Last week, a woman who is the poster athlete for perseverance and grit had her career and livelihood ripped away from her. She did nothing wrong. She did not use performance enhancing drugs like many of her competitors. Instead, she has had her private life blown open, with details about her inner organs discussed around the world. She has stood tall throughout it all. A role model for young men and women to emulate. Caster Semenya was met with a decision from the Court of Arbitration of Sport (CAS) that she can no longer compete in the women’s category at the Olympic level. Semenya is an athlete with  XY chromosomes with a  Difference of Sexual Development (DSD).

The decision stings. It leaves everyone who reads it with a visceral feeling of unease. It’s unfair. But that doesn’t mean that it was the wrong decision.

If you read about the Semenya case, you’re going to be met with extreme positions on both sides of the aisle. You’ll hear cries of unfairness, of why testosterone doesn’t matter that much, or why don’t we regulate height or weight or some other biological advantage? Some of these claims might be true, many are hyperbole at best.

The easy choice in this debate is to gravitate towards emotional rhetoric. To splash our anger across the Internet. The hard choice is to take the time to understand the complexity of the topic. To challenge ourselves to wade into the unclear world of sex and performance; a world that makes it clear that this might be the most difficult issue of modern sport.

To make sense out of this issue I want to try to answer a few questions that are at the heart of the controversy and some of our misunderstandings. Contrary to the title, this is not the be all end all Semenya document. But what I hope it does is clear up parts of the issue and get everyone to realize that quick takes and superficial analysis will not cut it on this topic.

  1.  Do we need to separate men and women in sport?
  2.  How do we divide sport currently and why do we utilize testosterone as a marker?
  3.  If it’s such an advantage, why isn’t Semenya way faster than all other females?
  4.  What is a DSD and why are people questioning Semenya’s sex?
  5.  What about other genetic advantages like Michael Phelps “lactic acid” or Shaq’s height?
  6.  Should we be concerned with the IAAF’s ruling?
  7.  Is this Discrimination?

The Separation of Men’s and Women’s Sport

We don’t separate by gender in math or science. Nor do we do so in just about any other field. Then why do we do so in sport? Shouldn’t we want equality?

The fastest 400-meter time run by a woman in all of 2018 (48.97 seconds). Over 5,000 males ran faster in the same year. That’s how it is across athletic domains. It’s a harsh reality. If we didn’t divide by sexes, there would be no women’s sport at the professional level.

What’s the cause of that gap? A myriad of different things, but research largely points to the androgenization that occurs in males. We can see this very clearly when we look at pre-pubescent boys and girls. Across a variety of athletic domains from speed to power to endurance, up until the age of 12 or 13, boys and girls improve at very similar rates. There is a small gap in performances between the two, but the overlap between boys and girls is significant. The pictures below show such a change.

Once puberty hits, boys rate of improvement rockets upward while girls improvement continues on at a steady rate. By the time both groups reach their late teens, the gap between the best in each group reaches 10-12%.

And for the non-track types reading this, 10-12% is massive at the highest level. It’s hard to conceptualize what that means, but it’s a huge gap. Even a few percentage point differential can be insurmountable.

Now it’s important to note that this does not mean the gap between every male and female is 10-12%. My brother, a non-runner, for example, would be beaten in a mile by thousands of women, even if he trained as hard as he could for the event. Performance is complex, it’s more than just what sex you are that contributes to it. BUT the androgenization that comes with being a male is the most powerful performance enhancer we have.

My Take: Should we separate men and women in sport? Yes. Women’s professional sport is amazing and there is value in it.

How do we divide sport currently and why do we utilize testosterone as a marker?

Currently, the IAAF uses what’s on a passport or legal document. So if your passport says women, then you compete in the women category. The exception occurs if you are found to be someone who is 46-XY with a DSD.

The new rule kicks into effect IF someone competing in the women’s division has Testosterone levels in the male range AND is XY. It is NOT about regulating Testosterone levels alone.

Why does the IAAF utilize testosterone to separate sexes? The IAAAF used to use chromosomes, but several conditions (more on that shortly) make simple XX vs XY complicated. For instance, someone who is XY but has complete androgen insensitivity, receives little of the androgenization effects of testosterone. They should be allowed to compete as a female with little doubt.

Recognizing that it’s more complicated than simple chromosomes, the IAAF has switched to testosterone as a way to dilineate XY individuals with a DSD. Why? 

Because of all the surrogate markers for sex, testosterone has the least overlap between men and women. Think of it like this: If 99.9% of all women were shorter than 5’6 and 99.9% of men were taller than 6’3 then we wouldn’t be using testosterone to draw that line, we’d use height. That’s the reality of testosterone between differences between the sexes; nearly zero XX women have testosterone levels that reach that of men.

Again to repeat, the reason that testosterone is used is not because it is the sole determinant of athletic performance. It’s that it plays a large role in the performance differential between males and females, but more importantly, there is little overlap between XX females and XY males. It helps distinguish between the two. Testosterone acts as a surrogate marker.

My Take? Is testosterone the best marker? It appears like it’s the best we’ve got right now, but I’m not convinced that we can’t do better utilizing a multi-marker approach.

 What is a DSD and why are people questioning Semenya’s sex?

Depending on where you stand it’s easy to make an absolute claim. “Semenya is XY that means she can’t compete.” OR “Semenya is a WOMAN. She grew up as one.”

While you might not like this answer, on both sides, the answer is much more complex then these simple statements. Semenya’s gender is female. 100%. As a DSD, it gets much more complex when we venture into biological world of sex. Please don’t stop reading if you are angry at me for that statement, please read on.

Sport has been forced to try to draw a line in the sand between male and female, where instead of a clear-cut division, it’s a world of grey. Contrary to 5th-grade biology, we can’t simply divide sex by chromosome (XY or XX).  For 99% of us it’s cut and dry. But for the .018% of individuals with a DSD, there is uncertainty. They span the gamut between the defined categories of male and female, making it impossible to draw a line without harming someone.

How complex is this? Even the medical experts have to make informed but not entirely concrete decisions based on what sex a child should be raised as, according to the latest research. One thing of note though, is that for some DSD’s, research does point towards a ‘recommended’ sex from a health and biological standpoint.

Before we go on, please read this chart. It wonderfully demonstrates the complexity of DSD’s. As you can tell we are looking at chromosomal differences, hormonal differences, and resulting physical differences. All three can vary greatly depending on the specific DSD.

In Caster Semenya’s case, she is 46 XY with a DSD. According to reports, she has internal testes, which produce high levels of testosterone (in the male range). What specific DSD she has is not public knowledge. But what we do know is that she does not have complete androgen insensitivity. In such a case, the high circulating testosterone levels don’t matter, because they have little effect.

What we should recognize is the sheer complexity of it all. Do we define sex based on chromosomes? internal or external genitalia? Hormone level and impact? It’s difficult.

Somehow, we have to make a murky mess fit into the classification model that sport utilizes. Now that classification model could have 2, 3, 4, whatever. I’m open to arguments about that. But fitting edge cases into nice and neat categories is difficult.

My take? If you want to understand how complex this is skim through this research article. Here’s the bottom line, if you think there is an easy answer in either direction, you’re wrong.

What about other genetic advantages like Michael Phelps “lactic acid” or Shaq’s height?

First off, the idea that Phelps produces “half as much lactic acid” which has been cited in The Washington Post, The Daily Show, and elsewhere is complete and utter bunk. Sorry, as an exercise scientist, that statement makes zero sense physiologically AND based on the data that statement comes from, is completely false. I won’t belabor the point, but this myth bugs me to death.

When it comes to natural advantages though, the reason this is different is simple. Height, arm length, etc. are NOT dividers of categories.

Testosterone is a divider. When it comes to men versus women, we created a women’s category to protect against this exact biological advantage (androgenization). Men perform better because of it. So we need a women’s field. So the natural advantage that an XY DSD athlete has with high testosterone is the exact advantage that the women’s division is designed to protect female athletes against. It’s the dividing marker. This is not the case for height, arm-length, lactate threshold, etc.

But doesn’t it seem dumb to regulate a natural advantage?

If a 6ft 6 boxer who weighs 200 lbs wanted to compete as a lightweight boxer, he wouldn’t be allowed. His 6’6 height is something he can’t control. It’s natural and someone that tall can’t be thin enough to compete as a lightweight.

Why is this the case? Because in boxing, weight is a dividing advantage. Weight acts as a surrogate marker for the fact that bigger people are stronger, have longer reaches, etc. than smaller people. Without weight classes, no small boxers could compete. So divisions based on weight were created.

Of course, just like with testosterone, weight doesn’t account for every advantage. People can be tall and skinny or short and heavy and weigh the same. Some of those advantages are evened out, but not all. In boxing (and wrestling and MMA) we’ve decided that the best surrogate marker for the fact bigger people typically have an advantage is weight.

We do the same thing for age in most sports. Age is a surrogate marker for development. Which again, isn’t perfect, because some people hit their growth spurt at 13, others at 16.

You can argue over whether or not testosterone is the correct surrogate marker. That’s fair. But the argument that we are regulating a natural advantage doesn’t hold water. We regulate other natural advantages when they are dividers of a classification.

My Take? The Phelps/Shaq argument is simplistic. It’s a result of taking the superficial route of thinking about a problem. It’s easy to grasp so it makes you go “Oh yea! That makes sense. That must be right!” But then once you dig deeper, you realize that the argument doesn’t hold up that well.

For instance, if we follow this argument to its conclusion, this argument gets us nowhere. If we believe that we all natural advantages are the same, then we have to either say:

1.    All advantages should be regulated (which is impossible…or we’d end up with a million categories like the Paralympics)


2.     There should be a single classification. Everyone competes in the same division (Because the only reason we have a women’s division IS to regulate a natural advantage.)

The logic of the argument falls apart quickly.

Does Testosterone Improve Performance?

There’s even been some suggestion that testosterone doesn’t play as large of a role in performance as we’d expect. That’s largely bunk.

We also can surmise the impact of testosterone on performance from looking at two contrasting stories. One, where testosterone comes from outside the body (i.e. doping), and one where it comes from inside the body.

First, when East German athletes were doped to the gills, they discovered that in females it was normal to see an improvement of 5-10 seconds in the 800m. That’s enormous at the international level. Of course, that’s adding testosterone to the body, which is different than the body producing more testosterone, but it gives us an indicator of how powerful testosterone can be.

Combine that with the fact that back when these rules were originally set in place by the IAAF, Semenya went through a period where she was forced to reduced her levels in order to compete. She regressed from a 1:55 800m in 2009 to a 2:02 800m season best in 2014.

Combine all of this with the data we have on performance improvement in boys and girls during puberty, and the picture becomes pretty clear. Testosterone matters a lot. It’s not the be all end all to performance, but when it comes to the difference between sexes, it plays a large role.

The CAS asked the IAAF to prove that testosterone improves performance in every event. That’s where the much talked about IAAF led study that showed that performance was improved in the 400m, 800m, and two field events came from. That study is flawed (explained shortly). But that doesn’t mean that tesosterone doesn’t improve performance.

The IAAF was given a nearly impossible task when the CAS asked them to prove performance increases in every event. It’s extremely difficult to show what degree any factor improves performance. It’s akin to asking the IAAF to prove that EPO improves endurance performance. Sounds simple right? Just look at Lance Armstrong. Well, because of ethical reasons in study design AND the complexities of performance, that question is nearly impossible to answer. Look no further than the myriad of studies that show EPO has no impact. An actual study from 2013 concluded “there is no scientific basis to conclude rHuEPO has performance enhancing properties in elite cyclists.” That’s the problem we are faced with.

You saw this same phenomenon in the 1960s when doctors were telling athletes steroids don’t work based on their research, and athletes were screaming that they do.

Performance is incredibly complex, even for something as strong as Testosterone. That doesn’t mean it doesn’t play a large role. It just means it’s incredibly difficult to isolate the impact of it on performance. Even more so when we have to figure out what natural testosterones impact is (which means that we can’t do any intervention study with exogenous testosterone).

My Take? Without a doubt, testosterone improves performance. Zero doubt that it makes a difference in every event. While exogenous Testosterone is not the same as a natural advantage, there is a distinct reason why athletes in all event groups dope with Testosterone. There’s also a reason that it seems to impact women’s performance to a higher degree than men.

Yes, I understand why the CAS wanted every event studied for evidence. But to me, that’s akin to WADA requiring evidence that EPO or steroids improves every athletic feat. It’s an impossible task.

The IAAF should look at improvement rates of DSD athletes from a young age until post puberty. My strong inclination based on the few known cases is that they will follow the “boys” improvement curve (i.e. they will make massive jumps in performance during the teenage years.)

If it’s such an advantage, why isn’t Semenya way faster than all other females?

A refrain I’ve heard often is that if Semenya has such an advantage why isn’t she running as fast as men or the world record holder.

A few things:

Unfortunately, the record book is littered with times from doped athletes. In particular, there are many times from athletes from Eastern European countries who had state-sponsored doping, such as the current WR holder. If you take these times out, then Semenya’s time looks much faster.

Semenya’s performance range. Being medalist potential at 400m, 800m, and 1,500m is unprecedented in the history of sport. Zero people have done this before. Add in her new attempt at the 5k (16:04 5k at altitude beating a 15:04 performer) and her range is far beyond any female in history

Finally, performance is complex (I keep saying this!). There are many men who are slower than women. So we have to remember that Semenya’s advantages don’t simply put her at world class male levels. Without her testosterone advantage, she was around a 2:02 type performer. So in order for her to be at male levels of performance (let’s say 1:43 for 800m), she needs to be about 1:50 performance level with low testosterone, for her to run around that level at high testosterone.

Often, athletes with a DSD won’t even be world class. That doesn’t mean their testosterone doesn’t help them or give them a leg up. It just means there baseline talent in other areas for running fast, wasn’t that good. Think of the slowest kid in your high school gym class. Even if you gave him a huge advantage in one or two parameters, that’s not going to make him far and away faster than everyone else.

My Take? This argument is often given by those without a track and field background who don’t understand running performance very well. That’s not meant as an attack, just that when you are in the sport and have witnessed thousands of athletes perform, you understand the nuance of performance and what a second or two here and there actually mean.

Should we be concerned with the IAAF’s ruling?

The IAAF has certainly mishandled many parts of this situation to a large degree. While very few of us were sitting behind closed doors or at the CAS meeting, from the outside a few things are apparent.

First off, sex testing has a long (and tortured) history. There were fears of men competing as women as far back as the 1930s. The first sex testing at a major championship came in 1966 when there were fears that females from the Soviet Union were actually males. Sex testing was arcane and invasive and a reflection of their times. It progressed from crudely looking at genitals to XY chromosome testing, and each step along the way proved inadequate.

According to some reports, the IAAF may be utilizing physical examinations that are a bit extreme. The IAAF needs to balance medical care and diagnosis with privacy concerns of the individual. This issue should be treated with care and scientific rigor. Standard of practice should be applied. An independent panel may go a long way to figuring out the best way forward here.

Genetic tests might provide an answer in the future. For instance screening for androgen receptor genes to rule out androgren insensitivity. Or testing for the SRD5A2 gene to check for the Alpha-5 Reductase deficiency form of DSD. Genetic testing won’t solve the problem. But if more athletes who have high testosterone levels can be screened as step one, that would eliminate needless and intrusive physical testing. Also, it seems like an MRI or similar medical imaging would be preferential to invasive inspection.

Secondly, more consideration should be given to the impact of androgen suppressive drugs. There are side effects and implications to any drug taken. So this should be taken into consideration.

It’s worth noting that for many of the DSD where a female gender is chosen, routine doses of estrogen is standard medical practice. What I don’t think has been established clearly is, does the medical dose of estrogen reduce testosterone levels to the acceptable IAAF level? And what are the ramifications for such treatment?

So again, an independent medical group to set a standard of care for athletes who choose to take testosterone suppression drugs would be a wise move.

Third, the IAAF bungled their scientific study to determine how much an effect testosterone influences performance in certain events. The study appears to be flawed. It’s covered in depth here.

Lastly, the IAAF’s press releases and Q and A have been at times absurd. Any issue is a battle of PR. Sadly, that’s how it works in today’s sound bite culture. In a complex case like this, the role of the IAAF should be to clearly tell what the issue is and what the decision entails. That’s not the case here.

My take?

The IAAF made some big mistakes. They need to take a long and deep look at their medical procedures, as the reports of some of them are almost barbaric. This might be a case of the IAAF ‘won’ the argument in spite of what they did.

It’s important to realize that in this argument you can be for division by sex and the CAS decision, but against how the IAAF is enforcing it. It’s easy in complex arguments to throw the baby out with the bath water. Instead, we need to evaluate EVERY individual component of this case.

Is this discrimination against Semenya?

Let’s get this out of the way, Semenya is facing discrimination. That being said, this is the most tricky issue to deal with. And the most sensitive.  I don’t feel I’m qualified to give an informed opinion on this topic as a whole, but I will make two comments based on research.

First, Semenya has faced untold amounts of racism, sexism, and discrimination. 100% True. She has stood strong throughout it all and it is amazing.

One question that comes up is frequently is why do they seem to targeting people of color from Africa?

While it’s impossible to rule out racial bias, what we do know is this. With Semenya’s condition, early diagnosis of the issue can occur through medical testing. In the U.S. DSD’s like the one Semenya has are often discovered early on. When discovered early (before puberty especially) more options are available for surgery, medicine, etc.  For instance, in one such DSD, surgery typically occurs between 6 to 15 months of age. In US hospitals, there is a standard diagnosis procedure for babies.

Unfortunately, in many countries this isn’t the case. Early and accurate diagnosis does not occur until after puberty starts, which can complicate both gender identity and the medical treatment. Furthermore, early diagnosis impacts which sex an individual will be raised as. For instance, the standard of care for certain DSD’s at an early diagnosis may be to be a male as medical surgery or treatment is often recommended very early on. Yet, a delayed diagnosis in another country, will lead to the opposite sex.

So depending on where you are born and when diagnosis occurs, a medical recommendation for which sex the individiual should be raised as differs.

Lastly,  there are a lot of meme’s going around pointing towards the current world-record holder in the 800m, Jarmila Kratochvilova as an example of why this must be racially motivated. After all, if Kratochvilova with her very masculine appearance is the world record holder still, and didn’t seem to face such vitriol as Semenya, it must be race related. Kratochvilova was an athlete who competed in the 1980s for the Czech Republic and she was most likely a part of a state-sponsored doping program. History tells us that she did in fact face accusations of doping, gender issues, and more. The Chicago Sun-Times ran a headline of “Is Czech Star really a She?” For decades, there have been calls of removal of her times, and Kratcohvilova has been shunned by the sport. And like Semenya, she is in many ways a victim. Only not of a biological quirk but of a state-sponsored doping system, which put her health at risk for life.

Again, that doesn’t mean that Semenya hasn’t faced racism. She has. But there are many examples of others who faced barbaric sex testing and discrimination, unfortunately.

So Where Are We Left?

Sports division of sex is binary. Gender is fluid. Biological sex isn’t fluid but it’s not a clear cut divide.

How do we fit all of that into a binary system? The answer is we can’t.

It’s easy to throw our hands up and say that we should do away with gender. But that would eliminate women’s sport.

Instead, we have to figure out where do we draw the line. There are two options as I see it:

A.    We let the athletes choose.


B.    We utilize some objective marker or markers to divide.

Choice A is easier, but it also opens up the women’s category for complication. Maybe it’s fear mongering that believes someone will abuse this category. That’s what some argue. Not that DSD and transgender are the same, but if we allow athletes to choose, then it seems hypocritical to allow everyone to choose, but then to regulate some (i.e. trans athletes). If we then eliminate the requirement for transgender athletes to lower testosterone to compete, then performance in the protected women’s category is hopelessly unfair. The above scenario is already playing out in some parts of the world, with non-Testosterone reduced individuals competing against XX women.

That’s one hypothetical scenario that some might argue I’m being unfair or discriminatory but remember that if we believe there should be a women’s division, then we are separating it for a reason (to give the large population of XX women  a chance to compete at the highest level)

If we choose B, then we are left with drawing a line in the sand where there is no clear place to draw it. Look again at the chart on sex and DSD classification. Where do you draw it? No matter where you place the line, someone will be discriminated against. Someone will be hurt.

The final option is to create more categories. Similar to what the Paralympics does. The issue here is that because of the incredibly small population of DSD athletes, athletes like Semenya would lose marketability and financial incentives because the fields and exposure would be tiny compared to the women’s division. Another solution is to simply have an OPEN category and an XX women category is an option.

So what do we do? It’s easy to complain. It’s easy to shout. But what’s the solution?

We need to come to terms with the fact that there are no easy answers here. There is no perfect solution that satisfies everyone. In dividing sex, we are trying to fit a square peg into a round hole. We have two distinct categories, where the reality is much more fluid and nebulous.

Wherever we draw the line, someone will be hurt and discriminated against. But at the same time, it’s important to have some sort of division. One that allows XX women to compete at the highest level. We’re caught in a conundrum, how do we protect the majority of women, while not causing harm to a very small but important minority of individuals who have a DSD. This is an ethical dilemma where no true answer lies. This is indeed a hard problem.

Regardless of what side you come down on, there is no answer that will provide that feeling of satisfaction. As humans, we are emotion driven creatures that crave some semblance of closure and fairness. This case will not provide either of those. That’s why it’s so important to face the issue with scientific rigor, to make sure that we stick to the facts of the matter. And above all else, we need to treat those involved with care and empathy.

To me, it comes down to realizing that it’s okay to hold seemingly opposing thoughts at once. That Semenya is an amazing athlete and a role model, but that there needs to be regulation for DSD athletes. And that none of that is fair.

Steve Magness is an author, coach, and sport scientist. His latest book is The Passion Paradox

Get My New Guide on: The Science of Creating Workouts


    1. Travis Stiles on May 9, 2019 at 10:48 pm

      I cant say I agree with everything in this, but i will say you did a damn fine job in what you put forward. Thank you sincerely for taking the time and putting in the effort.

    2. Eric on May 10, 2019 at 5:27 am

      Thanks for sharing this helpful overview. Similar to many controversial topics, people need to be better informed.

    3. Emma on May 10, 2019 at 10:29 am

      Hi Steve, thanks for this great overview.

      Something I’ve been wondering that I don’t think was mentioned in this post was the question of how does the IAAF plan to identify DSD athletes? Is it only when suspicions arise about a particular athlete based on their appearance? Should all female athletes competing in these events be tested to ensure a level playing field?

    4. Thomas Hamilton on May 10, 2019 at 5:05 pm

      Brilliant article that sheds great light on a bewilderingly complex situation, Steve. Many thanks for the viewpoints and research on this.

    5. Doug Consiglio on May 10, 2019 at 5:31 pm

      Great overview Steve!!

    6. Kate Minor on May 10, 2019 at 8:03 pm

      Sorry. Not female
      XY chromosomes
      no uterus
      no period
      no potential for menopause
      no chance of bearing children

      a vagina might make a woman, but not a female

    Leave a Reply