If you're going to sacrifice your body to run 26.2 miles all at once, shouldn't you at least run them as best you can? Eric Allen (USC) and Patricia Dechow (Berkeley) took a look at finishing times for 300,000 Chicago Marathon runners from 2003 to 2011, concluding, in the style of all behavioral economics papers, that marathons are just another thing that human beings don't do optimally.
If everyone tried to run his or her best race, the researchers argue, you'd expect to see a relatively smooth distribution of finish times across the large population of marathon runners. As you can see in the chart above—which shows finish time by minute—this is not the case. Instead of running their best race, marathoners seem to care more about hitting key round number targets.
The real fun comes in the gender breakdown though. Below are charts showing finish times (now in five-minute increments) for men and women, statistically significant "bunching" times are shown in black.
The women aren't immune to bunching (there's clearly an impetus to finish in under five hours), but they can't touch the spikes that male finishers have around 3:30, 4:00, and 4:30. You can almost picture the thousands and thousands of sweaty, semi-fit dudes fighting their way across the finish line at 3:59, and then refusing to talk about anything else for the next year.