About five years ago, bisexual people on the social media platform TikTok made a facetious observation that wearing cuffed jeans was associated with bisexuality. Earlier this year, more serious and consequential associations with bisexuality were reported by two researchers from the University of Michigan: Their paper in the journal Science Advances concluded that in males, a genetic signature of bisexuality was associated with risk-taking behavior and conceiving more children.
The study is one of many genome-wide association studies, or GWAS. The researchers analyzed DNA sequences from about 450,000 individuals in the UK Biobank, a large genetics database, to examine how they differed, then tested whether those differences were associated with measures of bisexual behavior, risk-taking behavior, and number of reported offspring.
The associations drawn by the study may sound uncomfortably similar to common stereotypes that bisexual people are greedy and promiscuous, and popular news articles about the study have often made this leap. At least 24 English-language news articles have been published describing the findings of the study, of which more than two-thirds propose promiscuity or unprotected sex as an interpretation for risk-taking behavior, even though these terms are not used in the journal article. Scientists like myself worry that this study and such headlines could embolden common stereotypes and worsen biphobia, stigma, discrimination, and disparities faced by bisexual (and “bi+” — which includes related identity labels) individuals, including more frequent accusations of infidelity, higher rates of experiencing domestic abuse, worse health outcomes, and even legal implications.
GWAS have given us a way to search for genes connected with human health, including genes associated with disease. But the associations identified are not always due to causal genetic effects. For instance, GWAS have found statistically significant genetic associations for musicality, being a morning person, and educational attainment. Although there may be some genetic contribution to these traits, non-biological factors can confound and overestimate the connection. Statistical models attempt to correct for confounders (such as measures of socioeconomic status), but a large number of small uncorrected effects can cause residual confounding that overestimates how much a trait is driven by genetics. This is especially challenging for traits that are polygenic, or influenced by multiple genes, and complex behavioral traits, such as sexuality. Despite repeated efforts over the past 30 years to identify a genetic explanation for homosexuality, no “gay gene” or set of genes consistently or robustly associated with homosexuality has been identified.
GWAS have given us a way to search for genes connected with human health, but the associations identified are not always due to causal genetic effects.
Another important detail underlying GWAS is how traits are classified. In the Science Advances study, a person was defined as bisexual if they reported having intercourse with at least one male and one female in their lifetime. This method of classifying sexual orientation for genetic association studies has been contested, as it would assume a genetic contribution for willingness to explore a sexual experience even once. However, the safety and opportunity to experience same-sex intercourse and the willingness to report that experience are highly influenced by societal factors.
In England, even private male homosexual activities were criminalized prior to 1967, and public promotion of homosexuality was prohibited until 2003. Therefore, even openly supporting the freedom for people to explore homosexuality has been socially and legally risky for the majority of most study participants’ lives.
Meanwhile, the classification of individuals with “risk-taking behavior” is central to the study, but scientifically dubious. A single question — “Do you consider yourself a person who takes risks?” — is used to classify participants. But as one might imagine, what each person considers “risk,” and therefore whether one views themself as a person who takes risks, can be deeply subjective and influenced by non-biological factors such as environment and cultural norms.
The study’s conclusions make me wonder whether the authors fully appreciate the extent to which stigma, discrimination, and social environment impact queer people. Despite longstanding evidence that a lack of societal acceptance is a principal cause of suicide among non-heterosexual people, the authors suggest that the same-sex behavior genes themselves have “costs to viability,” citing higher rates of suicide among non-heterosexual people. This logic suggests that the authors view suicidality among non-heterosexual people largely as an intrinsic genetic characteristic, rather than the result of stigmatization, discrimination, and violence from a heteronormative society.
Support Undark Magazine
Undark is a non-profit, editorially independent magazine covering the complicated and often fractious intersection of science and society. If you would like to help support our journalism, please consider making a donation. All proceeds go directly to Undark’s editorial fund.
To avoid these types of misinterpretations, research involving marginalized communities should broadly engage community members during the design, implementation, interpretation, and dissemination of the research. However, no queer or bi+ organizations appear to have been involved in this research, as none are listed in the acknowledgments. According to an article about the study in Science, the authors “informally discussed their findings with queer people” and “held a campus meeting that included scientists who belong to the LGBTQ+ community.” The details of this community engagement or how feedback was incorporated into the work does not appear to be publicly available.
Stereotypes of bisexual people have even affected legal proceedings. For example, in a 1984 child adoption case in Arizona, the court ruled that a bisexual man was not suitable for parenthood, citing concerns over “ambivalence in his sexual preference.” In 1993, a bisexual woman in Florida had custody of her child removed during divorce proceedings because of the alleged negative impact of not providing “a traditional home environment.” That decision was appealed and reversed on the basis that there was no evidence to substantiate the concern. However, a study that reports a genetic predisposition towards “risk-taking behavior” and “promiscuity” by bisexual individuals could be used to supply such evidence.
Shaky scientific evidence has often been entered in legal proceedings in attempts to hinder the freedoms of queer people. During the landmark Obergefell v. Hodges U.S. Supreme Court case in 2015, which ultimately established the federal right to same-sex marriage, the opposition took this approach by citing a study which reported that children with a parent who had a same-sex romantic relationship at any time during upbringing had worse outcomes than children raised by parents who did not. However, a thorough reanalysis of the underlying data found the results were “so fragile” that they were likely due to misclassifications and the author’s methodological choices.
As researchers, we have a responsibility to pre-emptively consider the broader social impact of our work.
Just last year, legislation in several states was put forth to limit gender expression and gender-affirming health care. Legislatures in Arkansas, Texas, and Alabama have cited or alluded to a small, observational study of parent reports that posits gender exploration among adolescents is contagious and can be described by the term “rapid-onset gender dysphoria,” which is not an established term in psychiatry or medicine. Severe concerns raised about the study methodology led to heavy revision of the original paper and retractions of similarly flawed related papers. A larger clinical study also failed to identify a novel phenomenon matching the description of rapid-onset gender dysphoria in adolescents. Still, the term continues to be used in legal proceedings intended to restrict freedom of gender expression.
Science is powerful: Research findings have and continue to influence the structure of our society, and the structure of our society influences our research. As researchers, we have a responsibility to pre-emptively consider the broader social impact of our work. While we’re accustomed to performing supplemental analyses to assess the robustness of our scientific conclusions, there is no parallel, explicit expectation to carefully and comprehensively consider the implications of our work on society.
Requiring such analyses may be a valuable standard to develop. They could include identification of communities potentially impacted by the work, data on engagement with those communities, feedback provided, and how that feedback was incorporated. These analyses could explicitly specify and dispel potential misinterpretations of the study upfront, and prompt more interdisciplinary collaboration and appreciation. Formalizing these assessments could improve communication and trust between scientists and the public, increase the applicability of research, and reduce both the risk that pervasive biases influence the research and that the research unintentionally fortifies pervasive biases.
Unlike a benign, silly association between cuffed jeans and bisexuality, associating risk-taking behavior and fecundity with bisexuality may both stem from and reinforce common negative, harmful stereotypes of bisexual individuals. Given the risk for harm and concerns about the study’s methods, this bi+ologist isn’t convinced that bisexual genes and risk-taking behavior are indeed “cuffed.”
Alexander Viloria Winnett is an eighth-year M.D.-Ph.D. student in the UCLA-Caltech Medical Scientist Training Program.
It is not the studies themselves that might reinforce stereotypes; it is the interpretation of the results that would have that potential. The only requirement for researchers is that they use accepted data gathering and data analysis methods and should be prepared to defend their work. What the author is suggesting is that researchers delimit their designs or prejudge their conclusions in order to avoid antagonizing marginalized groups. Such pre-judgement or delimiting is completely contrary to the aims and methods of science and smacks of political pandering.