The 110-Year “Schizophrenia Genetic Research” Train Wreck

The “genetics of schizophrenia” area of research is currently in disaster mode and awaits its endpoint. Instead of drawing the necessary conclusions, psychiatry and its psychiatric genetics subfield (International Society of Psychiatric Genetics [ISPG]) continue to speak—publicly—of progress, discovery, and being “driven by rapid advances in genomic technologies.” The reality, as I will show, is that the evidence psychiatry ceaselessly puts forward in support of schizophrenia as a genetic disorder is stunningly weak. This conclusion has important implications for prevention and intervention strategies, and also for other areas of psychiatry and human behavior in the larger context of the “nature-nurture” question.

In this article, I briefly review the lack of evidence that schizophrenia is a “heritable disorder” in the context of five decades of causal gene discovery failure. (The term “psychosis” is preferred by those of us who are critical of disease model approaches.) The first part of this article describes DNA-based (molecular genetic) failure and its evaluation by a well-known schizophrenia researcher. In the second part, I challenge the arguments of “Astral Codex Ten” blogger and psychiatrist Scott Alexander, who posted “It’s Fair to Describe Schizophrenia as Mostly Genetic” in January 2024.

I have been writing critically about genetic research on schizophrenia for many years, beginning with my 1998 dissertation. I summarized my dissertation’s findings and conclusions in a 1999 article, where I concluded, “Based on the weight of the evidence, it is predicted here that a gene for schizophrenia will not be found, because it does not exist.” I have continued to write about schizophrenia genetic research in articles, chapters, and four books, where I made similar predictions. In 2023, I published Schizophrenia and Genetics: The End of an Illusion, where I described the fruitless search for causal genes in the context of science’s “replication crisis.”

I published The Missing Gene: Psychiatry, Heredity, and the Fruitless Search for Genes in 2006. In that book, I concluded that genes for the major psychiatric diagnoses most likely do not exist. I predicted that the gene discovery and gene association claims of that era would not hold up. As it turned out, none of them did (see the linkage and candidate gene study discussion below). Looking back, the critics of 2006 were right and the leaders of psychiatry and psychiatric genetics were wrong (see my 2005 American Journal of Psychiatry exchange with leading psychiatric geneticist Kenneth Kendler here and here). Despite subsequent gene association claims based on the newer methods I will soon discuss, the existence of disordered genes that play a role in causing the major psychiatric diagnoses remains unlikely.

Critics often focus on original-source research publications. Relying on secondary sources and review articles by leading experts can perpetuate myths, as these accounts sometimes misrepresent basic facts and often endorse flawed, confirmation-biased original research. In some cases, authoritative writers and the authors of behavioral science textbooks, including Nobel Prize winners, appear unfamiliar with original research publications. Some leading authors discuss studies that don’t exist (see here and here).

Origins of Psychiatric Genetics

Psychiatric genetics began with Ernst Rüdin’s 1916 family study of “dementia praecox” (schizophrenia). Rüdin was a Swiss-German psychiatrist and founder of the psychiatric genetics field, as well as a 1905 co-founder of the German Society for Racial Hygiene. Rüdin, who believed that a single recessive gene caused schizophrenia, went on to establish the “Munich School” of psychiatric genetics. A major aspect of Munich School activity and research during the Third Reich involved calculating psychiatric “morbid risks” in support of the regime’s 1933 forced sterilization law and other racial hygienic (eugenic) measures. In a 2022 article, Kenneth Kendler wrote that Rüdin helped give “scientific prestige and credibility” to the regime’s “racial beliefs and policies.” For a detailed and chilling description of the work and deeds of Rüdin and the Munich School in the 1930s and ‘40s, I recommend Chapter 3 of The Nazi Symbiosis: Human Genetics and Politics in the Third Reich, a 2010 book by Sheila Faith Weiss.

In the 1920s, Rüdin performed a genetic family study of “manic-depressive insanity” based on over 600 families. He wrote a 160-page manuscript describing this study, but suppressed it because the results did not align with his genetic theories and worldview. The central fallacy of Rüdin’s early family studies was a failure to understand that psychiatric conditions can run in families for non-genetic reasons. Although most genetic researchers now correctly recognize that “running in the family” does not equal “genetic,” Rüdin and most of his eugenics movement researcher contemporaries and followers believed that it did.

The first schizophrenia twin study was published in 1928 by Hans Luxenburger, a Munich associate of Rüdin. As Kendler noted, Luxenburger was also “committed to racial hygiene.” Seventeen additional studies were performed by researchers in different countries (the most recent was published in 2018). The first schizophrenia adoption study was published in 1966. DNA-based gene-finding attempts began in the 1970s and continue to this day.

The Failed Linkage and Candidate Gene Eras

The two main methods used in an earlier era of psychiatric DNA-based research were linkage studies and candidate gene association studies. Both methods failed to identify causal genes. The authors of a 1983 DNA-based linkage study identified the gene responsible for Huntington’s disease, a genuine medical condition. This discovery gave rise to expectations that investigators would soon discover genes causing schizophrenia, bipolar disorder, major depression, and other major psychiatric (behavioral) conditions as well. From a 2013 vantage point, leading psychiatric genetic researcher Stephen Faraone recalled, “In the 1970s and 1980s, hope ran high as new methods in molecular genetics promised quick discoveries and answers to basic questions of etiology and pathophysiology.” Inspired by the Huntington’s discovery, some psychiatric genetic researchers of the 1980s mapped out family pedigrees and anticipated rapid gene discoveries and flights to Stockholm to receive their Nobel Prizes (see here, 19:20). However, despite high hopes and discovery-anticipation “euphoria” among psychiatric genetic researchers of the 1980s (including a highly publicized 1988 linkage study claim that proved unfounded), the outcome was failure.

The heyday of the failed candidate gene era (roughly 1995-2010), which overlapped with and followed linkage research, involved billions of dollars. According to top behavioral geneticist Robert Plomin in his 2018 book Blueprint, for schizophrenia alone, “over 1,000 papers reported candidate gene results for more than 700 genes.” The financial cost was approximately $250 million (or around $500 million adjusted for inflation), and Plomin asked rhetorically, “How can so many published papers have got it so wrong?”

As seen here, here, here, and here, over two decades of behavioral and psychiatric candidate gene discovery claims turned out to be false alarms. In other words, the behavioral candidate gene era was an unmitigated disaster. In Blueprint, Plomin described the era as a “flop,” a “fiasco,” and an “approach” that “failed everywhere.” Previously, Plomin had conducted his own candidate gene studies and claimed discoveries, and cited others’ supposed discoveries in his Behavioral Genetics textbook and elsewhere.

Nevertheless, candidate gene mythology continues in the social media, the internet, and popular works. For example, journalist Robert Kolker published his bestselling 2020 book Hidden Valley Road: Inside the Mind of an American Family about a Colorado family in which six of the ten male offspring were diagnosed with schizophrenia. Kolker described the work of two persistent psychiatric researchers, who, he said, identified schizophrenia risk genes with the aid of the family’s blood samples. However, the two genes Kolker called “discoveries” (CHRNA7 [patent now expired] and SHANK2) were actually candidate gene non-discoveries. Journalists and researchers understand that careers are made, prizes are won, endorsements from famous people are received, and books and potential movie deals are sold by stories about discoveries, not by failure or inconclusive results.

Here are 15 examples of books from the era that reported positive linkage findings and “promising” candidate genes “implicated” as causes of psychiatric conditions (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15). Among them are The American Psychiatric Publishing Textbook of Schizophrenia (2006, edited by Jeffrey Lieberman and colleagues), Behavioral Genetics (2000, by Plomin and colleagues), and Behavioral Genetics in the Postgenomic Era (2003, edited by Plomin and colleagues). These days, if the authors and other researchers of the linkage and candidate gene era see these books on a library shelf, they might experience a moment of acute embarrassment.

The initial working draft of the human genome sequence was published in 2001, which many people hoped would lead to rapid gene discoveries in psychiatry. Faraone and his psychiatric genetics colleague Ming Tsuang wrote in 1999, “From the perspective of psychiatric genetics, the Human Genome Project (HGP) is an immense factory producing and refining the tools we will need to discover the genes that cause mental illness.” And according to leading genetic researchers Kathleen Merikangas and Neil Risch, writing in 2003, “Completion of the human genome project has provided an unprecedented opportunity to identify the effect of gene variants on complex phenotypes, such as psychiatric disorders.” But it didn’t happen.

Genome-Wide Association Studies, Polygenic Risk Scores, and Rare Variants

Current schizophrenia “gene association” claims are based on the supposedly “hypothesis-free” genome-wide association (GWAS) method, which dates back to the period 2005-2007. A GWAS might find an association (correlation) between genetic variants and schizophrenia, but it doesn’t provide evidence that genes play a role in causing it. Correlations are not causes and may be spurious or the result of systematic error, or they may simply be chance findings, as earlier behavioral linkage and candidate gene researchers learned the hard way. In 2022, Thomas Insel, the biologically oriented former director of the U.S. National Institute of Mental Health (NIMH), recognized that “in contrast to the mutations discovered for cancer or rare diseases, none of the genetic variants associated with mental illness can be considered causal.”

The polygenic risk score (PRS) method is a DNA-based technique that combines statistically significant and nonsignificant individual SNP associations (“single-nucleotide polymorphism” common variants) identified in a GWAS to produce a polygenic (composite) risk score. A large-sample 2022 GWAS by the Schizophrenia Working Group of the Psychiatric Genomics Consortium calculated a modest “European cohort” 7.3% schizophrenia PRS, which fell to 2.4% “when restricted to genome-wide significant SNPs.” These low scores were buried in the text and did not appear in the article’s abstract, and can be compared with dramatically different twin study-based estimates of 80% heritability (but see the twin study critique below). In any case, like a GWAS, a PRS identifies, at best, gene-behavior correlations, not causes. As noted by behavioral geneticist Eric Turkheimer in his 2024 book Understanding the Nature-Nurture Debate (see my review here or here), PRS studies “are a definitive step backward from trying to figure out which genes are responsible for the heritability of behavior.”

Another research area focuses on potential rare variants (defined as variants found in less than 1% of the population), such as copy number variants, or “CNVs.” Many studies claiming CNV-schizophrenia associations have appeared over the years, but they are subject to many of the problems found in other DNA-based methods. In 2020, Jonathan Flint and Kendler, two of the world’s leading psychiatric genetic researchers, wrote that the “early hope that CNVs would reflect the ‘royal road’ to understanding molecular genetic effects on schizophrenia has been disappointing.”

Commenting in 2008 at the height of candidate gene era excitement, psychiatric researcher Timothy Crow wrote that although the “schizophrenia gene pond” was in fact “empty,” a “chorus of reviews…pervades the literature and will convince all but the most inquisitive that a solid foundation of evidence supports the pathophysiological relevance of these candidate genes.” We should keep in mind Crow’s 2008 “chorus of reviews” comment when assessing current claims about genes “implicated” by GWAS, PRS, and CNV studies—same chorus, slightly modified lyrics.

A Leading Schizophrenia Researcher Describes Gene Discovery Failure

E. Fuller Torrey is a well-known schizophrenia researcher and the author of several editions of Surviving Schizophrenia: A Family Manual. While believing that genetic factors play some role, Torrey sees schizophrenia as a brain disease possibly caused by infectious agents (see Peter Simons’ 2024 Mad in America article, and my discussion of disease theories below). In a 2024 publication, Torrey caused a stir by describing gene discovery failure and suggesting that “since schizophrenia does not appear to be a genetic disorder,” the NIMH’s “research portfolio should be reviewed.” Torrey’s statement smacked of psychiatric heresy, as it knocked out a key component of Eugen Bleuler’s original 1911 conception of schizophrenia as a largely hereditary disease. (Bleuler recommended sterilization on eugenic grounds and later worried that dysgenic “racial deterioration” could occur if eugenic policies were not implemented.) The concept of schizophrenia as a hereditarily predisposed brain disease remains central to the psychiatric profession’s identity.

The decades-long gene-search endeavor, Torrey said, has been “like a search for a psychiatric yeti”:

The Human Genome Project was undertaken primarily to discover genetic causes and better treatments for human diseases. Schizophrenia was targeted since three of the project’s principal architects had a personal interest and also because, based on family, adoption, and twin studies, schizophrenia was widely believed to be a genetic disorder. Extensive studies using linkage analysis, candidate genes, genome wide association studies [GWAS], copy number variants, exome sequencing and other approaches have failed to identify causal genes. (Emphasis added)

Torrey wrote that “schizophrenia alone has probably been the subject of more genetic breakthrough announcements over the last thirty years than any other human disease.” And yet, despite spending over (a non-inflation-adjusted) 8 billion NIMH dollars in attempts to find causal “schizophrenia genes,” none have been found. A major paradigm-shattering event, it would seem.

Given his insider status and reputation as a leading biologically oriented schizophrenia expert and researcher, Torrey’s article motivated psychiatrist Awais Aftab to post two 2024 “Psychiatry at the Margins” articles on the topic of schizophrenia genetics (here and here). Displaying some confusion, Aftab concluded, “It may very well be the case that schizophrenia is heritable and polygenic, and yet there are no ‘genes for’ schizophrenia.”

Back in 2020, Aftab and I had an online exchange on schizophrenia genetics, where he wrote that my “criticizing the notion that schizophrenia has a substantial genetic component” is typical of “degenerative scientific programs.” Perhaps he is now ready to modify this earlier stance.

In response to Aftab and Torrey’s 2024 articles, psychiatrist/blogger Scott Alexander posted his article, “It’s Fair to Describe Schizophrenia as Mostly Genetic,” in January 2024. In the following sections, I focus on Alexander’s article and explain why it’s not fair to describe schizophrenia as “mostly genetic”—or even genetic at all.

Scott Alexander: “Schizophrenia is Mostly Genetic”

While providing no indication that he was familiar with the original-source genetic research publications, Alexander followed standard psychiatry textbook accounts by asserting or implying that “schizophrenia” is a valid concept and that it (1) is a disease, (2) is strongly heritable (80%) based on earlier twin and adoption studies, and (3) is “mostly genetic” based on analogies with lung cancer and other medical conditions. I will now briefly examine these claims.

1) Is Schizophrenia a Disease?

Alexander implicitly endorsed the disease model throughout his post. He cited Wikipedia in support of his claim that schizophrenia “is famously nearly the same prevalence in every society studied,” but other reports do not support this claim.

There are no laboratory tests for schizophrenia. A diagnosis is made based on a combination of factors, which may include family history, personal history, behavior, speech, self-report, and reports from others. Earlier speculation that schizophrenia is caused by a genetically linked deficiency of the neurotransmitter dopamine (the “dopamine hypothesis”) has gone by the wayside, a fate similar to the unsupported “serotonin theory” of major depression.

Psychologist Mary Boyle, the author of Schizophrenia: A Scientific Delusion?, argued that in the absence of direct evidence, psychiatry must resort to “smoke and mirrors” tactics to support its brain disease claims. Psychologist John Read quoted a passage from the 1913 edition of  Emil Kraepelin’s influential textbook on psychiatry, where Kraepelin wrote that the causes “are at the present time still wrapped in impenetrable darkness.” Read commented that the key phrase is “at the present time,” which has “been used ever since by researchers forever on the verge of finding the biological cause of schizophrenia.” Read contributed several important chapters to the second edition of Models of Madness: Psychological, Social and Biological Approaches to Psychosis, a 2013 book that he co-edited.

Psychiatry is a medical specialty that treats conditions it claims are brain diseases or “chemical imbalances,” but for which there is no scientific proof. If researchers find that changes in the brain are correlated with a schizophrenia diagnosis, it doesn’t necessarily mean that brain malfunction is the cause. Such interpretations can confuse cause and effect, since people’s life experiences can modify brain anatomy and function (neuroplasticity), and neuroleptic (“anti-psychotic”) drugs have been shown to cause brain shrinkage and other abnormalities.

2) Is Schizophrenia Strongly Genetic (“80% Heritable”) Based on Earlier Twin and Adoption Studies?

Schizophrenia twin research. Alexander believes that “the results of twin studies have been mostly corroborated by adoption studies, a separate methodology.” In the next section, I will discuss some problems found in schizophrenia adoption research.

All 18 schizophrenia twin studies were based on comparing the concordance rates of MZ (monozygotic, identical, 100% genetically similar) versus DZ (dizygotic, fraternal, on average 50% genetically similar) twin pairs. All pairs were reared together in the same family environment. (Apart from a few anecdotal reports, there are no schizophrenia studies of reared-apart twins.) Twin researchers interpret higher MZ versus DZ concordance rates genetically by assuming that both types of twins grow up experiencing “equal” environments. This assumption, however, is clearly false.

Most leading psychiatric twin researchers and twin study defenders, such as Kendler, understand that MZ pairs grow up experiencing much more similar environments than DZ pairs. Yet, they continue to defend schizophrenia twin research by using several arguments I examined in Chapter 4 of Schizophrenia and Genetics. I demonstrated that none of these arguments withstands critical examination. In 2025, I responded to Alexander’s post on the so-called “missing heritability problem,” where I concluded that the problem (the large gap between twin study and GWAS heritability estimates) is really a twin study misinterpretation problem.

Because the MZ-DZ “equal environments assumption” (EEA) is false, twin studies of schizophrenia and psychosis have likely recorded nothing more than research bias; MZ pairs’ more similar environments, treatment, and exposure to childhood social adversity; MZ pairs’ higher levels of identity confusion and attachment to each other; MZ pairs’ greater tendency to model their behavior on each other; and MZ pairs’ greater tendency to experience folie à deux (shared psychotic disorder) than DZ pairs.

There’s an interesting trend in schizophrenia twin research that is consistent with non-genetic interpretations of concordance rate differences. In all studies compiling such figures, the pooled concordance rate for same-sex DZ pairs is 2–3 times higher than for opposite-sex DZ pairs. Although both types of DZ pairs share an average 50% genetic similarity, they experience different twin relationships depending on whether they are same-sex or opposite-sex (see psychiatrist Don Jackson’s classic, though somewhat dated 1960 chapter).

Overall, in the methodologically superior twin studies published since 1962, the pooled pairwise schizophrenia concordance rates are MZ = 24% (151/641) and DZ = 4% (68/1,519). This result shows that when one MZ twin was diagnosed with schizophrenia, their genetically identical co-twin was not diagnosed with schizophrenia around 75% of the time. In the most recent study, MZ pairwise concordance was only 15% (12/81), and the DZ rate was 3% (12/367). These figures differ from those in psychiatry textbooks, which typically report 50% MZ concordance. In any case, because the EEA is false, MZ versus DZ concordance rate differences should not be interpreted genetically.

Alexander cited one study that supposedly tested the EEA in support of his belief that “the biases in twin studies are relatively small,” while arguing that some biases tend to “underestimate heritability” because “any measurement error shows up as environmental variation” (emphasis in original). He said that “if a construct is hard to measure (schizophrenia is so famously hard to define that we’re not even sure it exists), then twin studies will underestimate heritability.”

The above argument makes little sense, but it does serve to remind us that the reliability of schizophrenia and other psychiatric diagnoses is low. If researchers cannot reliably identify (diagnose) the psychiatric condition they are studying, it is not a valid condition for research purposes. Alexander’s recognition that schizophrenia is “famously hard to define” and might not even exist suggests that twin researchers conducted their studies not really knowing which twins “had” schizophrenia and which twins didn’t. Instead of supporting genetic interpretations of schizophrenia twin studies, Alexander’s statement provided an additional reason to view such interpretations with skepticism.

Schizophrenia adoption research. Turning to the schizophrenia adoption studies, Alexander believes that they “corroborate” the twin studies. We have seen, however, that due to environmental confounding and other factors, there is no valid genetic evidence from twin studies to corroborate.

The most widely cited adoption studies were conducted in Denmark by U.S. and Danish researchers using four different research designs. (The other schizophrenia adoption studies, by different investigators, were performed in the U.S. [Oregon] and Finland. I critically reviewed the Finnish study in 1999, and both studies in my 2004 book The Gene Illusion.) In theory (but not in reality, as seen in the first bullet point below), a psychiatric adoption study cleanly separates (disentangles) potential genetic and environmental influences on a condition because adopted children inherit the genes of their biological (birth) parents, but are reared by adoptive parents with whom they share no genetic relationship.

Although still routinely cited in psychiatry textbooks and other works in support of genetics, the Danish-American studies (published between 1968 and 1994) provide striking examples of massively flawed and p-hacked research. P-hacking is the practice of researchers telling “big little lies” by manipulating definitions, data, and comparisons—either openly or behind the scenes—to transform non-findings into publishable or confirmation-bias-supporting “findings” that fall below the conventional .05 level of statistical significance. As described by one group in 2015, the “widespread” practice of p-hacking occurs “when researchers collect or select data or statistical analyses until nonsignificant results become significant.” In some cases, investigators engage in p-hacking by continuing to collect data beyond their study’s planned data-collection stop-point until they obtain the desired results.

Among many potentially invalidating problem areas, to arrive at conclusions consistent with their strong genetic confirmation biases, Danish adoption study investigators Seymour Kety, David Rosenthal, Paul Wender, Fini Schulsinger, and colleagues:

• Dismissed or minimized the impact of environmental confounds such as the selective placement of children by adoption agencies, late separation, late placement, and the non-representativeness and restricted rangeof adoptive family environments.
• Changed their 1968 “Copenhagen Study” key group comparison at the last minute because the planned comparison did not produce statistically significant results in the genetic direction.
• For the same reason, performed a last-minute arbitrary reduction of their 1994“Provincial Study” index and control adoptee (proband) groups, respectively, from 42 to 33, and from 42 to 24.
• Broadened the definition of schizophrenia “as widely as it may have ever been reasonably conceived before” (Rosenthal) to include diagnoses such as “inadequate personality,” “borderline schizophrenia or pervert,” “uncertain borderline schizophrenia,” and “not too vulnerable schizoid,” and then narrowed the definition when necessary, over a 26-year period. Textbook authors and even top schizophrenia experts often call these diagnoses “schizophrenia,” thereby misleading their readers.
• Made diagnoses based on transcribed “interviews” with adoptees and relatives that, in many cases, never took place (documented in 1984 by Lewontin, Rose, & Kamin, and by Kendler & Gruenberg). They never mentioned these “pseudointerviews” in their numerous publications. Of the interviews they did conduct, they believed they could determine whether a reluctant, unfriendly, or hostile Danish interviewee was in or out of the “schizophrenia spectrum” based on a five-minute interaction at the interviewee’s front door (see the Paikin et al. chapter here).
• Temporarily removed a study-defined “schizophrenia spectrum” diagnosis to achieve statistically significant results in their 1975 “biological paternal half-sibling” comparison, a p-hacked comparison they claimed supplied “compelling evidence that genetic factors operate significantly in the transmission of schizophrenia.”
• Found no genetic influences on schizophrenia at their “Adoptees Study” 1967 data-collection stop-point (see the 1968 Rosenthal et al. chapter here). Refusing to accept these negative results, they called their findings “preliminary” and continued the study to the point at which they achieved statistically significant results, based on an extremely broad definition of schizophrenia that included “pseudoneurotic borderline,” “manic-depressive psychosis,” and “schizoid character.” They published these massively p-hacked results in 1971. Mainstream psychiatry embraced this 1971 publication while ignoring a later 1978 Danish-American article showing no statistically significant researcher-defined “consensus diagnosis” schizophrenia or schizophrenia spectrum differences between the index and control adoptee groups (see Table 3 here). Therefore, despite borrowing heavily and openly from the p-hacking playbook over a 10-year stretch, the researchers’ final 1978 Adoptees Study results were negative.

Building on the work of earlier critically minded scholars, in Chapter 6 of Schizophrenia and Genetics (and earlier publications including here, here, and here), I documented these and many other invalidating problem areas in the Danish-American adoption studies.

In the 1970s and 80s, Theodore Lidz, a Yale University psychodynamic psychiatrist who studied the families of people diagnosed with schizophrenia, wrote or co-wrote three devastating critiques of the Danish-American studies that should have buried these studies for good (see here, here, and here). However, mainstream psychiatry largely ignored Lidz’s analyses at a time when the profession was attempting to establish itself on 1980 DSM-III medical model principles amid a rhetorical “biological revolution in psychiatry” that was welcomed and promoted by the drug companies.

In the 1980s and 90s, based on 1980 DSM-III diagnostic criteria, U.S. psychiatry produced “independent analyses” of the original Danish-American studies (for example, here, here, and here). The authors of these analyses concluded that they confirmed the original findings, but to do so, they used some of the same tricks used by the Danish-American researchers (see Chapter 6 here). The main fallacy of these analyses was psychiatry’s assumption that the original Danish-American studies provided valid scientific evidence in support of genetic influences on schizophrenia, when in fact they did not. The original studies, therefore, produced no genetic findings to uphold or confirm.

The Danish-American schizophrenia adoption studies are perhaps the longest-running example of openly and blatantly p-hacked research ever seen in the behavioral sciences, and as such, should be retracted. The p-hacking practices the researchers engaged in during a nearly 30-year period are clearly seen in their publications, which suggests that p-hacking fraud was an accepted and normal aspect of psychiatric research of that era. The perpetrators were even awarded prestigious prizes for their deeds. If Alexander is willing to read or re-read the original papers produced by these studies (for example, those found in books published in 1968 and 1975, and articles published in 1971, 1974, and 1994), I am prepared to explain to him, paragraph-by-paragraph, where these studies went terribly wrong.

Alexander and Aftab, and to a lesser degree Torrey (none of whom mentioned my views and publications), believe that schizophrenia is to some degree “heritable.” This belief has led to confusion, which is especially evident in Aftab’s 2024 posts. Since we supposedly “know” from family, twin, and adoption studies that schizophrenia is 80% heritable, how do we make sense of gene-discovery failure? I argue that we can make sense of it only: (a) by recognizing that schizophrenia family and twin study results can be interpreted environmentally; (b) by recognizing that the schizophrenia adoption studies were massively flawed and in most cases p-hacked (including Leonard Heston’s often-cited 1966 U.S. [Oregon] study); and (c) by abandoning the long-controversial practice of calculating heritability estimates, which are based on unsupported assumptions and do not record the “strength” or “magnitude” of the supposed genetic influence on psychiatric conditions.

3) Is the “Smoking Causes Lung Cancer” Analogy Relevant?

A major theme of Alexander’s article was that we should view schizophrenia as a mostly genetic condition in the same way we view lung cancer as being mostly caused by smoking cigarettes. There are several problems with this argument, some of which I list below:

1. Lung cancer is a disease of the human body, whereas schizophrenia (psychosis) is not.
2. The evidence that schizophrenia is caused in part by genetic influences is weak to non-existent (p-hacked and confounded family, twin, and adoption studies), and 50 years of failed attempts to identify causal genes support this viewpoint.
3. Based on Point B, it is reasonable to assume that genes that cause schizophrenia do not exist and that researchers and others have been misled by twin and adoption studies, thereby solving the “missing heritability problem.”
4. Because many environmental risk factors for schizophrenia and psychosis are now known and mentioned by Alexander, we could just as easily view schizophrenia as a “mostly environmental ”
5. According to twin researchers, there are “moderate genetic influences on lifetime [cigarette] smoking practices.” Therefore, based on Alexander’s logic, one could conclude that lung cancer is a mostly genetic disease. Undoubtedly, the tobacco industry would view such a conclusion favorably.

Conclusions

In a 2012 article, I wrote that the probable outcome of psychiatric DNA-based (molecular genetic) research would be that genes will not be found because they do not exist, and that psychiatric genetics eventually would be recognized as a “null field” of science. As John Ioannidis defined it in 2005, a null field is an area of research “with absolutely no yield of true scientific information….The extent that observed findings deviate from what is expected by chance alone would be simply a pure measure of the prevailing bias.” I think most insiders would now agree that psychiatric candidate gene research is a scientific null field. Most likely, the same fate awaits current psychiatric genetic GWAS, PRS, rare variant research, and other DNA-based methods.

People diagnosed/labelled with schizophrenia have low reproductive rates and lower life expectancies, or what Torrey in Surviving Schizophrenia called the “schizophrenia paradox—the continuing existence of schizophrenia despite a low fertility rate and a high mortality rate.” And the DSM-5 acknowledged that “most individuals who have been diagnosed with schizophrenia have no family history of psychosis.” These additional red-flag warnings should have alerted researchers decades ago that they risked wasting their careers by searching for what Torrey aptly called a “psychiatric yeti.”

Psychiatrist Allen Frances, former Chair of the 1994 DSM-IV Task Force, wrote in 2011 that “schizophrenia is admittedly a flawed construct with limited descriptive and explanatory power. It is…wildly heterogeneous with dozens of different presentations and probably hundreds of different causes (none of them known).” The failure of schizophrenia genetic research is inseparable from the failure to reliably define the concept of schizophrenia itself.

When psychiatry finally acknowledges that “schizophrenia” (psychosis) is not “genetic” or a “disease,” society will part ways with biological diversions and inappropriate medical approaches. The behavioral sciences and society will instead focus on environmental causes, non-medical interventions, and prevention, with a necessary political change component. People might still wish to take prescribed drugs, or others might encourage or pressure them to do so, but the process should be completely transparent, including informing people in writing of potential withdrawal problems and side effects. No one should be told they are being prescribed drugs to treat a genetically influenced or caused schizophrenia brain disease, because no such disease exists.

Non-medical approaches emphasize the role of relational, social, and economic adversities, including various types of oppression, and recognize how these impact our psychological well-being and our bodily capacities and responses. An example of this approach is the “Power Threat Meaning Framework” (PTMF), developed by Lucy Johnstone, Mary Boyle, and others. In a 2020 introductory book, the authors described the Framework’s “overall message” as follows:

All forms of adversity and distress are more common in social contexts of inequality and other forms of deprivation, discrimination, marginalisation and injustice. This evidence does not support the individualisation of distress, either medically or psychologically. Instead, it implies the need for action, primarily through social policy, at the earliest possible point, before the destructive and self-perpetuating cycles are set in motion.

The schizophrenia/psychosis causality puzzle pieces will fit together much better once we knock the “genetics” pieces off the table. It is more likely, however, that psychiatry will develop new methods and produce new sets of claims that will “kick the can down the road” and keep careers, research funding, and drug company “consultant fees” intact (for example, see the lengthy 2022 schizophrenia GWAS “competing interests” statement). Ultimately, however, psychiatry will be unable to avoid its “empty gene pond” day of reckoning.

In a 1992 article, psychiatric genetic researcher Michael Owen asked, “Will schizophrenia become a graveyard for molecular geneticists?” He reassured himself and his colleagues that “talk of graveyards is premature.” Perhaps such talk was premature in 1992, but 33 years and billions of wasted dollars later, I respectfully answer Owen’s question in the affirmative. The time has come for the null field “genetics of schizophrenia” research program initiated by Ernst Rüdin 110 years ago to close up shop, which will enable scientists and society to prioritize understanding the real non-genetic causes of psychosis by focusing on human experiences, not the human body.

****

Note: I would like to thank Mary Boyle, Mike Jones, and Jonathan Leo for their helpful feedback and suggestions on earlier drafts of this article. As always, all views expressed in this article are my own and do not necessarily imply endorsement by others.