Five-ring Circus

Gender Games

The Olympics has a new way to test whether athletes are men or women. Is it fair?

Caster Semenya of South Africa prepares before the start of the women's 400m during the Yellow Pages Interprovincial final on March 03, 2012 in Cape Town, South Africa.
Caster Semenya’s sex became a question mark amid speculation about her masculine-looking physique

Photograph by Shaun Roy /Gallo Images/Getty Images.

The International Olympic Committee has a new rule this year for deciding who gets to compete as a woman. The IOC’s hope is to avoid a spectacle like the one that followed South African runner Caster Semenya’s victory in the women’s 800 meters at the 2009 World Championships. Semenya’s sex became a question mark amid speculation about her masculine-looking physique. She wasn’t allowed to compete for almost a year, and then she was reinstated without a clear explanation. Nobody wants to repeat that. The question, though, is whether the science is clear enough to justify the Olympics’ latest approach—and the answer is probably no.

Here’s the IOC’s plan: Women who test in the male range for testosterone, and whose bodies respond to the hormone, may not be eligible to compete as females. “May not,” because these women may also be allowed to lower their testosterone levels medically, as some say Semenya is now doing, though the IOC hasn’t confirmed that. The committee does not specify what testosterone level is disqualifying, in part because individuals’ measures can fluctuate. “We’ll leave those decisions with the experts,” Arne Ljungqvist, the chairman of the IOC’s medical commission, told the New York Times.

The goal is to level the playing field, preventing people who identify as women, but have an unfair man-like advantage, from dominating women’s sports. Historically, sports authorities have required female athletes to be examined naked or to undergo chromosomal testing, in which they looked for a structure called a Barr body, indicating the presence of more than one X chromosome. At races, Semenya grew “accustomed to visiting the bathroom with a member of a competing team so that they could look at her private parts and then get on with the race,” as Ariel Levy wrote in The New Yorker.

Bad idea. Women with a disorder called congenital adrenal hyperplasia can have ambiguous genitalia, including enlarged clitorises that may look like small penises. They might fail an exam for sex even though they have ovaries and a uterus, and generally identify as females. Then there are people with a condition called complete androgen insensitivity syndrome. They are chromosomally XY, the normal makeup for men, but their bodies don’t respond to testosterone. So they develop female genitals and breasts, but have testes, not ovaries. They would fail a test of femaleness based on chromosomes, though they also generally live their lives as women. There are other conditions with other complexities, and it all confounds the idea of a test based on one simple factor.

That surely includes the latest testosterone standard. The IOC zeroed in on testosterone levels, it said, because the performance difference between male and female athletes may be “mainly” due to the hormone. It’s true that the normal female and male ranges for this hormone are clearly distinct. Yet women with conditions like congenital adrenal hyperplasia and complete androgen insensitivity syndrome sometimes test in the male range. That includes small but significant numbers of female Olympic athletes. In 2000, British endocrinologist Peter Sonksen analyzed testosterone levels for around 650 Olympic athletes representing a random selection across sports. His results haven’t been published, but he gave me a scatterplot of his data, and it shows almost 5 percent of the women testing in the male range for testosterone and more than 6 percent of the men testing in the female range. In other words, T levels are not diagnostic for sex.

The International Olympic Committee apparently thinks that’s OK. The point is to identify women whose high T levels have “performance-enhancing effects” with regard to “strength, power and speed” that give them an unfair advantage over other female competitors. (They will make allowances for athletes whose bodies do not respond normally to the hormone.) It’s widely accepted, after all, that testosterone builds muscle mass.

But little research supports the view that elite athletes with naturally high testosterone levels fare better in sports than those with lower levels. “We really don’t know,” says Allan Mazur of Syracuse University. “Logical assumptions about hormones don’t always turn out to be true.” In Sonksen’s study, more than 25 percent of the male Olympians had T levels below the normal male range. If testosterone is all-important for sports, how could that be?

And even if naturally high testosterone does offer an advantage, why should women who have it necessarily be treated differently from athletes with other mutations or medical conditions that confer, say, greater endurance or extra height? Katrina Karkazis of Stanford and Rebecca Jordan-Young of Barnard have persuasively made this case, noting that some high-level athletes have genetic or cellular variations that increase blood flow to their skeletal muscle or improve their aerobic capacity and endurance. Some professional basketball players have a condition called acromegaly, which makes them unusually tall. Or consider Secretariat, the thoroughbred who ran the Kentucky Derby in under two minutes, whose heart was more than twice the normal size. Olympic medalists and Triple Crown winners just don’t have garden-variety physiologies. How is the advantage to a female athlete of being 6-foot-3 different from the advantage of having high testosterone if both are rooted largely in her genetic makeup?

This is where the IOC’s task starts to seem harder than pulling off a triple twisting Yurchenko. You can see the temptation to rely on the legal classification of an athlete’s sex in her home country rather than going with a universal test. Yet in the knock-down, cheat-prone world of elite sports, some biologically based criteria are surely needed. Here are some distinctions we could fairly draw: People who are chromosomally XX and living as legal women should compete as female Olympic athletics, no matter what their bodies look like and no matter their testosterone levels. Under the current policy, XX-ers with naturally high testosterone could be ineligible to compete as women, and that seems wrong. Some XY-ers should also be deemed women for the purposes of sport, like those with complete androgen insensitivity syndrome and many with partial androgen insensitivity syndrome, too. Yes, this last category would entail some hard judgment calls, but at least it’s a smaller group. And rather than a narrow focus on testosterone, a more encompassing algorithm that takes into account chromosomes, genitals, gonads, and hormones seems like the only plausible approach. It’s nice to imagine that a simple test could also be a fair test. But in sports as in life, that’s often just not true.

Read the rest of Slate’s coverage of the London Olympics.