Gender Differences in Ultra Running?

I know I may be beating a dead horse at this point, but the controversy over Run Rabbit Run’s allowance of ‘Tortoise’ women an extra hour to finish the grueling 100 mile race (unless they want to “run with their man”) spurned my own investigation into men’s and women’s finish times in 100 mile (or similar) races. Specifically, I was curious as to whether the average finish times for men and women actually differed, and if that difference was substantial enough to support such a policy. I think this question warrants investigation because it reinforces the stereotype that women in particular are not physically capable of finishing 100 mile races in the allotted times. This may cause barriers to entry for some women: by reinforcing a narrative that the 100 mile distance is simply “too difficult” for their bodies.

Among the ‘elite’ crowd (i.e., top 5 or so runners), men are faster, although that gap is closing. However, what about the average runner? After all, that’s who we’re talking about with these cutoff times. I decided to look into the average finish times for men and women at 11 prominent trail races: GA Death Race 2018 (only non-100 miler), Orcas 100 2018, Javelina Jundred 2018, Run Rabbit Run 100 (tortoises), Cascade Crest 100 (2017), Cascade Crest 100 (2018), Bear 100 2018, Wasatch 100 2018, HURT 100 2018, Hardrock 2018, and Western States 2018. I used Selenium with Python to scrape ultrasignup results tables and converted the tables into a Pandas dataframe. These are the basic functions I used:

*index_url is the page of results from ultrasignup you are looking to scrape
find the indices for the start of the DNF and DNS lists (if available)
Make a dataframe of just the finisher results. DNF and DNS lists may be analyzed separately.

After scraping 11 races, I concatenated the individual race results into a larger dataframe. In aggregate, the average finish time for males and females for these races did not differ (t(1741) = 0.75, p = 0.46).

There could be some important variance that this test did not parse out. Specifically, maybe the conditions of the race itself influence gender differences (i.e., technicality, weather, elevation, etc…). Thus, I decided to examine individual races of varying course conditions. Since I’m examining multiple races, this would increase my chances of finding a ‘significant’ result. In fact, using a significance criterion of alpha = 0.05, I would expect 1 out of 20 examinations to be different ‘just by chance’. To counteract inflation of this “type I error”, I used a Bonferonni correction for multiple comparisons.

Among all the races examined, one race emerged with a gender difference: the Javelina Jundred 2018. Maybe this is due to the race being more “runnable” — conferring an advantage for the men due to increased dependency on running in the more “aerobic” range? Regardless, I was surprised I didn’t see more, particularly among “tough” races like Hardrock or the Bear 100. Furthermore, I was disappointed that I had internalized a belief that women are slower at many of these 100 mile races, when evidence suggests otherwise. The results support the notion that most women have the grit, stamina, and physicality to push through 100 milers on par with most men (I use ‘most’ to distinguish from the subset of elite men and women).

It is also notable that, when the information was available, an analysis of the proportion of men and women in the DNF category did not differ from the proportion of men and women entrants. However, the reasons for those dropouts are unknown. Additional data on whether more women ‘time out’ (as opposed to dropping out for other reasons) compared to men in 100 milers could further support the claim that timing is not the issue when it comes to attracting women to the distance.

One could argue that women attempting 100 milers feel less confident than men, and thus only enter the race when they are absolutely sure they are prepared. Thus, women may be performing at a higher percentage of their full potential. But why would that matter for gender-specific time limits? Do we want to support more women ‘winging it’ in a race by supplying more generous cutoff times? Clearly women are capable of meeting the standards. Personally, I think this argument doesn’t go toward supporting anything other than reinforcing a belief that men are inherently stronger — the toxic thinking that prevents some women from trying the distance in the first place.

The conclusion I draw from this examination is to look toward how we can support more women and women-identified athletes in the sport. We should be engaging in discourse regarding what social barriers exist. This post points to a start for that discourse, and hopefully the results I’ve presented provide fuel to keep the fire of increasing women’s participation in ultra-running roaring.