In the waning months of the Obama Administration, the Departments of Justice and Education advised schools and colleges that gender identity discrimination was to be considered a form of sex discrimination covered by Title IX (the federal sex non-discrimination law that applies to all federally-funded educational programming, including competitive college sports). Within a month of taking office, Trump’s administration rescinded that compliance letter, in a stroke erasing any explicit protection for transgender student-athletes. Perhaps, though, where the Obama Administration really went wrong was in not going further to name sex-segregated sports as a source of “gender identity” discrimination.

What is gender discrimination and how do sex-segregated sports trigger it? In my book, Beyond Trans: Does Gender Matter?  I use the term “sex identity discrimination” instead of “gender identity discrimination” because I think it is more precise. The traditional understanding of sexism in the U.S. involves stereotypes about what a person can and cannot do because she is a woman or he is a man. Sex-identity discrimination involves such stereotyping, but it also involves judgments about who does and does not belong in the very categories of female and male. With this refinement in mind, we reveal how people who may not self-identify as transgender can still be subject to sex-identity discrimination when others perceive them as not adhering to the gender norms associated with being a “real woman” or a “real man.”

Because an athlete’s self-understanding of their sex-identity may be at odds with an administrator’s judgment, we should be very careful about how, if at all, we invoke sex classification in sports policy.

In Beyond Trans, I suggest a liberation from the rote presumption that maleness and femaleness always—or even often—matters. More often than not, gender isn’t clearly or rationally linked to specific policy goals. When sex classification is relevant, the onus should be on policy makers to provide a clear definition of the terms “sex” and “gender” within the contexts of the policy, a clear statement of the policy goal at issue, and a clear explanation of why it is necessary to invoke sex classification in order to achieve the specific policy goal.

One of the legitimate policy goals at stake in competitive sports is “fair play.” There is a longstanding presumption that permitting men and women to compete against one another in competitive sports, especially sports involving physical contact, would not be fair—that men would always have a competitive advantage over women based upon “biological” sex differences. But not all sex-related characteristics provide a competitive edge in sports. For example, it matters nary a bit whether you have a penis or vagina, a beard, or soft skin for how fast you run or how high you jump.

Instead, when we get more specific, we can say that certain sex-related physiological features, such as muscle mass, lung capacity, and aggression, are relevant to ‘fair play.’ These mutable characteristics are affected by the amount of functional testosterone in an athlete’s body, not by the presence or absence of genitalia.

The NCAA has followed the International Olympic Committee  to decree that athletes wishing to compete as female must prove that their testosterone levels fall within the normal range for women. If their testosterone exceeds levels for female competition, the athlete must compete against males or will be excluded from competition. By adopting this administrative policy, the NCAA is effectively using hormone levels instead of sex identity as its sorting criteria. Perhaps the most logical next step would be to get rid of the sex markers altogether, leaving a simple division into categories based on testosterone levels. Consider that some transgender women use estrogen therapy to achieve some of the secondary sex characteristics associated with female sex identity, such as an increased ratio of fat to muscle and the production of additional breast tissue, just as some transgender men use testosterone therapy to achieve some of the secondary sex characteristics associated with male sex identity. Other transgender people don’t use hormone therapy at all, and some people identify as genderqueer, gender fluid, or non-binary rather than “transgender.” Further, a not insignificant number of people are born intersex, meaning that their chromosomes, hormones, and/or physiology are not dyadic (unequivocally “male” or “female”). With all of these possibilities present in the field of play, can a male/female dichotomy work as a functional policy? Taking gender and sex identity out of the game may level the field significantly.

If we were to replace sex-segregation with testosterone levels in Title IX sports, we would, of course, not be making a trivial change. Part of the sense of “losing tradition” is tied up with losing the social aspect of gender-based affinity that is considered formative to young athletes: girls and boys and women and men working together toward common goals and competing to achieve them. But this social aspect claim holds most validity at the recreational level. When it comes to elite competition such as NCAA Division 1 sports and Olympic play, competition takes precedence over all else. While it seems simple, it comes down to the single question of whether gender matters—and if it does, when, why, and how? Replacing male and female competitive sports with hormone levels seems, on its face, radical, but the NCAA and other elite sports organizations have already begun rewriting the rules in favor of fair play.