The NYC Marathon

New York’s annual marathon took place yesterday. Watching a bit of it on television with my friends, I was struck by the much earlier starting time for women than men. Specifically, professional women started running yesterday at 9:10 AM, while professional men start running at 9:40 AM. (This information comes from the runner’s handbook.) I wanted to get a sense of how much this head start depended on real differences in their performance, because I found it very hard to imagine why professional women would run significantly slower than professional men.

Of course, I have seen discussions of the speed difference between men and women before, but I was still very surprised by it yesterday. To get a sense of the scope of the differences, I found some data this morning from the ING Marathon website and made a quick density estimate plot, which you can see below:

hours_gender.png

It’s clear that men and women had quite difference average speeds yesterday, and that their times had very different distributions. Of course, these plots are each based on 100 observations, so I’m hesitant to make any strong conclusions. Having confirmed for myself that there are real differences in the performance of men and women, I have to confess that I still find it surprising.

For those interested in following up on this, the code I used to produce this plot and the data set I used are both available on GitHub. I’m sure there are other interesting questions one can ask of this data beyond simple comparisons across genders.

5 responses to “The NYC Marathon”

  1. John

    Maybe men are less willing to do something they’re not competitive at. There are certainly physical differences in men and women, but I doubt they come close to explaining the difference between the two graphs. I imagine you’ve captured a psychological difference between men and women. It would be interesting to sample men and women who *thought* about running in the NYC marathon.

  2. Paul Thompson

    This has exactly nothing to do with interest, enthusiasm, or any such nonsense. What it has to do with is stride length and height.

  3. Marc J

    … and heart size and blood cell count…

    By the way, if you intend to compare the 1000-1100 top of both genders to test if attitude plays a role, you should consider the total counts of both genders. If significantly more man are taking part, this might also be misleading in the sense that you compare persons with ndifferent attitudes (still professional men vs already funtime running females, maybe).

  4. Bill M

    thanks for plotting up that data–as an aspiring half-marathoner I’ve looked at similar data–unfortunately the NYC marathon results website doesn’t let you easily download their entire dataset–so you were only plotting the top finishers. As I recall the median time for men at marathon distances is around 4 hrs.

    If you look at the qualifying time requirements for elite runners (http://www.ingnycmarathon.org/entrantinfo/apply.htm) there certainly is a difference between men and women’s time distributions for these distance events. It is interesting to note the demographics of the 2010 race (from the ING NYC marathon results page):

    Starters
    Men 29,097
    Women 16,253
    Total 45,350

    Finishers
    Men 28,757
    Women 16,072
    Total 44,829

    Weather
    45°F, 48% humidity, wind 3 mph

    Finishers by Age Group
    Men Women Total
    18-19 109 73 182
    20-24 605 568 1,173
    25-29 2,438 2,612 5,050
    30-34 3,814 2,872 6,686
    35-39 4,975 2,775 7,750
    40-44 5,724 2,782 8,506
    45-49 4,267 2,026 6,293
    50-54 3,424 1,351 4,775
    55-59 1,638 595 2,233
    60-64 1,046 271 1,317
    65-69 323 76 399
    70-74 140 16 156
    75-79 37 4 41
    80-84 9 2 11
    85-89 2 0 2

    Oldest runners
    Joy Johnson (San Jose, CA), age 83, 8:04:59
    Sheldon Zinn (Brooklyn, NY), age 87, 8:08:41