethnic, religious, physical resemblance, adoptee status, economic class, birth cohort, and gender cohort
4. According to Mukherjee, “Bouchard’s staff was repeatedly struck by the similarity between the twins.” This has been a common theme in mistraphile publications over the decades—that the researchers were “struck” by their “findings.” The word “struck” suggests that the researchers were surprised and impressed by MZA behavioral similarity. It implies that they had no genetic biases as they observed and evaluated twins, and that they might even have had environmental biases. Subtle linguistic maneuvers of this type since 1979 have helped the MISTRA researchers’ claims achieve greater legitimacy. Although popular writers usually portray the Minnesota and other behavioral genetic twin researchers as objective scientists with little interest in the social and political implications of their findings, most were ideologically committed to the position that genetic influences on behavior are pervasive and important, and then interpreted their twin data in favor of genetics to promote this belief. In a 2009 interview published in Science, Bouchard discussed how in the early 1970s he was won over to the behavioral genetic perspective and Arthur Jensen’s hereditarian theories of intelligence.8
Human genetic research results do not speak for themselves, and are clearly subject to the genetic biases of the researchers who report and interpret them. MZA behavioral similarities could be, but rarely are, interpreted as showing the power of non-familial cohort effects and other shared environmental factors in shaping human behavior (see Point 3). In response to a statement by MISTRA researcher Nancy Segal, in her 2012 book Born Together—Reared Apart: The Landmark Minnesota Twin Study, that “science rests on data, not dialogue,”9 critic Michael Rossi replied,
“Whatever Bouchard and Segal’s protestations, it’s stories about data—interpretations of observation, assumptions underlying test design, explanations of statistical correlation—on which the Mistra scientists based their findings of ‘genetic’ behaviours given that not a single gene was actually examined in the study.”10
5. Mukherjee correctly reported that the famous 1990 MISTRA Science article sample consisted of 56 MZA pairs, and 30 DZA pairs. However, although the researchers had from the beginning designated DZA pairs as the MISTRA control group,11 in this Science article Bouchard and colleagues failed to publish correlations for these 30 DZA pairs. Although Mukherjee claimed that the researchers had presented “a comprehensive analysis,” and “had collected data from fifty-six reared-apart identical twins [MZAs] and thirty reared-apart fraternal twins [DZAs],” he did not mention that in their Science article the researchers failed to report correlations and other “collected data” produced by their own designated DZA control group. Apart from a few isolated commentators, no one ever criticizes this crucial omission.12 Bouchard and colleagues’ arbitrary decision to not publish their DZA correlations was based on their unconvincing claim that they could not do so “due to space limitations and the smaller size of the DZA sample (30 sets).”13
In their Science article, the MISTRA researchers wrote that “the study of IQ is paradigmatic of human behavioral genetic research,” and that,
“Monozygotic and dizygotic twins who were separated early in life and reared apart (MZA and DZA twin pairs) are a fascinating experiment of nature. They also provide the simplest and most powerful method for disentangling the influence of environmental and genetic factors on human characteristics.”14
To this day, however, the MISTRA researchers have failed to publish and analyze their “simple and powerful” full-sample DZA IQ correlations, even though they published such correlations for personality and most other MISTRA-studied non-IQ behavioral characteristics (traits) in the 1980s and 1990s. The most likely reason, as I showed in Chapter 6 of The Trouble with Twin Studies, is that the MISTRA’s full-sample MZA IQ correlations apparently are not higher than the unpublished full-sample control group DZA IQ correlations at a statically significant level. This finding invalidates Bouchard and colleagues’ conclusion that IQ “is strongly affected by genetic factors.”15
6. Mukherjee wrote that the MISTRA MZA pairs’ “physical and racial environments were…broadly different.” It is not completely clear what he meant by this, but I am unaware of the Minnesota researchers ever having made this claim.
7. According to Mukherjee, the MISTRA researchers determined that genetic factors are important by comparing MZA psychological test score correlations (IQ, personality) versus those of MZTs (reared-together monozygotic twins). “By comparing separated-at-birth twins against twins brought up in the same family,” he wrote, “Bouchard could untwist the effects of genes and environment.” However, Bouchard and colleagues did not use MZA versus MZT comparisons to assess the role of genetic influences.16 They concluded only that MZA-MZT comparisons showed that “common rearing enhances familial resemblance during adulthood only slightly and on relatively few behavioral dimensions.”17
8. On what basis, then, did the Minnesota researchers conclude that their results should be interpreted in favor of genetics? The answer is that they simply assumed that all MZA behavioral resemblance is caused by genetics, and that an above-zero MZA correlation “directly estimated” the “heritability” of the characteristic (trait). The researchers’ main assumption, as stated in their 1990 Science publication, was that MZAs experience “no environmental similarity.”18 As seen in Point 3 above, this assumption is utterly false. Mukherjee helped legitimize this obvious folly with his false claim that MZA behavioral similarity can “only reflect hereditary influences.” The MISTRA researchers recognized that some of the assumptions upon which they based their conclusions were “likely not to hold,”19 but claimed that “several combinations of violations of assumptions can act to offset each other.”20 Such unscientific genetically-biased speculation—that numerous false assumptions magically all cancel each other out in favor of genetics— allowed them to conclude what they already strongly believed: that above-zero MZA correlations, worked into “biometric model fitting” statistical procedures, showed that genetic factors have a strong and pervasive influence on most areas of human behavior.
9. The researchers also bypassed a critical step in the process of assessing the potential role of genetic factors. According to Segal,
“The simple comparison of the MZ (or MZA) and DZ (or DZA) intraclass correlations is an important first step in behavioral-genetic analysis because this demonstrates whether or not there is genetic influence on the trait” (italics added).21
Here, Segal recognized that the MZA correlation must be significantly higher than the DZA correlation as a preliminary “first step” in determining whether or not genes have any influence on traits (characteristics) such as IQ and personality. Because MZA pairs are more similar to each other genetically than are DZA pairs (100% versus 50%), a mean (average) MZA behavioral trait correlation not higher than the corresponding DZA correlation at a statistically significant level suggests that non-genetic factors alone are responsible for raising both correlations above zero, since MZAs’ greater genetic resemblance did not lead to their greater behavioral resemblance. And yet, after reviewing their data, Segal, Bouchard and their colleagues decided against making this necessary “important first step” determination of “whether or not there is genetic influence on the trait.”
10. “Religiosity and faith,” Mukherjee wrote, “were also strikingly concordant; twins were either both faithful or both nonreligious.” He most likely was referring to two MZA “religiosity scale” correlations reported in the MISTRA Science article. A MISTRA study of religious interests, attitudes, and values by Waller, Bouchard, and colleagues also appeared in 1990.22 Here the researchers assessed religious characteristics through the use of religiosity scale correlations, which were factored into the researchers’ model fitting procedures. MZA correlations on various tests ranged from .39 to .55. Assuming that the tests are valid, this indicates only that twins’ levels of religiousness varied together to a certain degree—for any number of possible reasons—but it does not mean that “twins were either both faithful or both nonreligious.” It is noteworthy that the DZA correlation was negative for two of the five scales, a result that runs counter to both genetic and environmental predictions.
Like other MISTRA publications, Waller and colleagues’ study was based on several false or questionable assumptions, the most prominent of which was the false assumption that “MZA and DZA correlations are due entirely to genetic factors.” They also assumed that genetic and environmental factors do not interact, even though, as several behavioral genetic researchers now recognize, genes and environments do interact.23 Mukherjee also overlooked obvious real-world examples showing that religiousness is a product of peoples’ social and political environments. To cite two of countless such examples, we can compare the level of religiousness in the Soviet Union in the 1950s, versus the same level currently found in the former Soviet republics. Or we could compare levels of religiousness currently found in North Korea and South Korea. But Mukherjee, like the behavioral geneticists whose work he champions, largely ignores real world examples and makes obviously false claims about genes and environments on the basis of indirect inferences from misinterpreted twin study results.
11. Like most authors over the past four decades, Mukherjee discussed well-publicized pairs such as the “Jim Twins.” These selectively reported (“cherry picked”) stories, however, prove nothing about genetics. According to Wikipedia, cherry picking is “the act of pointing to individual cases or data that seem to confirm a particular position, while ignoring a significant portion of related cases or data that may contradict that position.” Twins also had incentives to invent various “spooky” stories they told the researchers, and to exaggerate their degree of separation and behavioral similarity.
12. Mukherjee discussed the selectively reported MISTRA pair Daphne Goodship and Barbara Herbert from London. By the twins’ account, one twin was “the daughter of a lower-middle-class municipal gardener,” while the other was “the daughter of a prominent upper-class metallurgist.” Commenting on the twins’ different social class backgrounds, Mukherjee wrote, “Due to the rigidity of the class structure in 1950s England,” these twins “might as well have been brought up on different planets.” He thereby claimed that social class differences between a pair of London-dwelling female twins, growing up in the same era, were akin to having “been brought up on different planets.” Similarly misleading hyperbolic accounts of selected twins’ environmental differences are found throughout the mistraphile literature. One example is an early 1980s account of a male MISTRA MZA pair from Louisiana, who reportedly shared several behavioral similarities despite being “divided physically” by the apparently unbridgeable “mighty Mississippi” river!24
13. Although logic dictates that the MISTRA researchers’ claims in the 1980s and 1990s should have been followed by the discovery of genes for behavior at the molecular genetic level, decades of attempts to identify “genes for IQ” or “genes for personality” have failed to bear fruit. In 2014, Bouchard recognized that the results of attempts to uncover genes for intelligence (IQ) “have been dismal in comparison with expectation.”25 The following year, behavioral geneticist Eric Turkheimer conceded that “scientists have not identified a single gene that would meet any reasonable standard as a ‘gene for’ schizophrenia, intelligence, depression, or extraversion.”26
As seen here and in Part One, Mukherjee’s descriptions of the MISTRA in his New Yorker article and in The Gene: An Intimate History were inadequately researched, leading to many false or unsubstantiated statements about the study and its MZA sample. His writings help perpetuate myths about behavioral genetic and psychiatric genetic research in general. Unfortunately, Mukherjee’s account of the MISTRA is just the latest in a long series of similar publications that have appeared since 1979, where the works of critics are usually unread, ignored, dismissed, or distorted, and the false claims of supposedly unbiased apolitical researchers are legitimized. In the process the general public, and even sizable chunks of academia, have been massively misled.
* * * * *
- Mukherjee, S., (2016, May 2), Same but Different: How Epigenetics Can Blur the Line between Nature and Nurture, The New Yorker.
- Mukherjee, S., (2016), The Gene: An Intimate History, New York, Scribner, p. 387. Unless otherwise noted, all references to and quotations from this book are based on pages 380-387.
- Newman et al., (1937), Twins: A Study of Heredity and Environment, Chicago, University of Chicago Press. For a critical review of these earlier TRA studies, see Joseph, J., (2015), The Trouble with Twin Studies: A Reassessment of Twin Research in the Social and Behavioral Sciences, New York: Routledge, Chapters 2 and 3.
- Bouchard et al., (1990), Sources of Human Psychological Differences: The Minnesota Study of Twins Reared Apart, Science, 250, 223-228.
- Joseph, 2015, Chapter 2. See also Kamin, L. J., (1974), The Science and Politics of I.Q., Potomac, MD: Erlbaum; Taylor, H. F., (1980), The IQ Game: A Methodological Inquiry into the Heredity-Environment Controversy, New Brunswick, NJ: Rutgers University Press.
- Joseph, 2015, Chapter 2.
- McGue, M., & Bouchard, T. J., Jr., (1984), Adjustment of Twin Data for the Effects of Age and Sex, Behavior Genetics, 14, 325-343.
- Holden, C., (2009), Behavioral Geneticist Celebrates Twins, Scorns PC Science, Science, 325, 27.
- Segal, N. L., (2012), Born Together—Reared Apart: The Landmark Minnesota Twin Study, Cambridge, MA: Harvard University Press, p. 305.
- Rossi, M., (2013), Consider Jack and Oskar, [Review of the book Born Together—Reared Apart: The Landmark Minnesota Twin Study, by N. L. Segal], London Review of Books, 35 (3), 346-349.
- Segal, 2012. Segal wrote on page 12 that “Bouchard’s decision to use DZA twins as controls was made in a very early memo to the ‘Twin Research Team.’” See also Segal’s Footnote 69 on p. 343.
- Two commentators who did criticize this omission were Kamin, L. J., & Goldberger, A. S., (2002), Twin Studies in Behavioral Research: A Skeptical View, Theoretical Population Biology, 61, 83-95.
- Bouchard et al., 1990, p. 223.
- Bouchard et al., 1990, pp. 223-224.
- Joseph, 2015, Chapter 6. The quotation is from Bouchard et al., 1990, p. 227.
- Like Mukherjee, in my own publications published through 2010 I committed the error of writing that the MISTRA researchers based their conclusions about genetics on MZA versus MZT comparisons.
- Bouchard et al., 1990, p. 227.
- Bouchard et al., 1990, p. 224.
- McGue, M., & Bouchard, T. J., Jr., (1989), “Genetic and Environmental Determinants of Information Processing and Special Mental Abilities: A Twin Analysis,” in R. Sternberg (Ed.), Advances in the Psychology of Human Intelligence (Vol. 5, pp. 7-45), Hillsdale, NJ: Erlbaum.
- Johnson et al., (2007), Genetic and Environmental Influences on the Verbal-Perceptual-Image Rotation (VPR) Model of the Structure of Mental Abilities in the Minnesota Study of Twins Reared Apart, Intelligence, 35, 542-562, pp. 548–549.
- Segal, 2012, p. 62.
- Waller et al., (1990), Genetic and Environmental Influences on Religious Interests, Attitudes, and Values: A Study of Twins Reared Apart and Together, Psychological Science, 1, 138-142.
- Waller et al., 1990, p. 139. For an example of a behavioral genetic researcher recognizing the impact of gene-environment interaction on the field’s statistical models, see Johnson, W., (2010), Understanding the Genetics of Intelligence: Can Height Help? Can Corn Oil?, Current Directions in Psychological Science, 19, 177-182.
- Cassill, K., (1982), Twins: Nature’s Amazing Mystery, New York: Atheneum, p. 183.
- Bouchard, T. J., Jr., (2014), Genes, Evolution and Intelligence, Behavior Genetics, 44, 549-577.
- Turkheimer, E., (2015), Arsonists at the Cathedral, PsycCRITIQUES, 60 (40), 1-4. DOI: http://dx.doi.org/10.1037/a0039763. For my response to Turkheimer, see my November, 2, 2015 Mad in America blog posting.