In his bestselling 2016 book The Gene: An Intimate History, cancer physician and researcher Siddhartha Mukherjee chronicled the “story of the gene” and the history of genetic research and theories, detailing discoveries as well as horrible abuses and crimes committed in the name of genetics and eugenics.1 He weaved his own family history of psychosis and mental disorder into the story, and he believed that heredity was a major cause of these disorders, as well as a major influence on human behavioral differences (variation) in general. There are some interesting areas in the book, including his discussion of twin research as having its origins in eugenics and the German “racial hygiene” movement, an area that I covered in more depth in my 2004 book The Gene Illusion. Previously, Mukherjee had won a Pulitzer Prize for his book Cancer: The Emperor of All Maladies, published in 2011. The Gene: An Intimate History has also received awards, and will be the subject of a three-hour U.S. Public Broadcasting System (PBS) documentary by the famous filmmaker Ken Burns, entitled “Ken Burns Presents The Gene: An Intimate History.” It is set to premiere in April, 2020. (Brief video preview HERE.)
In three online articles I published in 2016, however, I showed that Mukherjee had only superficial knowledge of the main behavioral genetic study he championed—the famous “Minnesota Study of Twins Reared Apart,” or “MISTRA”—and that he got many of the basic facts wrong about this study.2 Turning to schizophrenia twin research, which related directly to Mukherjee’s discussion of his family history, I documented his numerous errors and obvious lack of knowledge in this area.
Due to the acclaim and publicity this book continues to receive, in addition to the upcoming PBS documentary (more below), here I will summarize and add to the main points I made in these 2016 articles. This is especially relevant today given that, despite what the media tells us, the decades-long search at the molecular genetic level for genes that cause differences in behavior, and for genes that cause schizophrenia, depression, and other major psychiatric disorders, has been a bust.
There are two main ways that twins have been used for behavioral research purposes: (1) the “twin method,” which is based on reared-together MZ (monozygotic, identical) and reared-together DZ (dizygotic, fraternal) twins, and (2) “twins reared apart” or “TRA” studies (also known as “separated twin studies”), which study MZ (and sometimes DZ) twins who supposedly were reared apart from each other in different family environments.
Twins Reared Together (The Twin Method). The twin method compares the behavioral resemblance or psychological test-score correlations of MZ versus same-sex DZ twin pairs. MZ pairs share a 100% genetic resemblance, whereas DZ pairs share an average 50% genetic resemblance. Schizophrenia and other psychiatric twin studies are based on comparing MZ versus DZ “concordance rates.” Concordance means that both twins are diagnosed/labeled with the same disorder. Genetic interpretations of the usual twin method finding that MZ pairs behave more similarly than DZ pairs are based on the “equal environment assumption” or “EEA,” where twin researchers must assume that MZ and DZ pairs grow up experiencing roughly equal behavior-shaping environments. Since the 1930s, critics have made a solid case that the EEA is false, and that MZ-DZ twin method comparisons may have recorded nothing other than the more similar treatments and environments, emotional bonds, and identity confusion experienced by reared-together MZ pairs.
Twins Reared Apart (TRA Studies). The six published TRA studies investigated twin pairs who were said to have been reared apart in different families. Such pairs are extremely rare, more so now than in the past. Contemporary TRA researchers claim/assume that any behavioral resemblance (correlation) between reared-apart MZ twin pairs, known as “MZA” pairs, can only be caused by the identical genes they share. By far the most celebrated and widely cited investigation is the Minnesota twin study (MISTRA), carried out by psychologist and behavioral geneticist Thomas J. Bouchard, Jr. and his colleagues, whose most frequently cited publications appeared in the 1980s and 1990s. Mukherjee focused on the most famous MISTRA article, which was published in 1990 in the prestigious journal Science. In this article Bouchard and colleagues claimed/assumed that environmental influences did not contribute to the significantly above-zero MZA IQ and personality correlations they reported, or that if environmental influences did play a role, that they should be counted as genetic influences. Critics, on the other hand, have argued that the MISTRA and the other five TRA studies contained major flaws and biases. (Links to my 2018 critical analysis of the MISTRA: FULL VERSION; ABRIDGED VERSION.)
Mukherjee’s Positions. Based largely on twin method and TRA study findings, Mukherjee concluded that researchers had discovered that sexual preference, temperament, personality, anxiety and most other aspects of behavior “have become progressively encircled by genes,” and that “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” (p. 387). In the “Manifesto” appearing towards the end of the book, he wrote, “In exceptional cases…extraordinary measures, such as genetic selection, or directed genetic interventions, are justified” (p. 482).
Although the genetic basis of schizophrenia is seen as a proven fact in the behavioral sciences and in the public imagination, critics have shown that the studies upon which such beliefs are based are massively flawed, and are based on numerous false assumptions.3 (See my 2017 e-book Schizophrenia and Genetics: The End of an Illusion.) Mukherjee called schizophrenia a “genetic disease” (p. 442), and echoing mainstream positions in psychiatry he saw it as “a complex, polygenic illness, involving multiple variants, multiple genes, and potential environmental or chance triggers” (p. 300).
The following points summarize and bring together my 2016 descriptions of the major errors found in Mukherjee’s arguments in favor of the importance of genetic influences on human behavioral differences, arguments that were based largely on twin studies.
Mukherjee’s Discussion of Twins Reared Apart Studies and the MISTRA
Bouchard Was Not the First Person to Study Reared-Apart Twins
Although Mukherjee implied that in 1979 Bouchard was the first researcher to use reared-apart twins to find “a way out of the impasse” caused by the twin method’s long-running EEA problem (p. 381), individual MZA pairs have been studied since the 1920s and 1930s, and the first of six TRA studies was published in 1937.4
Most Twins Were Not Separated At Birth, and Most Were Only Partially Reared Apart
In a 2016 article he published in The New Yorker, Mukherjee wrote that the MISTRA researchers “gleaned” their MZA sample by “scouring thousands of adoption records and news clips.” In fact, the MISTRA twins were a “collection of cases,” most of whom volunteered to participate in the study after hearing about it through media reports.5 TRA studies based on volunteer twins are plagued by similarity biases because twins had to have known of each other’s existence to be able to participate. Like the famous “Firefighter Pair,” some twins discovered each other because of their behavioral similarities. A much less biased method of obtaining MZA pairs is to locate them through national twin registries, where researchers can identify pairs who had been separated near birth, and who had never met each other or been aware of each other’s existence. The MISTRA had to rely on a similarity-biased volunteer sample because national twin registries did not exist in the United States.
Unlike the authors of the original three TRA studies, who together provided more than 500 pages of case-history information for the 75 MZA pairs they studied, the MISTRA researchers provided very little published case-history information. They have always prohibited independent reviewers from inspecting their raw data, yet they continue to insist that we must accept their claim that genetic factors play a major role in causing behavioral differences among billions of human beings—past, present, and future—on the basis of their interpretation of twin data and case-history information that they have always kept secret.
Mukherjee wrote that the MISTRA MZA pairs were “separated at birth” (p. 381). The MISTRA criteria, however, stated only that twins had “been separated by four years of age,” and that they had “spent their formative years apart.” The mean age at separation was 218 days, and ranged from 0 to 1,644 days (4.5 years).6 As seen in the earlier 1937 Newman et al. and 1962 Shields TRA studies, when growing up, most MZA pairs who participate in volunteer-based TRA studies such as the MISTRA lived nearby to each other, had varying degrees of contact and regular visits with each other, and/or were reunited and lived together for periods of time. As adults, many pairs formed or re-formed a relationship with each other many years before they volunteered to participate in the study. As I and others have documented, the earlier Newman et al. and Shields studies provided abundant evidence showing that most volunteer MZA pairs were only partially reared apart.7
There Are Many More Types of Non-Genetic Influences That Contribute to MZA Behavioral Resemblance
MZA behavioral similarities, according to Mukherjee, “could have nothing to do with nurture; they could only reflect hereditary influences—nature” (p. 381). This statement is completely false. In addition to volunteer biases, and separation that usually is only partial, reared-apart twins share a common prenatal environment. Leading behavioral genetic researchers have recognized that “an adverse prenatal environment can have detrimental effects on later behavioral and health outcomes.”
Postnatally, sex-differences have an important impact on behavior. Males and females are treated differently and are socialized to behave and think differently, meaning that same-sex pairs—which MZAs always are—will behave and think more alike for this reason alone. MZAs share the same ethnicity and skin pigmentation, which, like sex differences, leads to similar intra-pair experiences of oppression or privilege. MZAs also share a striking physical resemblance and a similar level of physical attractiveness.
MZA behavioral similarities are also influenced by cohort effects, which refer to similarities in age-matched people’s behavior, preferences, beliefs, physical condition, and other characteristics that are caused not by heredity, but by experiencing stages of life at the same time in the same historical period and cultural milieu. In a June, 2019 article, I listed over 30 environmental similarities and cohort influences shared by MZA pairs.
Mukherjee wrote that in the MISTRA, there were “frequent [environmental] discordances between two individual twins (one reared in a poor family, another adopted by a wealthy family)” (p. 381). There is no evidence that this occurred “frequently,” and the MISTRA case-history information remains unpublished and under lock and key. This has enabled the researchers, since the early 1980s, to write about and to provide to the media only selected MZA pairs that seemed to support their claims (more below).
The 1990 MISTRA Science Publication
Mukherjee correctly reported that the famous 1990 MISTRA Science article sample consisted of 56 MZA and 30 DZA (reared-apart DZ or fraternal) twin pairs (p. 381). However, although Bouchard and colleagues had designated DZA pairs as the MISTRA control group at the beginning of the study in 1979, in this Science article, which remains the main MISTRA IQ publication, they failed to publish correlations for these 30 DZA pairs.8 Although Mukherjee wrote that the researchers presented “a comprehensive analysis” (p. 381), like most commentators he overlooked the fact that the Science article did not report any DZA results. Bouchard and colleagues suppressed the correlations and other data produced by their 30 control group DZA pairs, most likely because their MZA IQ correlations were not higher than the corresponding DZA correlations at a statistically significant level. (More on this important point HERE.) To this day, the full-sample MISTRA DZA control group IQ correlations, based on three different tests, remain unpublished.9
Mukherjee Did Not Understand How the Researchers Arrived at Their Conclusions in Favor of Genetics
“By comparing separated-at-birth twins against twins brought up in the same family,” Mukherjee wrote, “Bouchard could untwist the effects of genes and environment” (p. 381). Mukherjee got it wrong on this point too, since Bouchard and colleagues did not compare reared-apart MZs versus reared-together MZs for the purpose of determining whether, or how much, genetic factors influence a behavioral trait (such as IQ or personality). For a description and analysis of the problematic methods they did use to make this determination, see my 2018 analysis of the MISTRA.
Stories About Individual Cherry-Picked MZA Pairs Prove Nothing About Genetics
A few months before the appearance of Mukherjee’s book, I wrote an online article about the sensationalized stories of individual MISTRA MZA pairs who supposedly displayed remarkable and “spooky” behavioral similarities. These include the “Jim Twins,” the “Firefighter Twins,” and “Oskar Stöhr and Jack Yufe.” Since 1979, their stories have been told countless times by researchers, textbook authors, journalists, and others, and a reared-apart triplet set was featured in the 2018 movie Three Identical Strangers (see my review of this movie HERE). Insofar as they are cited in favor of genetic influences on behavior, however, these stories constitute of one of the great “scientific” smoke-and-mirrors deceptions of the past half century. For a number of reasons, these selectively reported “cherry-picked” stories prove nothing about genetic influences on behavior, and are promoted mainly to hoodwink the public into supporting genetic determinist claims. Mukherjee (pp. 382-384) jumped on board this long-running train of deception and retold these twins’ stories in support of the MISTRA “findings,” even though their stories are much more appropriate for TMZ.com than they are for publication in supposedly “scientific” books and articles.
Mukherjee’s Discussion of Schizophrenia Twin Research
The Twin Method’s “Conceptual Gridlock” Strangely Does not Apply to Schizophrenia Twin Studies
Behavioral research results do not speak for themselves, and the key question in any twin study is how we should interpret the results. Mukherjee stressed the importance of performing TRA studies because, as he correctly observed, genetic interpretations of twin method results are “intrinsically flawed” because reared-together MZ pairs are treated more similarly by parents, teachers, and others than are reared-together DZ pairs. This led to a “conceptual gridlock” in studies of reared-together MZ pairs, he wrote, because “geneticists knew” that studying such pairs involved the “impossibility of unbraiding the twisted strands of nature and nurture” (p. 380, italics added). In other words, “geneticists knew” that the twin method’s MZ-DZ equal environment assumption (EEA) is false. So far, so good, although it is usually twin research critics, and not “geneticists,” who make this point. But in his discussion (p. 443) of the schizophrenia twin studies—which used the very same twin method—Mukherjee somehow forgot all this and claimed, amazingly, that these studies showed that “it was impossible to deny a genetic cause” of schizophrenia.
Schizophrenia Twin and Genetic Research Did not Begin in the 1970s
“The first clues about the etiology of schizophrenia,” Mukherjee wrote, “came from twin studies. In the 1970s, studies demonstrated a striking degree of concordance among twins” (p. 442). Strangely, Mukherjee cited a 1977 autism twin study publication as the single source of his claim, whose authors did not discuss schizophrenia twin research.10 The “first clue” schizophrenia twin study did not appear “in the 1970s,” but rather in 1928, and ten more were published between 1929 and 1969. And in fact, no original-data schizophrenia twin studies were published in the 1970s.11 Mukherjee (p. 298) said that Irving Gottesman’s schizophrenia twin study was published in 1982. The actual publication year was 1966.12
Mukherjee discussed the “enormous National Academy of Sciences (NAS) study [that] published [twin] data definitively linking schizophrenia to genetic causes,” which used “the twin method pioneered by Galton…” (p. 298). But as we just saw, elsewhere in the book he contradicted this claim when he wrote that “geneticists knew” that the twin method was unable to “unbraid the twisted strands of nature and nurture.” He said that the NAS study found that “identical twins possessed a striking 30 to 40 percent concordance rate for schizophrenia” (p. 298). But there was nothing “striking” about this result, since other researchers had reported similar or higher schizophrenia concordance rates since the 1920s. (Two examples: Rosanoff et al., 1934; Kallmann, 1946.)
The original NAS twin study was published in 1969 by William Pollin and his colleagues, and was based on twins identified through a U.S. Veterans Administration registry.13 The results that Mukherjee cited were taken from a 1983 update/reanalysis of this study by Kenneth Kendler and Dennis Robinette, who obtained additional records that enabled them to increase the original MZ sample of 69 to 164.14 The final 1983 NAS study pairwise concordance rate was 18.3% (30/164), but Kendler and Robinette emphasized 30.9% MZ concordance by using the “probandwise” concordance method, which double-counts the number of concordant pairs. The 1969 NAS study pairwise MZ concordance rate was only 13.8% (11/69), which led Pollin and colleagues to conclude that the “role of the suggested genetic factor appears to be a limited one.”15
“Fleets” of 1980s Schizophrenia Twin Studies That Never Existed
According to Mukherjee, “throughout the 1980s, fleets of twin studies strengthened the case for a genetic cause of schizophrenia” (p. 443). Apart from the above-mentioned 1983 NAS update, however, only one original-data schizophrenia twin study was published in the 1980s, by Koskenvuo and colleagues in 1984. This study is almost never mentioned by schizophrenia twin researchers, or found in textbooks or in schizophrenia genetics review articles, perhaps related to the fact that, similar to Pollin, Koskenvuo found a modest 11% (8/73) MZ concordance rate. (A 2018 Danish study found a similarly modest schizophrenia MZ concordance rate, this time 14.8%, or 12/81 pairs.) In “study upon study,” Mukherjee wrote about these non-existent (and of course unreferenced) “fleets” of 1980s schizophrenia twin studies, “the concordance among identical twins exceeded that of fraternal twins so strikingly that it was impossible to deny a genetic cause” (p. 443).
In a frequently cited 2003 schizophrenia twin study analysis by leading genetic researchers, the only 1980s twin study listed was Kendler and Robinette’s 1983 NAS study, and the only 1970s twin study listed was Pekka Tienari’s final 1975 report based on his 1963 Finnish study.16 Regardless of what Mukherjee claimed, Koskenvuo’s unknown 1984 investigation, and the 1983 NAS study update, were the only original-data schizophrenia twin studies published between 1970 and 1990.
How Mukherjee was able to conjure up multiple “first clue” 1970s schizophrenia twin studies, followed by additional “fleets” of such studies in the 1980s, I haven’t a clue.
The 2020 Ken Burns Documentary
In July, 2018, PBS announced that it would premiere “Ken Burns Presents The Gene: An Intimate History” over three nights in the spring of 2020. According to the press release announcing what is likely to be a widely discussed film, Burns will “use science, social history and personal stories to weave together a historical biography of the human genome while also exploring the stunning breakthroughs in understanding the impact genes play on heredity, disease and behavior.”
In the fall of 2018 I was contacted by Ark Media, a company involved in the project. I was told that they were interested in hearing more about my criticism of the MISTRA. One or more people there had read my 2016 articles about Mukherjee’s book, and in November, 2018 a senior producer interviewed me by telephone. In the course of this conversation it occurred to me that he was more interested in getting background information than he was in my critique of the MISTRA or my exposure of Mukherjee’s careless scholarship in the area of behavioral twin research, and I have not been contacted by the project since then.
“Junk science props up totalitarian regimes,” wrote Mukherjee, “and totalitarian regimes produce junk science” (p. 128). But “democratic” (neoliberal) capitalism needs and produces its own brand of genetic junk science, which is currently supplied in abundance by the fields of behavioral genetics and psychiatric genetics. This helps explain why the elites, including billionaire Bill Gates who loves Mukherjee’s book, look favorably upon the promotion of dubious behavioral genetic claims in general, and the “Minnesota Study of Twins Reared Apart” in particular.
In 2019 I showed that in his 2018 book Blueprint: How DNA Makes Us Who We Are, leading behavioral geneticist Robert Plomin provided a false account of the history of genetic research and theories in psychology.17 Plomin airbrushed from history the eugenic theories and practices of the first half of the 20th century, and even denied the existence of genetic research in psychology prior to the 1960s—the very same troubling history that Mukherjee discussed in his book. But since the major conclusions in both books support the interests of the politically and economically powerful, both are promoted and/or publicized by the corporate media, and books and articles by the critics are ignored or are marginalized. That’s pretty much how things work these days. I hope I am wrong, but I expect Burns’ documentary to continue this trend.
The methods and findings of some of the most famous studies produced by American psychology, such as “The Milgram Experiment,” “The Stanford Prison Experiment,” and “Being Sane in Insane Places” have been called into question in recent years. Unethical and biased research practices in psychology have come to light, in addition to undisclosed conflicts of interest and a non-replication crisis, and there are even cases of long-undetected outright fraud in psychology. Although the EEA and the MISTRA have thus far avoided in-depth journalistic or objective mainstream academic scrutiny (i.e., not performed by twin researchers), such scrutiny, to put it mildly, is long overdue.
Although most of what he wrote about schizophrenia twin research was wrong—just as much of what he wrote about the MISTRA was also wrong—Mukherjee’s status as a Pulitzer Prize winning medical authority contributes to the misinformation and mythology surrounding genetic research in psychiatry, psychology, and the social and behavioral sciences in general. Mukherjee clearly knew little about TRA research, or about the evidence supporting his claim that schizophrenia is a “genetic disease,” yet he thought he knew enough to reach strong conclusions in favor of genetics in what would become an award-winning book that was turned into a Ken Burns documentary, and to even speculate about the need for “genetic selection” in the future. As the saying goes, a little bit of knowledge can be a very dangerous thing.
- Mukherjee, S., (2016), The Gene: An Intimate History, New York, Scribner. ↩
- My three 2016 online articles about Mukherjee’s book were “Reared-Apart Twin Study Mythology: The Latest Contribution” Part One and Part Two, and “Schizophrenia Twin Research as Reported in The Gene: An Intimate History—Getting the Facts Straight.” ↩
- For additional evidence that the twin method’s equal environment assumption does not hold for psychosis and schizophrenia, see Fosse R., Joseph J., & Richardson, K., (2015), A Critical Assessment of the Equal Environment Assumption of the Twin Method for Schizophrenia, Frontiers in Psychiatry, 6:62, 1-10. doi:10.3389/fpsyt.2015.00062 ↩
- Newman et al., (1937), Twins: A Study of Heredity and Environment, Chicago, University of Chicago Press. ↩
- Segal, N. L., (2012), Born Together—Reared Apart: The Landmark Minnesota Twin Study, Cambridge, MA: Harvard University Press, pp. 35-36. ↩
- Segal, 2012, pp. 37-38. ↩
- Joseph, J., (2015), The Trouble with Twin Studies: A Reassessment of Twin Research in the Social and Behavioral Sciences, New York: Routledge, Chapter 2, Tables 2.1 and 2.2. ↩
- Segal, 2012. MISTRA researcher Nancy 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. ↩
- The abstract (summary) of the 1990 MISTRA Science article stated that the study tested “more than 100 sets of reared-apart twins or triplets,” but the main article reported correlations for less than 50 sets. It is simply amazing that this confusing (and confused) publication, which was based on the obviously false assumption that MZA pairs experienced “no environmental similarity,” and where the DZA control group results were suppressed, made it through the supposedly rigorous peer-review process of one of the world’s leading scientific journals. ↩
- The autism twin study publication cited by Mukherjee was Folstein, S. E., & Rutter, M., (1977), Infantile Autism: A Genetic Study of 21 Twin Pairs, Journal of Child Psychology and Psychiatry, 18, 297-321. ↩
- In 1973, Margit Fischer published the final report of Fischer and colleagues’ 1969 Danish schizophrenia twin study. ↩
- Gottesman & Shields’ 1966 schizophrenia twin study was featured, and described in much more detail, in their 1972 book Schizophrenia and Genetics: A Twin Study Vantage Point. ↩
- The original NAS schizophrenia twin study was Pollin et al., (1969), Psychopathology in 15,909 Pairs of Veteran Twins: Evidence for a Genetic Factor in the Pathogenesis of Schizophrenia and its Relative Absence in Psychoneurosis, American Journal of Psychiatry, 126, 597-610. ↩
- Kendler, K. S., & Robinette, C. D., (1983), Schizophrenia in the National Academy of Sciences-National Research Council Twin Registry: A 16-year Update, American Journal of Psychiatry, 140, 1551-1563. ↩
- Pollin et al., 1969, p. 608. ↩
- Sullivan et al., (2003), Schizophrenia as a Complex Trait: Evidence from a Meta-Analysis of Twin Studies, Archives of General Psychiatry, 60, 1187-1192, p. 1188. ↩
- Joseph, J., (2019, September 3rd), A Blueprint for Genetic Determinism, (Web blog post, Mad in America “The Gene Illusion”), https://www.madinamerica.com/2019/09/blueprint-genetic-determinism/ ↩
Mad in America hosts blogs by a diverse group of writers. These posts are designed to serve as a public forum for a discussion—broadly speaking—of psychiatry and its treatments. The opinions expressed are the writers’ own.