Riddles of Sex Difference Science: Q&A with Sarah Richardson

In a new open access article recently published in the journal Philosophy, Theory, and Practice in Biology, GenderSci Lab Director Sarah Richardson extends and develops her framework of “sex contextualism” (Richardson 2022). Sex contextualism offers an alternative to commonplace essentialist approaches to interpreting sex-related biological variation in biomedical research (see: “No Sex Without Context”; Research Handout; Teaching Slidedeck).

The article, “Sex and the Riddle of Variability,” focuses on a scientific debate over differences in trait variability between male and female laboratory rodents. Through an analysis of this debate, Richardson demonstrates how researchers make judgments about “what counts as sex” when “designing, analyzing, and interpreting sex-split studies using rodents.” These judgments, she argues, result in “a context-specific construct of sex” (Richardson 2025, p. 2). This way of thinking about sex contrasts with a common understanding of sex as a simple variable revealed by comparing males and females – an understanding encoded in research policies that increasingly encourage scientists to include male and female organisms in research.

In this Q&A for the GenderSci Lab blog, Richardson digs into sex contextualism, the “riddle of variability,” and implications for research design, policy, and public understanding of sex difference science:

Rather than thinking of “sex” as a cookie cutter… the sex contextualist understands sex as an artisanal practice, composed of the elements at hand that make sense for the research question and materials.

- Sarah Richardson

GSL: What do you mean by the “riddle of variability”?

Richardson: Scientists engaging in laboratory research seeking to locate sex differences in a trait or outcome of interest face a riddle: which variation counts as “sex” differences, and which is background? 

The simple example of the estrous cycle, which I focus on in this article, reveals the riddled nature of sex in biomedical laboratory research design using model systems. The rodent estrous cycle is clearly a sex-related trait, reflecting changing levels of ovarian hormones in relation to ovulation. If estrous state is part of what the researcher thinks is “sex,” then research design requires including an adequate number of cycling animals to statistically power an analysis of each of 3-4 estrous phases, measuring hormonal status, and analyzing for variation across these differently-sexed states. But if, as many researchers who advocate for sex inclusion research policies have argued, estrous cycle does not need to be routinely measured in sex difference designs, what counts as “sex” could be a simple comparison of the mean trait value for males and females. 

Riddles are often solved when we challenge our preconceived categories. Take, for example, the following: A basket contains two eggs. Two people each take an egg each, resulting in the basket containing one egg. How is this possible? The answer is that the second person takes the egg with the basket, playing with the reader's assumption about what counts as "taking an egg." (Thanks to Mia Miyagi for this fun example of a riddle!)

I argue that this riddle of variability is not limited to the case of the estrous cycle. It is a profound feature of biomedical reasoning about sex differences using experimental designs in which binary sex categories are confounded by other forms of variation. In practice, the riddle is “solved” by human judgments about what “counts” as sex and what does not. Thus, as I write, “sex does not simply ‘fall out’ of comparisons of males and females as a coherent biological variable that causes differences between individuals or subgroups. It is operationalized through a series of assumptions about what ‘intrinsic sex’ is” (Richardson 2025, p. 10).

GSL: Is one sex more variable than the other, and why does it matter?

Figure 1 The coefficient of variation measures the dispersion of variance around the mean for each population or subgroup.  Image from: https://datatab.net/tutorial/levene-test​

Richardson: This question has a long history, intertwined with cultural ideas about differences between human men and women. For example, Darwin famously claimed that males are more variable than females, which he thought explained “man’s attaining to a higher eminence, in whatever he takes up, than can woman.” Meanwhile, medical researchers have long asserted that females are more variable than males due to ovarian hormone cycles, in this case making the more homogenous males more ideal research models.

In everyday laboratory practice, group differences in variability are a practical issue for research design, and there is no consistent pattern as to which sex is more variable than the other across different traits. When using male and female animals in laboratory experimental research with the intention to compare the two groups to ascertain sex differences, it is vital to account for possible quantitative differences in the amount of variability within a sex for a particular trait. Otherwise, differences in means for a trait could be misinterpreted, leading to false positive reports that overestimate the presence and magnitude of sex differences (Figure 1). Complicating matters, a range of factors related to or confounded with “sex” might influence a group’s variability for a given trait, including group vs. individual housing, rearing, reproductive behavior, ovarian hormone status related to estrous phase, age/developmental stage, and body size.

GSL: If this is an old debate, why has it recently revived?

Richardson: As funders and institutions introduce policies which require researchers to include male and female organisms in all research and to disaggregate and compare data by sex, such as the  NIH with its “Sex as a Biological Variable” policy and, more recently, the UK’s MESSAGE project, these types of comparisons between sexes have become more and more common. These policies have good intentions, seeking to advance women’s health and increase the rigor and replicability of preclinical research. But they also create a scenario where a one-size-fits-all study design is applied to diverse models, species, and biological systems. Getting at sex differences by comparing male and female organisms in the same research design sounds intuitive, but it’s actually not clear whether this is always the best way to study sex-related biology. On one side of the debate are researchers who say that, on balance, it’s still better to run these sex comparisons than not. On the other side are researchers who claim that these binary research designs risk obscuring meaningful sex-related variation, favor the reporting of differences in means between sex categories over the more direct study of mechanisms such as hormones, and yield results that are invalid or of limited clinical utility.

GSL: When comparing male and female biomedical data, should researchers include the estrous cycle in their operationalization of sex?

Richardson:  Sometimes the estrous cycle will be included as part of the construct of sex, and other times it will be excluded. As I write in the paper, “A sex contextualist approach does not tell us whether one way of constructing sex as a biological variable is more objectively true, sound, or empirically adequate than another,”; hence, “both choices – to exclude or include estrous cycle measures in sex-difference research design – are examples of sex contextualism in practice” (Richardson 2025, p. 9).

The key to the contextualist approach is that sexes are not interpreted or presented as essential categories or natural kinds, but rather, as operationalizations by the researcher. Rather than thinking of “sex” as a cookie cutter that the researcher applies in any kind of context to produce a single form, the sex contextualist understands sex as an artisanal practice, composed of the elements at hand that make sense for the research question and materials. 

As should be clear by now, one thing that distinguishes sex contextualism from other frameworks for defining sex categories in biomedicine is that rather than specifying what construct of sex a researcher should use, a sex contextualist expects that definitions of sex will be pragmatic, context-specific, and often plural. Making the contextual dimensions of how researchers use the concept of sex visible is important because it allows diverse users of research findings, from other researchers to pharmaceutical investors, science journalists, clinicians, and the general public, to better evaluate features such as generalizability and proximity to mechanistic explanation of any inference about sex differences.   

GSL: For non-researchers who hear of biological sex difference claims in areas such as sports performance, STEM professions, or leadership and risk-taking propensity, what insights does sex contextualism offer?

Richardson: Science literacy in this area requires critical reading of these many contested claims. Needless to say, the existence of a sex difference should in most cases be irrelevant to the distribution of social goods. Many popular claims about human sex differences represent selective appeals to the authority of biological science to advance a particular political viewpoint. There are many useful touchstones for gaining this literacy – the most important of which are: (1) to acknowledge the history of bias in studies of sex differences with a tendency to overstate their magnitude and uncritically incorporate gendered social stereotypes and assumptions (as in Darwin’s views about male brilliance, or associations between maleness and aggression and femaleness and parental care) into research design; (2) to recognize that for almost all sex differences there is significant overlap between males and females, with overwhelming similarities between the groups; (3) that sex difference findings often change across time and place, showing that they are sensitive to context and, in particular, to gendered social norms and expectations; and finally, (4) that there is considerable variability, fluidity, and diversity among women and among men, including a not-insignificant number of people who fall outside of this binary on one or more traits. 

While I developed the concept of sex contextualism thinking specifically about how to understand sex in biomedical laboratory research, I do believe sex contextualism adds to our broader toolkit for critically assessing scientific claims about sex. Across studies, researchers conceptualize “sex,” “male,” “female,” “sex difference,” and sex-related variables in plural and sometimes incommensurate ways. Thus, a finding of a “sex difference” should be understood as reflecting a context-specific (meaning species, laboratory, discipline, etc.) definition of that construct. Consumers of sex difference science need to ask, “How is sex operationalized in the study?” which might include questions such as, “In what model system? In what research context? Do study constructs and inferences uncritically incorporate gendered stereotypes or assumptions? What variables were included? How was variability measured and represented?”

GSL: Can you tell us more about the NSF-funded project of which this paper is a part?

Richardson: Last year, I began a three-year National Science Foundation-funded project to develop additional case studies of how sex is plurally and contextually operationalized in laboratory research models. The project responds to a great deal of interest among researchers in how to answer calls to consider sex in their preclinical research, how to interpret findings of apparent sex differences, and how to responsibly communicate their results. In addition to close readings of studies tackling questions related to sex – such as that considered in the “Sex and the Riddle of Variability” paper, “Is estrous cycle necessary to model when studying sex differences?” – the project involved site visits and interviews with scientists.  

Sadly, the NSF funding, which supported a postdoctoral scholar, graduate student research assistants, research travel, and consultation with experts in the field, was revoked on May 15, 2025, as part of a mass termination of research grants to Harvard University. We are hopeful that legal challenges to these arbitrary cancellations will be successful and that we will be able to restart the work soon.

Original open access research article: Richardson, S. S., (2025) “Sex and the Riddle of Variability”, Philosophy, Theory, and Practice in Biology 17(1): 3. doi: https://doi.org/10.3998/ptpbio.5492 

SUGGESTED CITATION

Richardson, S.S, (2025). Riddles of Sex Difference Science: Q&A with Sarah Richardson. GenderSci Lab Blog. https://www.genderscilab.org/blog/riddlesofsex

STATEMENT OF INTELLECTUAL LABOR

Sarah Richardson drafted and revised the blog post. Marina DiMarco, Mia Miyagi, and Hannah Niederriter provided feedback and edits.