What The Sports Gene Gets Wrong

As readers here know, I really like The Sports Gene by David Epstein, formerly of Sports Illustrated and now at Pro Publica. The book overall is excellent and has already motivated several posts -- here and here. In this post I pick a nit with respect to the book's treatment of testosterone and athletic performance with respect to determining who is judged eligible to compete in women's athletics, a complicated issue that deserved to be treated more comprehensively.

Back in spring 2012, students in my graduate seminar on Science and Society were tasked with coming up with a policy for determining who would be allowed to participate in the women's events in the then-upcoming 2012 Olympics in London. The case that we were looking closely at was of course that of Caster Semenya of South Africa, a female athlete whose treatment by the sporting community was embarrassing and undignified.

At that time the International Olympics Committee adopted a policy (following the lead of the International Association of Athletics Federations or IAAF) which used testosterone levels in and usage by the body as the criterion for determining eligibility for participating in women's events.The policy called for a panel of experts to make judgments in contested cases on the following (the policy is here in PDF):

The Expert Panel shall examine all available information and establish (i) whether the investigated athlete’s androgen level, measured by reference to testosterone  levels in serum, is within the male range, and if so, (ii) whether such hyperandrogenism is functional or not.

The Sports Gene notes this policy as well as a similar policy adopted by the NCAA and notes that "testosterone has been deemed the source of the male athletic advantage." Epstein observes that testosterone may not be all there is to the advantage, as women who are unable to process testosterone -- call androgen insensitive -- are actually over-represented in sports, "vastly" so according to Epstein.

It is here where I pick my nit. Unlike the in depth and nuanced exploration of reasons behind Jamaican sprinting success found in The Sports Gene, the role of testosterone in female athletic success -- which is just as complex, contested and nuanced is given short shrift.

Epstien sums up the science on this issue as follows:

No scientist can claim to know the precise impact of testosterone on any individual athlete. But a 2012 study that spent three months following female athletes from a range of sports--including track and field and swimming--showed that the elite-level competitors had testosterone levels that consistently remained more than twice as high as those of nonelites. And there are powerful anecdotes as well.

Fair enough, as far as it goes. But it doesn't go far enough.

In a critique of the IAAF and IOC policies published in 2012, Karkazis and colleagues (here in PDF) made a case against using testosterone as a criterion for determining eligibility for participating in women's sports events:

The new policies rest on the notion that the difference in athletic performance between males and females is “predominantly due to higher levels of androgenic hormones in males resulting in increased strength and muscle development” (IAAF 2011c, 1). Both policies rely in particular on testosterone levels as the mark of unfair advantage. Although it may be surprising, given that this is a popular belief and is stated as fact in both IAAF and IOC statements (IAAF 2011d; IOC 2011), the link between athleticism and androgens in general or testosterone in particular has not been proven. Despite the many assumptions about the relationship between testosterone and athletic advantage, there is no evidence showing that successful athletes have higher testosterone levels than less successful athletes.

Say what?

In sum, there is a great deal of mythology about the physical effects of testosterone and other androgens (Fausto-Sterling 1985; Jordan-Young 2010). Likewise, mental effects of androgens are often implied to give an additional boost to athletes, but placebo-controlled studies of testosterone show that increasing testosterone (above minimum functional levels) has no effects on mood, cognitive performance, libido, or aggression (Bhasin et al. 1996; Bhasin et al. 2001; Kvorning et al. 2006). Optimal levels of testosterone is one of many factors that is necessary for athletes to achieve their own “personal best,” but comparing testosterone levels across individuals is not of any apparent scientific value.

Karkazis et al. may indeed be on to something that the role of endogenous testosterone is vastly overstated. Perhaps instead the suggestive research by Christopher J. Cook and colleagues (replied on by Epstein) could be more correct in identifying endogenous testosterone as a key determinant of elite female athletic success. Clearly there are uncertainties and disagreements in the basic scientific understandings. As The Sports Gene well explains, such complexities are much the norm with respect to human athletic achievement.

Science issues aside, the IAAF and IOC policies may indeed represent a pragmatic compromise that is both legitimate and fair. Certainly the Karkazis et al. recommendation to used national legal definitions of what it means to be female is pragmatically untenable. Karkazis et al. do make a compelling case that the expert advisory process which developed the guidelines fell short in important respects, including its composition, procedures and scope.

As with the case of Jamaican sprinting success and east African long-distance success, the issue of sex differences and advantage in athletic competition is complicated -- more complicated than its treatment in The Sports Gene. Of course, this is but one nit in what is overall an excellent book, and one that will be required reading in my upcoming spring 2014 course.