Cross-Sectional Age Comparisons Understate Individual Performance Decline in Endurance Athletes: A Within-Runner Analysis of 264,213 Boston Marathon Finishes

For decades, scientists and coaches have tried to map the exact moment when the human body begins to lose its endurance. They want to know how fast a runner slows down as they get older, a question that matters to athletes planning their competitive futures and to researchers studying how we age. The standard way to answer this has been to look at a single race day, line up all the finishers, and compare the average time of a forty-year-old against the average time of a sixty-year-old. This method assumes that the difference between those two groups is simply the result of getting older. However, this approach overlooks a subtle but powerful filter: the people who keep running. If slower runners tend to quit the sport earlier than faster ones, the group of sixty-year-olds in any given race is not a random sample of everyone who was forty years old two decades prior. Instead, it is a group that has survived a long process of elimination, leaving behind only the most durable and consistent performers. This creates a distorted picture, making the aging process look gentler than it actually is for the individual.

A new study published in 2026 by researchers at the University of Michigan and the University of North Carolina challenges the traditional view by looking at the same runners over time rather than comparing different people at a single moment. The team analyzed fourteen years of results from the Boston Marathon, a race run on the same course every spring. They gathered finish records from 2001 to 2014, a period that included over 280,000 race finishes. To track individuals, the researchers linked runners across different years using their names, sex, country, and the year they were born. This allowed them to build a detailed history for 15,465 runners who had competed in at least three different marathons, creating a total of 65,490 individual race results to study. By following these specific people as they aged, the researchers could measure how much a single runner slowed down year by year, independent of who else was in the race that day.

The results revealed a stark difference between the two ways of looking at the data. When the researchers compared different runners of different ages in the same race, as is commonly done, they found that at age 45, a runner's time increased by about 1.28 minutes per year. However, when they tracked the same individuals over time, the decline was much sharper. For the same age of 45, the individual runners were actually slowing down by 1.62 minutes per year. This gap widened significantly as the athletes got older. By age 55, the cross-sectional method suggested a slowdown of 2.13 minutes per year, while the individual tracking showed a decline of 3.05 minutes. The difference became most dramatic at age 65. The traditional method estimated a slowdown of 2.98 minutes per year, but the data following the same runners showed they were actually slowing down by 4.48 minutes per year. In other words, the standard way of measuring aging understated the rate of decline by roughly 50 percent for runners in their mid-sixties.

The study also identified the mechanism behind this discrepancy: selective dropout. The researchers examined who returned to run the Boston Marathon in subsequent years and found that slower runners were much less likely to come back. In any given race, the fastest 25 percent of finishers had an 84 percent chance of returning for another race, while the slowest 25 percent had only a 77 percent chance. Furthermore, the likelihood of returning dropped with age; for every year a runner got older, their chance of coming back decreased slightly. This pattern means that as the years go by, the older age groups in the race become increasingly filled with only the most resilient athletes who have managed to avoid injury or loss of motivation. When a researcher compares today's sixty-year-olds to today's forty-year-olds, they are comparing a group of survivors to a group that has not yet been filtered by time. This survivorship bias flattens the curve, making it appear that aging is less severe than it truly is for the average person.

The findings held true even when the researchers adjusted for specific conditions, such as the extreme heat of the 2012 race, which slowed down the entire field that year. They also found that the steeper decline was consistent across both men and women, though men tended to slow down slightly faster than women at every age. The study suggests that the age-graded standards currently used to compare athletes across different decades are likely too optimistic. These standards are built on the performance of the remaining competitors in an age group, not on the trajectory of an individual athlete. Consequently, an athlete aging into their sixties may find that their personal decline outpaces the standards they are trying to meet, not because the standards are too strict, but because the standards themselves are based on a group that has been artificially preserved by the sport's dropout rate.

Ultimately, this research demonstrates that to understand how an individual's performance changes with age, one must follow that individual over time. Comparing different people at a single point in time provides a snapshot of who is left in the race, but it fails to capture the true speed at which the human body loses its endurance. The data from the Boston Marathon shows that while the sport may look like a gentle slope of decline when viewed from the stands, the reality for the runner on the course is a much steeper descent. For those interested in the biological reality of aging, or for athletes planning their own futures, the lesson is clear: the path of the individual is often far more demanding than the average of the crowd suggests.