Women Slow Down Less Than Men in Marathon Heat
Across 91 marathons, women's finish times slow about 20% less than men's in the same heat, even matched for pace and age.
Ask enough runners whether men or women handle a hot marathon better, and you'll get confident answers in both directions. We checked it against 91 marathons' worth of results, more than two decades of race days, and found a real, consistent gap: when it's hot, women's finish times slow about 20% less than men's.
The gap holds when each sex is measured against its own race-day weather, when men and women from the same race and year are compared, when they're matched by how fast they normally finish, when they're matched by age, and after checking whether heat was simply pushing more women than men to drop out before finishing. None of those checks made it go away.
Why we didn't just compare hot-day finish times
The obvious way to check this would be to take everyone's actual finish time on a hot day and sort men and women by how slow they ran. We didn't do that, because on a hot day the group that "normally runs 4:00" fills up with people who'd normally run 3:40 and got hurt by the heat. Sorting people that way means sorting them partly by how much the heat already hurt them, which is the exact effect we're trying to measure.
Instead, the unit of comparison is a position in each sex's field, not an individual runner. Runners aren't tracked from year to year. Take the woman finishing at the 30th percentile of the women's field and the man finishing at the 60th percentile of the men's field: each position gets a race-normal time, the average time run at that position across all of that race's years. If those two positions both normally run around 4:00 at a given race, we compare them to each other. Heat on any one day can't move a position like that, which is what makes the comparison safe.
On top of that pace matching, we checked the gap by 10-year age band, since women finishers in this data run younger than men on average (median age 36 vs. 40). We also checked whether heat was changing who shows up in the results at all: whether the female share of finishers, and each sex's total finisher count, moved as the weather got hotter. It didn't move enough to explain anything.
How much heat actually costs, in degrees
WBGT, wet-bulb globe temperature, is the heat-stress number that factors in humidity and sun along with air temperature, rather than air temperature alone. It usually reads lower than the air temperature on a race morning. We binned every race-year by how hot the runners' hours on course actually were, then compared each sex's median finish time in that band against the same race's own cool years (WBGT under 50°F).
Both sexes slow down as the bins get hotter, and men slow down more in every one. At WBGT 65°F and up, the hottest band in this data, men ran about 7.1% slower than their race's cool years. Women ran about 5.7% slower. Put in minutes: for a man and a woman who'd each run 4:00:00 on a cool day, 7.1% is about 17 minutes added and 5.7% is about 14 minutes added. Same heat, about three minutes less lost.
The pattern isn't universal race to race. Of the 154 race-years hot enough to compare (WBGT 62°F or higher in the women's window, with at least two of that race's own cool years on record), the women's gap to the men narrowed in 108 and widened in 46. The overall direction is clear across the dataset. Any single hot race could still go either way.
This isn't one race's story. The gap shows up across 91 marathons, and it's still there once men and women are matched by normal finish time and by age, not just in the raw medians.
The gap holds once you match pace and age
Comparing men's and women's finishing positions that normally run the same time at a given race, women still slow less: about 0.069 percentage points less per °F of WBGT than men (men's matched slope works out to roughly 0.35% per °F, so this is close to the 20% gap we found from the unmatched medians above). That's a little smaller than the unmatched 0.075-point gap, so pace matching trims the finding slightly rather than growing it. Matching by 10-year age band instead gives a bigger gap, about 0.087 points less per °F.
Splitting the pace-matched comparison into finish-time bands shows the gap at every pace from 3:15 to 6:30, not just in the middle of the pack. It ranges from about 0.037 points (runners who normally finish 4:45–5:15) to about 0.091 points (5:45–6:30, though only 10 races have enough runners of both sexes finishing that slowly to check at all). The age breakdown tells a similar story from 18 through 59: the gap shows up in every decade. The 60-and-over band is the exception: the estimate there is close to zero, but with only 34 races behind it and a wide range either way, that group is simply too small to say anything.
We also checked two ways heat could fake this result without a real sex difference behind it. Slower runners aren't simply more heat-sensitive in this data: men and women sit at nearly the same point on the pace-vs-heat-sensitivity curve, so pace differences between the sexes can't be doing the work (removing pace from the model directly barely moves the gap). And heat doesn't appear to be pushing more women than men out of the finishing results either: the female share of finishers barely shifts as it gets hotter, and both sexes' finisher counts fall by a similar, noisy amount.
What might explain it
Our data shows the pattern, not the cause. The most credible explanation in the running literature is pacing, not physiology: men are more likely to start faster than their fitness supports and pay for it later, and heat seems to punish that habit harder. Deaner et al. (2015) found men slowed by about 15.6% in the second half of a marathon, against 11.7% for women, and men were about 2.8 times as likely to slow by 30% or more. Trubee et al. (2014) looked directly at heat: comparing Chicago's hot 2007 race against its cool 2009 race, women paced better than men, and the pacing gap between the sexes widened in the heat. We're calling this likely, not proven. As far as we could find, no published study has compared the sexes at matched pace in a hot marathon the way we did here for finish time.
Our result agrees with two earlier studies. Vihma (2010) found warm-weather effects less evident for women in Stockholm marathon results from 1980 to 2008. Weiss et al. (2024) found the temperature effect on finish time "more pronounced among men" across two decades of Berlin Marathon results. El Helou et al. (2012), looking at 1.79 million finishers across six major marathons, found the decline with temperature "analogous" in men and women. None of the three compared men and women at the same pace.
Physiology probably plays a smaller role than pacing. Sex differences in sweat response are mostly explained by body size and surface-to-mass ratio rather than sex itself, and women sweat less than men only at very high heat loads (Notley, 2017, and Gagnon and Kenny, 2012). That's a real difference, but on its own it doesn't obviously predict a 20% gap in marathon slowdown.
What this means on race day
Don't read this as permission for women to shrug off the heat. Every band in our data, for both sexes, got slower as WBGT climbed, and at WBGT 65°F and up that's a real hit to anyone's finish time, roughly 6 to 7%. The gap here is in degree, not in whether heat matters: women lose less time than men do, not none. The practical takeaway, for both sexes, is the one pacing research points to anyway: start conservative on a hot morning and hold it, rather than banking time early that the heat will take back later. Our guide to running a marathon in the heat covers pacing, hydration, and gear for exactly these conditions, and our RunScore post covers how we price heat into a race-day forecast in the first place. If nobody's showing you WBGT for your own race, dew point is the next best number to check.
More guides
- How to Run a Marathon in the Heat: pacing, hydration, and gear for a warm race day
- Why We Rebuilt RunScore From 6.3 Million Marathon Finishes: the dataset this post pulls from
- Stress-Testing Our Marathon Heat Number: how we stress-tested the heat slope this post relies on
- Dew Point Is the Marathon Stat That Predicts Your Race Day: the number to check when nobody's showing you WBGT
- How Weather Affects Your Marathon Pace: the full pace-cost model heat is one part of
- Find your race: the live RunScore and heat forecast for your next start line