The Olympics have come and gone,
left nice memories and a few controversies. I think the controversy that most
needs addressing is the case of “hyperandrogenic” athletes, basically female
athletes with very high testosterone. To me the regulations that the International Association of Athletic Federations (IAAF)
implemented (subsequently overturned by the Court for Arbitration for Sport (CAS)) are possibly an example of choosing
the wrong metrics and this can have drastic consequences. Trying to exclude
people without the right metrics also allows room for disgusting comments like “I
am glad to be the first European and the second White” (https://www.rt.com/sport/356879-olympics-jozwik-second-white/).
Furthermore, it could be that the wrong question is being asked.
First, let’s have some quick
background. Women’s athletics is usually separated from men’s competition
simply for questions of “fairness”, to make it a real competition. As an
illustration, Paula Radcliffe’s Marathon record is one of the most outstanding
records in women’s athletics (2:15:25) (although run in a mixed gender race),
so far ahead of the rest (next are 2:17:18 and 2:17:43):
The men’s record is 2:02:57, the
400th men’s time is 2:07:37. Ms Radcliffe’s time would place her around 5000th
fastest marathon. So it makes some sense to have separate competitions for men
and women. But why am I writing this post? It gets worse than disgusting comments.
At least 4 women have had their gonads removed and their clitorises
partially removed in order to be allowed to participate in IAAF (athletics
ruling body) sanctioned events.
“Doctors in Nice reported in 2013
that they operated on four hyperandrogenic women, ages 18 to 21 and from rural
or mountainous regions of developing countries, cutting out gonads and partially removing their clitorises. The
athletes were told that surgery would most likely dent their athletic
performances but allow them to continue competing, the doctors reported. They
said the women were allowed to resume competition one year later.” http://www.espn.com/olympics/story/_/id/17306565/female-olympian-details-process-testosterone-level-measures-high-hyperandrogenism.
This stems from the idea that a
line needs to be drawn between the 2, men and women. According to IAAF, the
line is at 10nmol/L of testosterone;
anything above that is considered to be in the male zone. Removing the gonads
and partially removing the clitoris are designed to reduce the naturally occurring testosterone level
to below the threshold.
As shown above the IAAF threshold
is just at the lower range for average adult men. It would seem reasonable to
set the threshold there.
However there are 2 issues that
immediately spring to mind.
First, if we assume that the
distribution of testosterone among males and among females are normal, there
still are long tails; basically the neat minimums and maximums are not as distinct/absolute
as you might think. While most men will fall into the range of around 10 to 37
nmol/L, there will be a few below 10 and a few above 37. The same concept
applies to women: some below .5 and some above 2.5nmol/L. Using the minima and
maxima for the average adult as the +-3 standard deviations range for a normal
distribution:
Hence, instead of nicely
separated groups, we are likely to have overlapping distributions; these would
be out of the ordinary, exceptional people.
This brings me to my second
point; what are top athletes but exceptional people?
If testosterone is indeed a major
factor in explaining consistently exceptional athletic performance, this means
that top athletes should have more testosterone than the average adult, and
thus be concentrated towards the upper ends of the distributions for each
gender. The distribution of testosterone among the athletes would not be random
sampling from their respective populations, but they would be heavily biased
towards members with high testosterone. In other words, the measurements for
the average woman and the average man should not be used as a gauge; it would
be better to use past Olympians data to set a benchmark. After all, we should
be comparing with people with similar characteristics.
Would that solve the issue of
metrics?
Yes and no.
I think that the metric might be
acceptable (again assuming that testosterone is the main difference maker), but
the question might be the wrong one. One of the issues that was raised in front
of the CAS is that of discrimination against women. Basically the single threshold
only affects women, men with lower or especially higher testosterone are
allowed to compete among men, but women with higher levels are blocked from
participating among women...
I think that the question that
should be asked it not where is the line to be drawn between men and women, but
there should there be more than 2 categories so that competitors are more
equally matched (just like in boxing for example)?
My conclusion in this example is
that finding the right metric is critical, but also that if it is hard to find
the right metric, sometimes the question itself should be revisited.
Post Scriptum: More readings on the issues surrounding
hyperandrogenism:
Informative articles with historical persepctive:
The surgeries:
Legal Opinion:
Details about the case brought before the CAS:
Paper on gender and fair play, because there’s more to this
question than data:
P.P.S Opinion pieces:
Ignorance is responsible for the mess:
Disgusting remarks: “I’m glad I am the first European, the
second White”
A final chart from me: the best times since 2009 for the finalists of
the women’s 800m at the Rio Olympics in order of finish from left to right,
with runners names and ages (Data from IAAF):
I just inserted a quadratic trend; you can interpret for
yourself whose performances 'show the most one-way drastic improvement over that span of time'.
(Age explains why some athletes do not have that long a history of best times).
