Limitations of the Male/Female Binary for Studying the Influences of Sex‐ and Gender‐Related Factors on Health

ABSTRACT

Momentum has been building for several decades around the value of incorporating sex and gender considerations in biomedical, clinical, and health research more broadly. In that period, there has been a proliferation of guidelines, policies, definitions, methods, and conceptual frameworks for doing so, which is both constructive and challenging: the diversity of concepts and methods generates knowledge that highlights different aspects of the phenomena under study, but at the same time, it can create inconsistency and fragmentation around the operationalization and interpretation of research attending to sex and gender considerations in health. A male–female binary approach to examining how sex and gender influence health has predominated in many domains, and although this has value for helping to identify health disparities related to sex and gender, there are also some important limitations of an uncritical overreliance on male–female comparisons; three case studies from the biomedical literature are used to help illustrate some of these limitations. Ultimately, there is no single correct approach to addressing sex and gender in health research. I contend that the most crucial element is that researchers need to bring careful and critical attention to the incorporation of sex and gender considerations in ways that are appropriate for their research context and understand and articulate the limitations of their chosen approaches.

1. Introduction

Calls to account for sex and gender in health research began to emerge in the 1970s and 1980s, arising out of the women's health movement and second‐wave feminism, with growing recognition of the ways that the neglect of women in medical research had fostered gender disparities in health outcomes (Criado‐Perez ²⁰¹⁹). Starting in the mid‐1980s, the inclusion of women in health and medical research began to be encoded in policy with initiatives such as the Black Women's Health Imperative (1984), the Inclusion of Women and Minorities in Clinical Research policy (1986), establishment of the NIH Office of Research on Women's Health (1990), and the 1993 NIH Revitalization Act which mandated inclusion of women in all clinical trials (Office of Research on Women's Health ²⁰²¹).

In the 21st Century, exhortations to address sex and gender in health research have only become stronger and more frequent and are increasingly directed at researchers outside of clinical research, including those doing basic experimental laboratory research. In 2000, the Canadian Institutes for Health Research were established, including the Institute for Gender and Health, and in 2009, the CIHR began requiring all applicants for funding to respond to questions about their integration of sex and gender considerations in the proposed work (Johnson et al. 2014; Sharman and Johnson ²⁰¹²). In 2014, the US National Institutes of Health signaled that they were moving to require all NIH‐funded research to incorporate males and females (Clayton and Collins 2014); since 2014, that position has shifted in a variety of ways, narrowing the mandate to those studies that are “preclinical,” rather than to all biomedical research (Clayton 2016; Miller et al. ²⁰¹⁷; Tannenbaum et al. ²⁰¹⁶). Similar policies and guidance are also being applied in the European Union (European Commission. Directorate General for Research and Innovation ²⁰²¹; European Research Executive Agency, n.d.). It is also becoming more common for health‐related journals to have specific requirements for authors to address sex or gender in some way, as exemplified by the SAGER guidelines (Heidari et al. 2016, 2024; Peters et al. ²⁰²¹).

As a graduate student in the late 1990s and early 2000s, much of my training as a biomedical scientist took place in the midst of these developments, when sex and gender considerations were on the radar in biomedical research but were not yet widely operationalized, especially in pre‐clinical studies. My own work at that time was focused on understanding the molecular and cellular mechanisms underlying allergic sensitization in the respiratory tract, principally using experimental mouse models and some in vitro cell culture work (Ritz et al. 2000, 2002, 2004; Stämpfli et al. ¹⁹⁹⁸). At the time, I knew that sex and gender were relevant to the phenomena I was studying (e.g., it was clear in the immunological literature that gonadal hormones had a substantial influence on white blood cell function), and I heard the calls from multiple quarters to pay attention to sex/gender considerations in research. In the early 2000s, I attended a conference in Toronto entitled Biology As If The World Mattered, which was comprised mostly of feminist science studies scholars but also a few working scientists, including Deboleena Roy; like Roy (2004), I also identified as a feminist and felt strongly about the importance of gender equity in health and wanted my research to reflect my convictions. And yet if you were to look at my published experimental laboratory work in that period, you would be hard pressed to identify much of anything that suggested I was a scientist who thought sex/gender considerations were important. Although there is a stereotype that animal research relies disproportionately on male animals, the extent to which this is true depends to a substantial extent on the discipline. Although neuroscience research tended to prefer male animals, immunological research was less likely to report the sex of the animals used at all (Beery and Zucker 2011). In our lab we always used female mice, and this seemed to be the case for most other groups using similar models to the one we used; I can't speak for the other research groups, but in our case it was at least partly because of resource considerations—female mice could typically be housed at a higher density than male mice and so the housing costs were lower. But it was also because most other researchers in our field seemed to be using female mice, and so that was the norm in our biomedical niche. Little if any of my published work talked about sex or gender aside from mentioning the sex of the mice we used in the methods section.

This gnawed at me, and the more time that went by, the more I felt like a fraud for claiming to care about equity and health but not actually addressing sex/gender in my work explicitly. Yet every time I confronted this tension and tried to come up with an approach that I could use in my laboratory work, I seemed to come up short of anything that would substantively address these kinds of questions and was also feasible given the experimental context I was working in. I considered including male and female mice in my studies or measuring or supplementing or blocking hormones, but my intuition was that these didn't seem adequate to try and capture the complexity of sex or gender (though at the time I struggled to try and coherently articulate these concerns). I combed the literature voraciously trying to find approaches that felt like they were meaningful and appropriate to the context of laboratory research but was never able to find the kind of guidance I was looking for.

At one point when I was a junior faculty member, I remember expressing my frustration about the limitations I was perceiving in my efforts to address sex‐ and gender‐related issues in my animal models and cell culture experiments, and a more senior colleague suggested that I could do epidemiological work instead. But that would have required me to undergo extensive retraining and change my whole research focus and approach, and for me the whole point was that I wanted to figure out how to address sex and gender as an experimental scientist, not to become some different kind of scientist. This proved to be a crucial moment of insight for me, recognizing that there were probably plenty of other biomedical scientists who also wanted to better account for sex and gender in their work but without diverting the central focus of their work. Indeed, I could see among my colleagues that there were many scientists who were not necessarily averse to incorporating sex and gender considerations into their work, but did not particularly want to become researchers focused on questions of sex and gender; this is the community I have come to understand as the core audience I want to serve with my scholarly work in this area.

In the remainder of this paper, I'll offer a perspective that I hope will be valuable for health researchers in any discipline who acknowledge the value of addressing sex and gender considerations in their work (and who are responding to funder mandates to do so!), but who don't necessarily see themselves as sex and gender researchers. I will discuss a few case studies that I have found to be particularly useful in highlighting the limitations of relying on binary male–female framings for understanding sex and gender influences on health, and some commentary on the ways that I believe our scholarship could be substantially enhanced by taking a more contextual and mechanistic approach to understanding how sex and gender influence health. This paper aims to emphasize the value of bringing a more critical perspective to the use of male–female comparisons in biomedical research, and to consider designing experiments that investigate the influences of sex‐ and gender‐related factors rather than simply comparing sex/gender categories.

2. Limitations of the Sex/Gender and Male/Female Binaries

The conceptual distinction between sex and gender arose originally to try and help mitigate the problem of biological essentialism—the tendency to attribute any differences observed between males and females to innate biological causes without adequate attention to the ways that social structures and norms about gender shape and act as determinants of health. This often manifests itself in the tendency to invoke hormones or chromosomes or the physiology of reproduction as the principal causes of male–female difference, in the absence of any consideration about how things like gendered occupational segregation, gendered norms of behavior, gendered patterns of exposure to stressors or risk factors, gender biases of health care providers, or other social mechanisms shape and modulate health outcomes.

In the context of biomedicine, many agencies, institutions, and scholars offer different definitions and frameworks for making a distinction between sex and gender, where sex is usually understood as relating to the material biological basis for categorization into male and female related to reproductive functions, and gender is understood as a construct representing the various social, political, psychological, cultural, and other factors that influence the experiences of the individual as related to the gender norms and ideologies in a given social context. I think that there is value in making the conceptual distinction, particularly because it explicitly draws attention to the ways that male–female differences can be produced by social as well as biological causes, which is important for challenging biological essentialism. At the same time, I recognize that in most cases it is not actually possible to draw a sharp line between sex and gender; like the nature/nurture debate, when we scrutinize carefully we find that everything is both nature and nurture, both sex and gender at the same time, irretrievably entangled with one another. This entanglement of sex and gender, and what that means for researchers, is an active area of scholarship, and not my primary focus here; there is a range of existing and growing literature on this subject for readers who wish to take a deeper dive into these interesting and important theoretical and practical matters (e.g., Ah‐King and Hayward ²⁰¹⁴; Ashley et al. ²⁰²⁴; Blaffer Hrdy ¹⁹⁸¹; Bleier ¹⁹⁸⁴; DuBois and Shattuck‐Heidorn ²⁰²¹; Fausto‐Sterling ^{2005, 2019}; Hird ²⁰⁰³; Hubbard ¹⁹⁹⁰; Kaiser ²⁰¹²; Krieger ²⁰⁰⁵; Schiebinger ²⁰⁰⁴; Schmitz ²⁰¹⁰; van Anders and Watson ²⁰⁰⁶; Velocci ²⁰²⁴). Sex/gender entanglement was also the central focus of a 2023 Ernst Strungmann Forum conference (for which I was on the Program Advisory Committee and a participant), the proceedings of which will soon be available as a book (forthcoming 2025).

The other central binary is that of male “versus” female, which is the vastly predominant way that sex‐ and gender‐related considerations are operationalized in the biomedical and health literature: scientists include both male and female cells/animals in their experiments, comparing them to see if female cells/animals and male cells/animals respond differently to the same experimental intervention.

Certainly it's not hard to understand why using a male–female comparison is an appealing approach. For one thing, the idea of a male/female binary, and of male/female difference is deeply embedded in our cultural commonsense and thus it's a readily available framework for thinking about sex/gender. We are socially saturated with flawed and spurious binary messages about gender—pink is for girls and blue is for boys, men are aggressive while women are nurturing, girls are collaborative whereas boys are competitive, men are from Mars and women are from Venus. Male/female comparisons also seem easy to implement experimentally, at least at first blush—when you place your mouse order, just order male and female animals, right?

But we must be careful here on both counts. That commonsense also makes it very easy for experiments using male–female comparisons to be influenced by our unconscious stereotypes about gender: Martin (1991) demonstrates how this has happened in reproductive science, Fine (2011) undertakes an in‐depth examination of this in neuroscience, and Jordan‐Young and Karkazis (2019) talk about this with respect to testosterone. When we have been so thoroughly acculturated to the ideologies of male/female difference, we are more likely to accept data that conforms with those cultural biases, and less able to recognize the existence of those biases, ultimately rendering us less objective. Human beings are “remarkably prone to believing evidence that confirms [our] already existing beliefs” (Shotwell 2016), and so if we are going to use male–female comparisons in our experimental work, it's crucial that we are judicious and conscientious about how our underlying beliefs about gender might influence the way we analyze and interpret our findings.

Additionally, many scientists have already recognized that doing male/female comparisons is not actually such a simple matter after all. Resource issues are often a focus of that concern, and they are indeed a part of it—incorporating both males and females in many cases would require more animals, more reagents, more people, more time, and more money. But those are not the only challenges. In the context of experimental animal models, we can usually order or breed male and female mice and put them into our experiments, but the conditions under which they are housed in the laboratory may well create artifacts that distort the findings; for example, the stressors a mouse experiences by being housed 4 to a cage may be quite different if they are 4 females than if they are 4 males due to some of the social dynamics that occur with laboratory mice. Where the research involves transformed cell lines, it is literally not possible to include both male and female cells, at least not in any scientifically defensible way; the processes for isolating or producing and maintaining these cell lines renders any attempt to compare them as representatives of “male” and “female” meaningless (Ritz 2017). When working with primary cell cultures it is possible to include cells from female and male subjects, but even still, the highly artificial nature of the in vitro environment means that there are many important aspects of what we mean by “sex” that are not present in a flask or petri dish, and so such model systems only represent some sex‐related factors (for those readers who would be interested in reading more about this, I've elaborated on these concerns at length in a previous publication; Ritz ²⁰¹⁷).

But even aside from these kinds of practical concerns, there are conceptual concerns about relying heavily on male/female comparisons as a way of understanding how sex and gender influence health. Male–female comparison implicitly places male and female in opposition to one another, expecting difference, and conceptually marking them out as opposites; this is something we do culturally all the time (see examples above), and tends to promote dichotomous interpretations of male–female comparisons even when it is not well justified by the data itself. When a male–female comparison of a health‐related measure is made and a statistically significant difference in means is found, in most instances the distributions overlap to at least some degree; there is often a greater degree of within‐sex category variation than there is between the sex categories, the distributions are very rarely truly dimorphic, and the effect size is rarely described. However, it is unfortunately very common for researchers to interpret and describe the findings as though they are dimorphic, even when there is substantial overlap in the male and female distributions. What this means is that if we uncritically rely on male/female comparisons, we may well end up making generalizations about these categories that do a disservice to subgroups of people within both categories, as well as failing to account for people whose embodiment does not align with the normative categories themselves.

Note however that I'm not saying that male and female aren't often useful categories or that there is anything inherently wrong with making a male–female comparison in research. From a biological perspective, sexually‐reproducing species require a complementarity of gametes, and these different gametes are associated with certain anatomical structures and reproductive physiologies in sexually‐reproducing plants and animals. But although the gametes involved are binary (ova vs. sperm), and there are certain traits that are often used as markers to assort people into male or female categories (penis, testes, ovaries, uterus, vagina, vulva, X or Y chromosomes etc.), this doesn't mean that the construct of sex is strictly binary. The constellations of cells, tissues, molecules, structures, and pathways that we associate with “sex” do not fall into two neatly divided populations nearly as often as we think they do (Clancy 2024; Dea ²⁰¹⁶; Fausto‐Sterling ^{2000, 2012}).

Over the years there are 3 case studies that I have found to be particularly useful in illustrating the limitations and dangers of being overly reliant on a simple male–female binary comparison as our principal approach for attempting to understand how sex‐ and gender‐related factors can influence health.

2.1. Case Study #1: Ferritin Depletion in Frequent Blood Donors

In 2017, Goldman et al. (2017) published a study examining iron status in frequent blood donors in Canada, which found that iron depletion (as measured by plasma ferritin levels) was more prevalent among women than among men. For first‐time donors and those who had not given blood in at least a year, iron depletion (as indicated by a ferritin level of less than 12 μg/mL) was present in 9.6% of women under the age of 46, 5.5% of women over the age of 46, and 0.5% of men. For those who had donated 1–3 times within the previous year, the prevalence of low ferritin levels went up to 50.4% for women under 46, 20.3% for women over 46, and 6.8% for men. Among donors who had given 4 or more times in the previous year, the rates were 48.8% for women under 46, 37.2% for women over 46, and 27.3% for men.

These results formed the basis for a change in policy at the Canadian Blood Services to modify the recall interval for blood donation based on sex/gender category. Prior to the 2017 study, all donors who met the eligibility criteria for donation were permitted to donate every 56 days. Based on the findings in the 2017 study, that policy was changed, allowing most male donors to continue donating every 56 days (the minimum hemoglobin threshold for male donors was increased to 130 μg/dL), but the recall interval was lengthened to 84 days for all female donors.

This is a noteworthy example to me because it highlights what I see as an important disconnect in the application of category‐specific interventions based on male–female comparisons. Indeed, it is true that the data in the Goldman study and in other previous studies all showed that the prevalence of iron deficiency in blood donors was higher among women than among men (Baart et al. 2013; Cable et al. ²⁰¹²; Rigas et al. ²⁰¹⁴; Salvin et al. ²⁰¹⁴). The problem, however, is that differences in prevalence between the two categories do not mean that the individuals within each of those categories are at similar risk. And indeed, a close read of the data reported in the Goldman study very clearly shows that not all women are at equal risk of ferritin depletion. The prevalence of iron depletion was higher among women under 46 than among those over 46—which we can plausibly speculate is related to menstrual losses and menopause, as was seen in the Danish (Rigas et al. 2014) and Australian (Salvin et al. 2014) studies (which specifically analyzed by menstrual status). At the same time, among male donors—who are still allowed to donate more than 4 times per year provided they meet the minimum hemoglobin levels—27.3% had ferritin levels below 12 μg/mL. In other words, more than 1 in 4 men who are frequent donors will continue to be at risk of iron depletion (I think it's also worth noting that among women of all ages who were donating 4 or more times per year, approximately 20% had normal ferritin levels, and could continue to donate safely on a 56‐day recall interval).

It is worth noting here that during blood donation, the same volume of blood is taken from all donors (approximately 450—500 mL) regardless of body size. But of course, the total blood volume a person has is proportional to their body size, at about 8% of body weight—thus, a 50 kg person (the minimum weight for a blood donor in Canada) would have approximately 4 L of blood compared to a 100 kg person who has about 8 L. A single blood donation thus represents about 12% of the smaller person's blood volume, but only 6% of the larger person. Thus, a smaller donor is losing a larger fraction of their total body iron than a larger donor with each donation, with obvious implications for the risk of iron depletion. Since women on average are smaller than men on average, it certainly makes sense that the average woman is going to be at higher risk than the average man, but that does not mean that all women and all men are at the same risk. Indeed, the Danish study (Rigas et al. 2014) confirmed that weight was a significant predictive factor for the risk of iron depletion for all groups—men, premenopausal women, and post‐menopausal women.

Functionally speaking, the updated CBS recall interval criteria mean that a 100 kg post‐menopausal woman is treated as though she is at the same risk for iron depletion as a 50 kg woman with heavy menstrual losses, and as though she is at higher risk than a 50 kg man. Both of these seem unlikely given the relevant mechanisms mediating iron depletion in this context, which have been documented in other published papers. In its current form, the determination of recall interval based on gender category means that some men at risk of iron depletion are not protected while some women are being extended protection that they do not need; if the policy on recall interval were revised to address the mechanisms underlying the risk for ferritin depletion (particularly menstrual status and weight, but could also consider other factors like meat consumption) it would allow the agency to direct the protective mechanisms to the individuals that need them regardless of gender category.

2.2. Case Study #2: Sex‐ and Gender‐Related Influences on COVID‐19

In the early days of the COVID‐19 pandemic, data began to emerge in a number of jurisdictions that suggested that men seemed to be at higher risk of severe COVID and death than women (Grasselli et al. 2020; Jin et al. ²⁰²⁰). This observation was taken up by both the lay media and the scientific community, with speculation about the causes of such a disparity and studies undertaken to examine it more closely.

Much of the discourse around this focused on biological factors—particularly about sex‐linked genetic factors and hormones associated with sex, especially testosterone (Gebhard et al. 2020; Giagulli et al. ²⁰²¹; Peckham et al. ²⁰²⁰; Pivonello et al. ²⁰²¹). And certainly, steroid hormones (including androgens, estrogens, and progestogens) do have the ability to modulate immune response and warranted investigation. At the same time, it is important to note that the reports of disparities in morbidity and mortality were not seen across all jurisdictions, population sub‐groups, and across time, which is what would be expected if the differences were due principally to sex‐linked biological factors (Danielsen et al. 2022; Krieger et al. ²⁰²⁰).

There were many studies that tried to examine these disparities, many of which relied heavily on male–female comparisons. A notable early example was published in Nature in August 2020 by Takahashi et al. (2020), entitled “Sex differences in immune responses that underlie COVID‐19 disease outcomes” where they compared immune parameters in male and female COVID‐19 patients with moderate disease. In the paper, they report that “the immune landscape in patients with COVID‐19 is considerably different between the sexes,” with “male patients had higher plasma levels of innate immune cytokines such as IL‐8 and IL‐18 along with more robust induction of non‐classical monocytes,” while “female patients had more robust T cell activation than male patients.” The authors concluded that the findings “provide an important basis for the development of a sex‐based approach to the treatment and care of male and female patients with COVID‐19,” and suggested that “vaccines and therapies to increase T cell immune responses to SARS‐CoV‐2 might be warranted for male patients, whereas female patients might benefit from therapies that dampen innate immune activation.”

The paper examined a huge range of immune parameters, but the abstract specifically notes that IL‐8, IL‐18, non‐classical monocytes, and T cell activation were different in men than in women. When one looks at the data, though, the picture is much less dimorphic than these descriptions suggest. Fortunately, the figures themselves show individual data points for each subject in the study, allowing the reader to appreciate and evaluate the distributions for themselves. In examining the graphs, the visualization of the data makes it clear that the differences they are talking about when they draw these distinctions are statistically significant differences in means, but the distributions for all of these parameters overlap considerably between male and female patients. For 4 of the 5 parameters, the difference in means appears to be principally driven by a small number of outliers, with the main body of the distributions largely overlapping for males and females. These are shown in the original paper in figures 1c (panels for IL‐8 and IL‐18), 2c (panel for ncMono), 3c (panel for CD38+ HLA‐DR+ CD8+ cells), and 3e (panel for PD‐1+ TIM‐3+ CD8+ cells) if you wish to peruse them for yourself. When I look at these figures, what I see is that most male and female patients are actually quite similar, with a handful of outliers; the most substantial difference seems to be in the case of IL‐18, where a subset of the male patients (approximately 1/3) had IL‐18 levels that were above 120 pg/mL, while the rest of the male subjects and most of the female subjects were below that level.

This is not to say that the differences Takahashi et al. observed are not valuable and interesting, and these observations could indeed help us to understand why men seemed to be at higher risk for severe symptoms and death due to COVID‐19. But when you look at the data distributions it is clear that it was (pardon the phrase) not all men. The same was true with respect to the evidence of T cell activation in women—a small number of female patients had higher T cell activation, but most women and men had very similar T cell activation. It may be useful to try and understand whether the male patients with elevated IL‐8, IL‐18, and ncMono were the ones with enhanced symptoms, or if perhaps there was any correlation of T cell activation with protection from severe COVID. But when I look at this data, I do not see a basis for treating men and women differently. If anything this data indicates that most men and women should be treated the same, and if these immune parameters are indeed mediating increased risk for severe disease, then we should be using those parameters themselves to assess risk.

Although I'm singling out one example here, I don't really mean to pick on Takahashi et al. This kind of interpretation of male–female comparisons is absolutely rife in the medical literature: a difference in means is observed between female and male, the language used describes them in absolute terms as “different” or “dimorphic,” and conclusions are made suggesting that differential treatment based on category might be warranted. It's a troubling pattern, and I would argue that it's also dangerous. When data distributions overlap substantially for females and males, targeting treatment based on sex category would mean that lots of people would be either denied treatment that would be helpful to them or provided with treatment that would not benefit them. Now, in cases where there is actually a dimorphic sex difference—that is, there is little or no overlap between the male and female distributions, then probably there is a stronger case for differential treatment, and in those cases—especially when the consequences are severe or high stakes, it may indeed make sense to target treatment categorically as a provisional approach while more research is conducted. But once we understand the mechanism(s) underlying a male–female difference, I contend that we would still be better off targeting the mechanism rather than the category.

2.3. Case Study #3: Occupational Ergonomics

In the early 1990s, one of my mentors in sex/gender and health—Karen Messing—was doing work with labour unions in Quebec related to gender and ergonomic issues in workplaces. In one instance, her team studied the gendered ergonomic implications of work undertaken by mechanics responsible for rebuilding diesel engines, 10 of whom were men, and one of whom was a woman (Courville et al. 1991). In an initial assessment of the workers' perceived exertion, difficulty, and pain experienced during different aspects of the procedure indicated that the sole female mechanic reported somewhat more pain and difficulty with certain aspects of the work than some of her male co‐workers (though male co‐workers did report varying experiences of pain and difficulty as well).

The engine rebuilds were normally conducted in pairs, and the female mechanic was normally paired with the same male colleague. The investigators made a number of biomechanical and morphological measurements of the workers (grip strength, height etc.), and found that—not unexpectedly—the female mechanic had less grip strength and upper body strength than her male counterpart. However, a detailed analysis of their work over the course of several days showed that only 4% of their time was spent doing work that required brute strength, and which did not seem to be associated with the pain experienced. One aspect that was substantially different between the two workers was the time spent in postures and positions that placed different kinds of strain on the joints (particularly the shoulder joint) due to their differences in height: the female mechanic spent considerably more time operating in a posture in which her arm was raised up above the shoulder and applying force in the operation of a tool due to her shorter stature; the same motion for her male counterpart did not have the arm above the level of the shoulder. As a result, the forces and strains on the shoulder joint were markedly different for those two mechanics doing the same job.

Notably, Courville et al. point out that although the female mechanic did indeed have lower grip strength and arm strength than her male counterpart, this was not likely to be the cause of the increased pain she experienced—her grip and arm strength were entirely adequate to perform the task. Her relatively shorter stature seemed to be driving the disparity. As a result, the ergonomists recommended providing alternative tools (e.g., with longer handles), or adjusting the height of the work surface relative to the diesel engine, so that the workers would be able to undertake the work at more optimal angles given their stature—in fact, they noted that the ability to use different sizes of tools and adjust the height of the work surface would benefit taller workers as well (and also noted that the tallest male worker reported back pain that was likely related to having to do some of his tasks in a bent position known to be a risk factor for back pain).

To me this is a powerful example for a number of reasons. First, given the disparity in grip and arm strength, it would have been very easy for the workers or researchers to slip into stereotypes about gender and strength and reach the erroneous conclusion that the female mechanic was simply not strong enough to do the work safely, but the investigators carefully resisted that bias; although the female worker was not as strong as her male counterpart on these measures, she did have sufficient strength for the task. Second, their analysis suggesting adaptations of the work environment for the shorter worker is not only relevant for the female mechanic in this case, but for any workers of smaller stature—and in fact, for all workers overall—workers of any stature could potentially be at risk of injury if their tools and posture related to the object of work create unnecessary strain.

3. How Can We Move Past the Male/Female Binary to Address Sex/Gender in Experimental Biomedical Research?

Although the case studies I've discussed highlight the limitations of uncritical overreliance on female–male comparisons for understanding the influence of sex and gender on health, I am not suggesting that we should not use female–male comparisons at all. Indeed, I think the case studies actually help to illustrate that male–female comparisons are very useful for helping to identify the existence of many sex‐ and gender‐related disparities in health. That being said, there are a number of things we should be conscious and careful about when we employ them. We must recognize the heterogeneity within the categories of “male” and “female” and attend to the distributions of the data. In many cases, the within‐group variation will be substantially larger than the extent of the mean difference between the groups. Moreover, many of the traits and factors we associate with sex are not static, but change and fluctuate dynamically across the lifespan, and in response to the environment. It's also valuable to recognize that sex is not a “cause,” and that “male” and “female” serve largely as proxies for mechanisms that we may not yet recognize or understand (Pape et al. 2024). “Being male” does not cause a decrease in antibody levels or an upregulation of GABA receptors or a delay in the flux of calcium ions across a membrane. A male/female comparison should really be considered as no more than an exploratory pilot study and interpreted in provisional, judicious, and cautious ways.

Scientists making female–male comparisons can be more conscientious about refraining from interpreting findings in essentialist, binary, or anthropomorphic ways. We can invite and entertain and articulate hypotheses that recognize the dynamic and complex nature of sex/gender, in ways that explicitly acknowledge that differences observed between males and females are almost always differences of degree, rather than absolute differences. Given that, we should report our data in ways that allow others to appreciate the variability within the categories and the overlaps between them (Ritz and Greaves 2022), as Takahashi et al. did in their study.

There are also other ways to address sex/gender aside from simple female–male comparisons. Even if we do not do anything differently experimentally than what we have always done, we can develop our expertise about the known influences of sex/gender in our scientific niches, and we can do a better job of explicitly identifying the limitations of our work with respect to sex/gender considerations.

Experimental scientists work with models all the time—in fact, one of the elements of our expertise as scientists is being able to choose an appropriate model to address a given question, and use that model responsibly with due regard for its strengths and limitations. No experimental model is perfect, and we use them anyway because they are useful for certain purposes. The key is not to over‐interpret or over‐extend our findings in light of the model's limitations, which scientists are usually pretty conscientious about. Addressing sex/gender in experimental biomedical research is really just another exercise in understanding the limitations of our models, and using and interpreting them accordingly.

I find Richardson's (2022) concept of sex contexualism particularly helpful here, which “recognizes the pluralism and context‐specificity of operationalizations of “sex” across experimental laboratory research” in order to rigorously design experiments that attend to the most relevant variables and mechanisms. In a sex contextualist approach, the researcher is called to contemplate how exactly sex is relevant to their research question, and design the experimental approach so as to query the mechanisms through which sex is hypothesized to influence the outcome of interest (Pape et al. 2024). For example, when I supervised a Masters student many years ago, she was using an in vitro model employing the Jurkat T cell line (a cell line derived from a male lymphoma patient) to examine whether the presence of estrogen modulated the relationship between nickel exposure and cytokine production by those cells. We were very deliberate in recognizing that our approach wasn't about “males and females”; we were using an immortalized cell line (which came from a male donor, but can't be said to “represent” males) and the hormone we were manipulating is not a “female hormone” but is present in all bodies. In going beyond a male–female categorical comparison and instead considering the influence of specific sex‐ or gender‐related factors, biomedical research will be better able to shed light on the mechanisms that drive health inequities.

At the end of the day, “gender conceptions…are an inevitable backdrop to the science of sex, and they can play a constructive role in science when they are subject to criticism” (Richardson 2013). Addressing sex/gender in our research isn't about developing a better understanding of gender and health, it's about developing a better understanding of health, period. At this crucial juncture in biomedical research, we must be attentive to the gender conceptions that we are bringing to our science, critique them, and nourish a science of sex/gender that promotes health equity for all.

Ethics Statement

The author has nothing to report.

Conflicts of Interest

The author declares no conflicts of interest.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.