Chapter 4
Irene and Maureen “were always much alike, short and thin with dark eyes. And dark straight hair. They were frequently mistaken for one another, and at one time one twin deputised for the other at the factory where she worked without anyone noticing. They always remained intimate and never troubled to make separate friends. Many ‘telepathic’ experiences are reported; if one fell down at one spot the other would too, and hurt the same knee. Maureen is said to have had pains when Irene had her baby… . At the present time both twins are attending a psychiatric clinic where observation has shown that in both symptoms are much worse just before the periods… . The first remark [in an interview], on being introduced by their mother, was: ‘Yes, we’re the twins, and we’re still ill. It’s a mysterious illness, ours.’ They seemed to share one illness between them.”
—A 37-year-old British reared-together identical (MZ) twin pair, both diagnosed with schizophrenia, as described by twin researcher Eliot Slater in his 1953 schizophrenia twin study151
“Twin studies…provide the bulk of the evidence for the widespread influence of genetics on behavioral traits.”
—Behavioral genetic researchers Robert Plomin and colleagues, in the 2013 edition of Behavior Genetics152
The Twin Method and Its Assumptions
Due to the limitations of family studies, and the spectacular and unexpected failure to produce genes that cause schizophrenia and psychosis, psychiatry and psychiatric genetics have fallen back on emphasizing previous classical twin methodcomparisons between reared-together MZ (monozygotic, identical) and reared-together same-sex DZ (dizygotic, fraternal) twin pairs. MZ pairs are said to share 100% of their segregating genes, whereas (like ordinary siblings) same-sex DZ pairs are said to share an average 50% of their segregating genes (see Figure 1). When both members of a twin pair are diagnosed with schizophrenia, the pair is concordant for the condition; when one twin is diagnosed but the other is not, they are discordant for schizophrenia. Twin pair “concordance rates” are presented as a percentage ranging from 0% to 100%. For example, after identifying 100 MZ pairs in which one twin is diagnosed with schizophrenia, if in 25 of these pairs the second twin is also diagnosed with schizophrenia, using the “pairwise method” concordance would be 25%. If a study finds that MZ pairs are significantly more concordant than DZ pairs for schizophrenia, twin researchers conclude that the difference is caused by genetics factors.
It should be emphasized that psychiatric twin studies are based on comparisons between reared-together MZ and DZ pairs, with each pair growing up together in the same family home. Other than a handful of individual cases of MZ pairs that have come to the attention of researchers since the 1920s, there have been no psychiatric studies of reared-apart twin pairs. The only possible exception was a 1990 Minnesota Study of Twins Reared Apart (MISTRA) study of substance abuse/dependence and antisocial personality.153
Social and behavioral science researchers often study people who grew up in environments that these researchers did not design, control, or observe. They must therefore make assumptions about the nature of these environments, and the conclusions they reach usually depend on the validity (truth) of these assumptions. An assumption is something taken for granted or accepted as true without proof. The project or investigation then treats it, and researchers arrive at conclusions, as if it were true. Whether an assumption is true or false can completely change the findings of a study—for example, a finding that the behavior or disorder is caused or influenced by genetic factors, or a finding that it is caused by non-genetic factors.
As seen in Figure 1, genetic interpretations of twin method results are based on several key assumptions. I will focus here on by far the most controversial and important one, which has always been twin researchers’ MZ-DZ “equal environment assumption,” also known as the “EEA” (see the bolded Assumption #6 in Figure 1). According to this assumption, MZ and DZ pairs grow up experiencing roughly equal environments, and the only factor distinguishing them is their differing degree of genetic relationship to each other (100% versus 50%). As one group of twin researchers put it, the EEA “is crucial to everything that follows from twin research.”154
Twin researchers’ belief that the EEA is valid allows them to argue that genetic factors explain the usual finding that MZ pairs behave more similarly (or correlate higher on psychological tests) than do same-sex DZ pairs. Twin correlations or concordance rates are then factored into model fitting procedures, which lead to the previously discussed controversial practice of producing numerical heritability estimates (0%-100%). The influences of “shared” and “unshared” environmental factors are calculated from what is left over from the heritability estimate.
Schizophrenia twin research dates back to 1928, when Luxenburger published his German study.155 The first American study was published by Aaron Rosanoff and colleagues in 1934.156 In the 16 schizophrenia twin studies published between 1928 and 1999, the pooled MZ concordance rate was about 40%, while the pooled DZ rate was about 8%. These figures can be compared with the general population rate of 1% or less. Looking at the 10 methodologically superior studies published between 1963 and 1999, the pooled rates fall to 23% MZ versus 5% DZ, meaning that in these studies the MZ co-twin of a person diagnosed with schizophrenia is not so diagnosed about 75% of the time.157
In 1946 Kallmann published a large, frequently cited, and controversial schizophrenia twin study, where he claimed to have investigated, made diagnoses, and determined twin zygosity (whether a pair is MZ or DZ) in the families of 691 twin pairs identified from New York state hospital records. His age-corrected concordance rates, based on 174 MZ and 296 DZ pairs, were 85.8% MZ and 14.7% DZ, which are the highest rates ever reported. Kallmann concluded that the predisposition to schizophrenia “depends on the presence of a specific genetic factor which is probably recessive and autosomal.”158
Nearly four decades later, Markku Koskenvuo and colleagues published their methodologically superior 1984 Finnish study, which was based on a large records-based sample consisting of 73 MZ and 225 same-sex DZ pairs. They found a modest 11% MZ concordance rate (8/73). DZ concordance was only 1.8% (4/225), which is not significantly greater than the expected population rate.159 Koskenvuo and colleagues concluded that their results “are in accordance with a hypothesis that postulates great environmental influence (high proportion of discordant pairs) with apparent genetic liability (high difference in MZ and DZ concordance rates).” They also recognized that the “greater environmental similarity between MZ than DZ twin partners can…bias conclusions” in favor of genetics. Because both the existence and results of this study are almost never reported in textbooks, or in the publications of authoritative schizophrenia genetics experts, this low concordance rate study by Koskenvuo and colleagues is largely unknown.
Mainstream textbooks and academic publications often report that pooled MZ concordance is about 50%, with the accompanying 80% heritability estimate based on twin studies. This 80% figure is then circulated widely, giving the misleading impression that schizophrenia is largely genetically determined.
The MZ-DZ Equal Environment Assumption:
The Achilles’ Heel of the Twin Method
There are several important methodological and theoretical problems found in schizophrenia twin research. These have included:
· The lack of an adequate and consistent definition of schizophrenia
· The questionable validity of the “schizophrenia” concept
· The questionable reliability of a schizophrenia diagnosis
· The use of non-blinded diagnoses
· That MZ pairs experience more similar prenatal environments than experienced by DZ pairs
· The use of diagnoses that were made (or were not made) on the basis of hearsay or sketchy information about unavailable or deceased twins
· The use of inaccurate methods of zygosity determination (whether a pair is MZ or DZ)
· The unnecessary, and potentially biasing, use of age-correction formulas
· The use of non-representative or small samples, or sample populations that were biased in favor of concordance (for example, psychiatric hospital populations)
· Inadequate descriptions of the researchers’ methods
· The use of the “probandwise” twin concordance method instead of the “pairwise” method, which leads to inflated concordance rates
· The inflation of concordance rates due to questionable reanalysis practices by genetically oriented researchers not involved in the original studies, which then become the inflated concordance rates reported by the authors of textbooks and other secondary sources
· Investigator bias in favor of genetic conclusions
Although these factors have undoubtedly inflated true concordance rates, especially in the older studies performed between 1928 and 1961, there is no doubt that the pooled MZ concordance rate is significantly higher than the pooled DZ rate for schizophrenia and most other psychiatric disorders and behavioral characteristics. The key question has always been: “Why?”
We have seen that the twin method’s crucial MZ-DZ equal environment assumption (EEA) states that reared-together MZ and DZ pairs grow up experiencing roughly the same environments. From the inception of the twin method in 1924, until the mid-1960s, the EEA was defined as the simple statement that MZ and same-sex DZ pairs experience roughly equal environments. To cite one example, in their 1954 textbook Human Heredity, genetic researchers James Neel and William Schull defined the EEA as follows:
“In the comparison of MZ with DZ twins the assumption is made that the individual members of a twin-pair enjoy comparable or equivalent environments.”160
This 1924-1960s original definition of the EEA is false, since MZs clearly grow up experiencing much more similar environments than DZs.161 Most studies show that MZ pairs are treated much more similarly, and experience more similar environments, than DZ pairs.162 In addition, MZ pairs experience much higher levels of identity confusion, attachment, and emotional closeness than experienced by DZ pairs, which will (presumably) lead to greater MZ behavioral resemblance.163 For example, schizophrenia twin researcher Einar Kringlen performed a “global evaluation of twin-closeness” based on 117 Norwegian pairs, and found that 65% of the MZ pairs had an “extremely strong level of closeness,” which was true for only 17% of the same-sex DZ pairs. Fully 90% of Kringlen’s MZ pairs had experienced “identity confusion in childhood,” which was experienced by only 10% of the DZ pairs.164 To the best of my knowledge, Kringlen’s findings have never been cited or discussed in the works of leading schizophrenia experts, or in any psychiatry or psychiatric genetics publication.
In light of the overwhelming evidence that MZ pairs experience much more similar environments than experienced by DZ pairs, twin researchers of the 1960s and 1970s were confronted with two main options: (1) abandon the twin method, including all previous conclusions in favor of genetics, because the EEA is false; or (2) redefine the EEA in an attempt to keep the twin method alive. Twin researchers should have chosen option #1, but they instead chose option #2.
Two Defenses of the EEA That Fail
In redefining the EEA, twin researchers have claimed since the 1960s that the twin method and the assumption are valid on the basis of two main arguments, which I have called “Argument A” and “Argument B.” Supporters of Argument A concede the point that MZ pairs experience more similar environments than experienced by DZ pairs, but maintain that the EEA is valid because MZ pairs “create” or “elicit” more similar environments for themselves because they are more similar genetically. Supporters of Argument B also concede the point that MZ pairs experience more similar environments than DZ pairs experience, but maintain that the EEA remains valid unless critics are able to identify “trait-relevant” aspects of the environment that MZ pairs experience to a greater degree than DZ pairs. Many twin researchers use both arguments in support of behavioral and psychiatric twin studies. An example of each argument is seen below:
ARGUMENT A
Twin researchers’ “twins create their more similar environments” argument
Example:
“It is important to note that if MZ twins are treated more alike than DZ twins, it is most likely associated with their genetically based behavioral similarities.”
—Twin researchers Nancy Segal and Wendy Johnson, 2009165
ARGUMENT B
Twin researchers’ “trait-relevant” argument
Example/Definition:
“The traditional twin method, as well as more recent biometrical models for twin analysis, are predicated on the equal-environment assumption (EEA)—that monozygotic (MZ) and dizygotic (DZ) twins are equally correlated for their exposure to environmental influences that are of etiologic relevance to the trait under study” (italics added).
—Twin researchers Kenneth Kendler and colleagues, 1993166
As my colleagues and I showed elsewhere, and as I showed in detail in Chapters 7 and 8 of The Trouble with Twin Studies, these arguments are illogical and do not hold up to critical examination.167 Argument A is based on circular reasoning, since the conclusion that MZ pairs behave more similarly because they are more similar genetically is based on a premise stating the very same thing. Twin researchers invoking Argument A, therefore, take for granted that genetic factors influence observed behavioral differences among humans, which is the very issue under dispute. In his 2015 review of The Trouble with Twin Studies, behavioral geneticist Eric Turkheimer did not dispute my contention that EEA defenses are based on illogical circular reasoning, nor has anyone else to my knowledge.168
In addition to his support for Argument B, Kendler wrote often that the EEA and the twin method are valid on the basis of the Argument A position that “the environmental similarly of MZ twins is the result and not the cause of their behavioral similarity” (italics in original).169 Apart from the circularity of this argument, the argument fails because even if true, higher MZ concordance for psychiatric disorders such as schizophrenia could still be completely caused by environmental factors. For the moment, let’s assume as a hypothetical example (1) that Kendler’s Argument A position is correct, (2) that a high level of mercury in the brain at a certain developmental stage causes schizophrenia, and (3) that a twin study finds a significantly higher schizophrenia concordance rate for MZ as opposed to DZ pairs. In this hypothetical scenario MZ pairs would be more similarly exposed to the environmental toxin mercury because, due to their inherited behavioral similarity and similar physical appearance, they spend much more time together, attend school together and have common friends and so on, and elicit more similar treatment by their parents and others. However, it would be mistaken to conclude, as Kendler and other Argument A supporters likely would, that mercury exposure should be counted as a genetic effect because “MZ twins create their more similar environments.” Twins developing schizophrenia because they are more similarly exposed to toxic environmental agents or treatments—even if this common exposure is caused by their inherited behavioral similarity—is an environmental effect having nothing to do with “genes for schizophrenia.” To assess the EEA’s validity, it is necessary only to determine whether MZ and same-sex DZ childhood and adult environments are roughly equal. Once it is determined that these environments are different, we must conclude that genetic interpretations of twin method MZ-DZ comparisons are invalid because they are confounded by environmental factors. Just like in family studies, the question of why these environments are different is irrelevant.
Supporters of Argument B attempt to reverse the burden of proof onto critics to show that MZ and DZ pairs differ in terms of their exposure to “trait-relevant” factors. According to Kendler, Tsuang, and colleagues, “It would seem that the burden of proof rests with critics of the twin method to demonstrate that ‘trait-relevant’ environmental factors are more similar for identical than same-sex fraternal twins.”170 As a group of psychologists dedicated to distinguishing good science from bad science pointed out in another context, however, “The burden of proof in science rests invariably on the individual making a claim, not on the critic. Proponents of pseudoscience frequently neglect this principle and instead demand that skeptics demonstrate beyond a reasonable doubt that a claim…is false.”171 Therefore, twin researchers bear the burden of proof for identifying specific and exclusive trait-relevant factors that contribute to the cause of the behavioral characteristic they are studying, and they continue to bear the burden of proof for showing that MZ and DZ pairs are not differentially exposed to these factors. Although most twin researchers now recognize that MZ and DZ environments are very different, they have been unable to identify specific and exclusive “trait relevant” environmental factors for schizophrenia and psychosis. As the above mercury level example shows, Argument B fails because environmental toxins, child abuse, racism, severe bullying, living in poverty, and countless other possible environmental factors could be of “of etiologic relevance to” schizophrenia. At the same time, twin researchers have failed to assess the impact of several identified “trait-relevant” environmental factors related specifically to schizophrenia and psychosis, which Roar Fosse and colleagues showed are more similarly experienced by MZ as opposed to DZ twin pairs.172
Furthermore, the circular Argument A position that twins create their own environments because they are more similar genetically potentially renders Argument B largely irrelevant. This is because, even if critics show that MZ pairs experience more similar trait-relevant environments than DZ pairs, twin researchers could still argue that MZ pairs “created” or “elicited” their more similar trait-relevant environments. The validity of the EEA and the twin method, therefore, rests mainly on the acceptance or rejection of Argument A.
It is noteworthy that genetic researchers do not invoke Argument A or Argument B when evaluating the results of family studies, despite the fact both family studies and the twin method compare groups experiencing very different environments. They arbitrarily choose to apply a standard to family studies that they choose not to apply to the twin method, and then arrive at entirely different conclusions about how to evaluate the results of each research method.
As I showed in my 2004 book The Gene Illusion: Genetic Research in Psychiatry and Psychology Under the Microscope, many leading psychiatric twin researchers of the post-World War II era recognized that environmental factors explain part—but in their opinion only part—of higher MZ concordance rates.173 For example, in the process of arguing that twin studies are “still our best method,” Kringlen wrote in 1976, “The total difference in concordance rate between MZ and DZ twins cannot be ascribed to genetic factors only. A series of studies of both normal and abnormal twins show that the environment of the MZ twin pair is more similar than the environment of the DZ twin pair.”174 And according to founding behavioral geneticists John Fuller and William Thompson, “Perhaps the best judgment is that part of the greater similarity between MZ cotwins comes from the fact that they are treated more alike by parents and associates.”175 Contemporary defenders of psychiatric twin research usually fail to acknowledge that most of their predecessors believed that some portion of MZ-DZ concordance rate differences is caused by environmental factors.
In a series of “EEA-test” studies spanning several decades, designed to test the validity of the equal environment assumption, twin researchers measured aspects of twins’ environmental similarity, such as whether they shared the same bedroom growing up, had common friends, were dressed alike, and so forth. Although they usually found that MZ pairs grow up experiencing much more similar environments than experienced by same-sex DZ pairs, these researchers typically concluded in favor of the EEA and the twin method on the basis of Argument A, Argument B, or both.176
Many critics, however, have argued that schizophrenia twin studies are no more able than schizophrenia family studies to disentangle the potential impact of genetic and environmental influences simply because MZ pairs experience much more similar environments, and much greater levels of emotional closeness and identity confusion, than experienced by DZ pairs. That is, in both family studies and twin studies, the results can be explained entirely by non-genetic factors.
Same-Sex vs. Opposite-Sex DZ Twin Concordance Rates
Additional evidence that schizophrenia twin concordance rate differences are caused entirely by non-genetic factors comes from differences between same- and opposite-sex DZ pairs. Pooled schizophrenia concordance rates across all studies that compiled such figures are 11.3% for same-sex DZs, but only 4.7 % for opposite-sex DZs.177 According to genetic theory same- and opposite-sex DZ schizophrenia concordance rates should be similar. But they are not.178
DZ Twin vs. Non-Twin Sibling Concordance Rates
We can also compare the behavioral resemblance (concordance) of DZ twins, versus that of the ordinary (non-twin) siblings of people diagnosed or labeled with schizophrenia. DZ pairs and sibling pairs share the same average genetic relationship to each other; the only difference being that DZ pairs are born at the same time, whereas ordinary siblings are not. In Gottesman’s widely reproduced “Figure 10” of the schizophrenia “grand average relative risk” percentages, the DZ twin risk is given as 17%, but as only 9% for siblings.179 An even larger difference is found in Slater’s 1953 twin study. Schizophrenia concordance among Slater’s same-sex female DZ pairs was 22.5% (9/40), while the rate among the non-twin siblings of a twin diagnosed with schizophrenia was only 4.6% (26/568). In this comparison, female same-sex DZ concordance was almost five times higher than the sibling rate.180
Although not addressed by Gottesman and most authors reproducing his Figure 10, genetic theory cannot account for these differences, since both sibling sets have the same genetic relationship to each other. Theories emphasizing DZ twins’ stronger psychological bond, more similar treatment, and greater physical proximity would predict a greater DZ twin versus sibling risk, and this is what we find in Gottesman’s Figure 10. Critics such as psychiatrist Don Jackson discussed in 1960 the “very striking finding that same-sexed fraternal twins, especially sisters, have a much higher concordance rate than ordinary sibs.”181 And in 1984, Richard Lewontin, Steven Rose, and Leon Kamin observed that “from an environmental viewpoint—and only from such a viewpoint—we would expect concordance among DZs to be higher than among ordinary sibs.”182 It follows that environmental factors impact MZ-DZ comparisons in a similar way.
One of numerous examples of an influential secondary source misreporting and misrepresenting schizophrenia genetic research is found in the following passage from psychiatrists Richard Keefe and Philip Harvey’s 1994 book Understanding Schizophrenia: A Guide to the New Research on Causes and Treatment. Arguing in support of the EEA in schizophrenia twin research, they wrote:
“Studies have been performed in order to examine the relative similarity of experience of identical [MZ] twins and same-sex fraternal [DZ] twins. These studies have found that the similarity of experience between these two types of twins is essentially identical. Since same-sex and opposite-sex fraternal twins have exactly the same concordance rate for schizophrenia, variations in environmental experiences between fraternal and identical twins as a cause of the difference in prevalence seems even more unlikely.”183
It is not true that the “similarity of experience between these two types of twins is essentially identical.” Almost all studies investigating this issue found the opposite, that MZ pairs experience more similar environments than experienced by DZ pairs. Most contemporary twin researchers now recognize that MZ and DZ environments are not equal, but they continue to uphold the EEA and the twin method on the basis of the previously discussed faulty “twins create their own environments,” and “trait-relevant” arguments. We also saw that, as opposed to being “exactly the same,” same-sex DZ schizophrenia concordance is 2-3 times higher than opposite-sex DZ concordance.
Keefe and Harvey also claimed that “the concordance rate between fraternal twins is about 15%, which is similar to the concordance rate between any two siblings.”184 Yet one page earlier, they provided a table with the title “Approximate Rates of Schizophrenia in Biological Relatives of Persons with Schizophrenia.” Their own table showed a DZ rate of 15%, but a sibling rate of only 6-9%—roughly one-half of the DZ rate.185 Despite Keefe and Harvey’s misrepresentation of twin study data in Understanding Schizophrenia, the book’s dust jacket featured an endorsement by Gottesman, at the time the world’s leading expert on the genetics of schizophrenia. Gottesman enthusiastically recommended the book, comparing it to Dr. Spock’s famous book on baby and child care.
A 2015 Twin Study “Cathedral”
In 2015, Dutch twin researcher Tinca Polderman and colleagues published a meta-analysis (analysis of combined studies) of 2,748 reared together twin studies performed in 39 countries between 1958 and 2012, which looked at more than 17,000 physical, medical, and psychological traits.186 They found that for most traits, MZ twin correlations were higher than the DZ twin correlations. They concluded that their “results provide compelling evidence that all human traits are heritable.” Turkheimer called this study “a cathedral [that] was erected in honor of twin studies in the prestigious journal Nature Genetics.”187 Amazingly, Polderman and colleagues failed to mention or defend the all-important EEA, that is, the assumption upon which genetic interpretations of MZ-DZ differences in all 2,748 studies were based. Because this assumption is false, it doesn’t matter whether researchers pool together the results of 5 twin studies, 500 twin studies, 2,748 twin studies, or a million twin studies. Two wrongs don’t make a right, and data from 2,748 twin studies—where each study, and the pooled results, are based on the same false assumption—provide no valid evidence in support of genetic theories of human behavioral differences.
Comparing twin research and accompanying heritability estimates to a “cathedral” is fitting. A cathedral is an awe-inspiring building with altars, statues, artwork, majestic arches and columns, and stained-glass windows, where worshippers are expected to accept on faith doctrines that have no scientific basis whatsoever. Behavioral genetics and psychiatric genetics publications are also impressive looking, and contain awe-inspiring statistics, models, and diagrams. Turkheimer wrote in a May 28th, 2015 blog posting that the Polderman study “represents an inconceivable amount of work. And the meta-analysis itself is beautifully executed. The graphs are striking, the numerical analysis is sophisticated.” These graphs and analyses certainly are spectacular and sophisticated, but they are more akin to a cathedral’s beautiful stained-glass windows because the critical assumption underlying twin researchers’ genetic interpretations of (pooled or non-pooled) MZ-DZ behavioral correlation differences is utterly false. As the British medical statistician Lancelot Hogben warned over 80 years ago, almost as if he had this study in mind,
“There is a danger of concealing assumptions which have no factual basis behind an impressive façade of flawless algebra.”188
Due to advances in technology since 1933, we could “upgrade” Hogben’s warning by tacking on the phrase, “…and computer-generated online or pdf color graphics, diagrams, and statistical analyses.”
A Pulitzer Prize Winner Gets the Basic Facts Wrong
In his best-selling 2016 book The Gene: An Intimate History, Pulitzer Prize winning cancer physician Siddhartha Mukherjee made definitive claims in favor of genetics on the basis of schizophrenia twin research, even though he was only marginally familiar with this body of research.189 Similar to mainstream psychiatry, Mukherjee saw schizophrenia as “a complex, polygenic illness, involving multiple variants, multiple genes, and potential environmental or chance triggers.”190
“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.”191 Strangely, he cited a 1977 autism twin study as the single source of this claim, whose authors did not discuss schizophrenia twin research.192 In fact, no new schizophrenia twin studies were published in the 1970s. The 1973 Danish study by Margit Fischer, the 1972 book by Gottesman and Shields, and Pekka Tienari’s 1975 final report on his Finnish twin sample were all based on twin samples first reported in the 1960s.193 Moreover, the “first clues” produced by schizophrenia twin research go all the way back to Luxenburger’s 1928 study, and at least eight more, including Kallmann’s and Slater’s studies, were published up to the 1970s.
Moving on to the next decade, Mukherjee wrote, “Throughout the 1980s, fleets of twin studies strengthened the case for a genetic cause of schizophrenia.” In “study upon study,” he wrote, “the concordance among identical twins exceeded that of fraternal twins so strikingly that it was impossible to deny a genetic cause.” These “fleets” of 1980s twin studies, in Mukherjee’s view, helped “bring sanity to the study of madness” because they helped overturn unsupported “seductive” psychoanalytic explanations of psychosis.194
In fact, no “fleets” of schizophrenia twin studies appeared in the 1980s. The only original-data schizophrenia twin studies published in that decade were the low concordance rate Koskenvuo et al. 1984 study, which we saw is almost never mentioned in the psychiatry and psychiatric genetics literature, and the 1983 Kendler and Robinette study discussed below. In a frequently cited 2003 schizophrenia twin study meta-analysis by leading psychiatric genetic researchers Patrick Sullivan, Kendler, and Michael Neale (over 1,500 citations as of August, 2017, according to Google Scholar), the only 1980s twin study listed was the Kendler and Robinette publication, and the only 1970s twin study listed was Tienari’s final 1975 report based on his 1963 study.195 The 1984 Koskenvuo study, as usual, was missing in action.
Mukherjee saw Kendler and Robinette’s 1983 “enormous National Academy of Sciences (NAS)” twin study as “definitively linking schizophrenia to genetic causes.” He claimed that this investigation found that “identical twins possessed a striking 30 to 40 percent concordance rate for schizophrenia.”196 But there was nothing new or “striking” about this finding, since earlier researchers had been reporting similar or higher concordance rates since the 1920s. The key question, as always, is how we should interpret these findings.
The original NAS schizophrenia twin study, which was based on U.S. military veterans identified through a twin registry, was published by American psychiatrist William Pollin and his colleagues in 1969.197 Recognizing that “nine major twin studies of schizophrenia” beginning with Luxenburger had come before them, Pollin and colleagues reported an MZ concordance rate of only 13.8% (11/80). Although they concluded that genetic factors play a role, they also concluded that, because “85 per cent of the affected monozygotic pairs in the sample were discordant for schizophrenia,” the “role of the suggested genetic factor appears to be a limited one.”198 Pollin and colleagues recognized that the twin method’s assumption that MZ and DZ environments are similar (the EEA) is “incorrect.”199 They concluded that “the variance between the two groups may be explained by environmental as well as by genetic hypotheses.”200
Kendler and Dennis Robinette’s 1983 study was a “16-Year Update” of Pollin and colleagues’ original NAS study. This update increased the original MZ sample from 80 pairs to 164 pairs, but only 30 pairs were concordant for schizophrenia, resulting in a modest combined total 18.3% pairwise concordance rate. In this “definitive” study, therefore, over 80% of the genetically identical MZ pairs were discordant for schizophrenia. Kendler and Robinette’s “striking 30 to 40 percent concordance rate” (to be precise, 30.9%) was based on these researchers’ use of the “probandwise” concordance method, which double-counts the number of concordant pairs and produces higher concordance rates. They did not comment on the validity of the EEA, and concluded that their results “support the etiologic importance of genetic factors in schizophrenia.”201 The combined pairwise MZ schizophrenia concordance rate for the only two schizophrenia twin studies published in the 1980s (Koskenvuo et al., Kendler & Robinette) is only 16% (38/237).
Ironically, in The Gene: An Intimate History Mukherjee focused on the doubts surrounding the EEA, and on the twin study “impasse” these doubts had created, when discussing what he saw as the need to study reared-apart twins. In this context, he wrote that MZ pairs are “reared in the same home, by the same parents, [are] often schooled in the same classrooms by the same teachers, [and are] dressed, fed, and nurtured identically.” Reared-together MZ pairs, Mukherjee understood, “offered no self-evident way to separate the effects of genes versus the environment.”202 Comparing reared-together MZ pairs to DZ pairs, in his view, only “partially solved” this problem. He noted the critics’ argument that MZ-DZ comparisons are “intrinsically flawed” because MZ pairs may be “treated more similarly…by their parents” than are 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.” 203 When Mukherjee turned specifically to schizophrenia twin research based on reared-together pairs, however, he did an about-face. His argument (in the same book) that MZ behavioral resemblance could be caused by the more similar environments they experienced was conveniently forgotten, and in the schizophrenia context he claimed that twin study results were “definitive.”
Although much of what he wrote about schizophrenia twin research was wrong—just as much of what he wrote about the Minnesota Study of Twins Reared Apart (MISTRA) and reared-apart twin studies 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 schizophrenia twin research, yet he thought he knew enough to write about it in a very positively reviewed book that could well earn him a second Pulitzer Prize. As the saying goes, a little knowledge is a dangerous thing.
Other Twin Study Designs
The Offspring of Discordant MZ Pairs Design
Two other types of schizophrenia twin studies should be mentioned. The first type studied schizophrenia rates among the biological offspring of discordant reared-together MZ pairs (one twin is diagnosed with schizophrenia, while the other is not). Psychiatric geneticist Margit Fischer introduced this method into twin research in 1971, after which appeared a frequently cited 1989 follow-up study based on Fischer’s pairs by Gottesman and Aksel Bertelsen (over 525 Google Scholar citations as of August, 2017).204 Genetic theories predict finding comparable schizophrenia rates among the offspring of these pairs, and this is what these researchers claimed to have found.
However, the offspring of discordant MZ pair design is unable to adequately separate potential genetic and environmental influences, and small samples make it unlikely that statistically significant differences will be found between these offspring groups. The discordant co-twins grew up in the same family, and like most MZ pairs they experienced similar environments and treatments. Most of the environmental factors that could have contributed to a schizophrenia diagnoses in one of the twins were probably also experienced by their “well” co-twin, who could have manifested deviant rearing patterns toward his or her offspring. There are also doubts about the accuracy of the diagnoses, especially since most of the offspring were not personally examined by the researchers, and there were no definite schizophrenia cases in Fischer’s total sample.205 Torrey considered the case material handed down by Fischer to be “markedly unsatisfactory,” and found that none of the MZ pairs “met the basic criteria for twins with clearly diagnosed schizophrenia in the index twin and verifiable (i.e., based on interview) normality in the co-twin.” Torrey argued that Gottesman and Bertelsen’s decisions about which twins to retain or remove were “entirely arbitrary,” and that their conclusion in favor of genetic factors “is premature at best.” More details on problems with these studies can be found elsewhere.206
Twins Reared Apart
Although no systematic schizophrenia “reared-apart” twin study has ever been published, there have been a few single-case reports of purportedly reared-apart MZ pairs concordant or discordant for schizophrenia. In psychologist Susan Farber’s 1981 review of these cases, she concluded that only nine MZ pairs qualified as having been truly reared-apart. However, even in these cases the twins were aware of each other’s existence, and had periodic contact. The pair that in Farber’s opinion was the “best separated set in the literature” (Craike & Slater’s 1945 “Edith & Florence”) was for her so “poorly separated” that the original authors wondered whether the twins’ knowledge of each other contributed to their “delusional systems.”207
Regardless of how many pairs of individual reared-apart MZ twins are reported concordant or discordant for schizophrenia, they do not provide scientific evidence in favor of genetics because, among other reasons, they are anecdotal reports that were not part of a systematic study. As the saying goes in science, “The plural of anecdote is not data.” Perhaps this is why Gottesman concluded in 1982, “After a quarter century of experience with twins reared together and twins reared apart, it is my conviction that twins reared apart are a wonderful source of hypothesis generation, but not a useful source for hypothesis testing.”208
Conclusion: Genetic Interpretations of Schizophrenia Twin Studies Must Be Rejected Outright
There seems to be no end to far-fetched and even comical “findings” from twin method MZ-DZ comparisons, which use methods similar to those used in schizophrenia research. Among these we find a twin study whose authors concluded in favor of a genetic basis for being a “born again Christian” (65% heritability), a twin study that found important genetic influences on tea and coffee drinking preferences, a twin study that found that the heritability of “loneliness in adults” is 48%, a twin study that found genetic influences on “problematic masturbatory behavior,” a twin study that found a substantial genetic component in British election voting choices, and a twin study of the “frequency of female orgasm,” which found that “overall, genetic influences account for approximately 31% of the variance of frequency of orgasm during sexual intercourse, 37% of the variance of frequency of orgasm during sexual contact other than during intercourse, and 51% of the variance of frequency of orgasm during masturbation.”209 These studies seem more appropriate for the Science and Technology section of the satirical news source The Onion than they do for publication in scholarly academic journals.
Most likely, twin studies of schizophrenia and psychosis have recorded nothing more than research bias, MZ pairs’ more similar environments and treatment, MZ pairs’ higher levels of identity confusion and attachment to each other, and MZ pairs’ greater tendency to experience folie à deux (shared psychotic disorder) than DZ pairs. As David Rosenthal concluded in 1979, because “many investigators have shown that the psychological aspects of being an identical [MZ] twin are quite different from those involving DZ twinship,” in “both family and twin studies, the possible genetic and environmental factors are confounded, and one can draw conclusions about them only at considerable risk.”210
Elsewhere, Rosenthal described a key aspect of the critics’ position as follows:
“Psychological factors unique to monozygotic twins, especially that of shared identity, have been described vividly by several investigators who maintain therefore that the equal environment assumption is ill-founded, that solely on psychological grounds one would predict a higher concordance rate for monozygotic twins, and that the inference of a genetic contribution to the disorder [schizophrenia] is not warranted based on such findings alone.”211
Leading psychiatric genetic twin researchers such as Kendler believe that the EEA has been validated by the EEA-test study results.212 And yet, Kendler wrote in a 1993 edition of a leading psychiatric journal, “With some uniformity, the available empirical evidence suggests that the EEA is probably at least approximately correct for the normative traits and psychiatric disorders studied.”213 This single sentence by the leading theoretical defender of twin studies and the EEA in psychiatry contains hedged terms such as “some uniformity,” “available empirical evidence,” “suggests,” “probably,” “at least,” “approximately,” and “normative.” It is hardly a ringing endorsement of the EEA.
Since the 1920s, the twin method has been based on the assumption that reared-together MZ and DZ twin pairs grow up experiencing similar environments, even though most people—including most leading twin researchers—now understand that MZ and DZ environments are very different. This has compelled these researchers to concoct illogical arguments to allow them to continue their work, and authoritative psychiatric texts and popular media outlets continue to endorse the original twin researchers’ mistaken interpretations of their results in favor of genetics. When the EEA’s validity was challenged by American criminologists Callie Burt and Ronald Simons in 2014, genetic researchers in that field felt the need to produce a 61-page article in response, even though they believed that the debate had been “settled” many years earlier. Their response, which required the collaboration of no fewer than 24 people, did nothing to counter the critics’ arguments against the EEA’s validity.214
It is crystal clear that the twin method is no more able than a family study to separate the potential influences of genes and environment on schizophrenia and other forms of behavior, leading to the conclusion that genetic interpretations of all past, present, and future MZ versus same-sex DZ twin method comparisons in psychiatry, and in the social and behavioral sciences in general, must be rejected outright.