In a previous post, we noted a sharp decline in the number of swims by athletes over age 12 (link). In this post, we’ll learn that the age-related decline in USA-S swims is due to the multiplicative combination of three factors: fewer athletes, attending fewer meets, and swimming fewer events. We’ll also discover that the age-related decline in athlete participation is more precipitous for women than men.
To provide accurate analysis of this decline, we’ll need to address the source LSC bias in our data. Recall that our data was sourced from 18 LSCs and 1 Zone. When an athlete from a different LSC appears in our data, we only see a small fraction of their swims. For example, when an athlete from Connecticut Swimming (CT) competes in an Eastern Zone age group championship, they will appear in our data set for that one meet per season but we don’t see any of their other swims that season. This censoring increases the number of athletes in our data while artificially decreasing participation measures, such as meets/swims per athlete. The effect is enhanced for athletes at the top of their age groups (aged 10, 12, 14, 18), who are more likely to attend championship meets. Due to a peculiarity of Eastern Zone qualifying times, the effect is particularly exaggerated for 15/Over men.
To correct for this censoring effect, we will -- for the purposes of this post only -- limit swims to those by athletes from one of our 18 data source LSCs. Many swims are missing an LSC code, or have an invalid LSC code, so this filter also has the unfortunate effect of excluding valid data source LSC swims. After this filtering, we’re left with 22.2 million swims:
All analysis in this post will use the reduced “Source LSC” data. This data is still subject to censoring, whose net effect is to reduce participation measures. For example, our data doesn’t include swims by athletes from source LSCs who compete in meets sanctioned by non-source LSCs and Zones. This censoring will affect athletes who compete across LSC boundaries, as well as those who compete in meets sanctioned by the Central, Southern, or Western Zones.
Regardless of which data set we use, the number of age group swims peaks at age 12 and drops rapidly thereafter. Here we show the Source LSC data only.
SCY swims decline 72% from their peak at age 12 to age 17, while LCM swims decline 57% in the same period.
Female swims decline 75% from their peak at age 12 to age 17, while male swims decline 57% in that same period.
The decline to age 18 is more severe. However, we won’t consider the age 18 swims due a censoring effect in our data. High school seniors may swim their last short course season as a 17 year old, even if they compete in USA-S past their high school graduation. Including 18 year old short course swims in our analysis would therefore undercount athlete participation.
The number of Source LSC age group athletes is relatively constant from ages 10 to 12 and thereafter declines. The total decline from the peak at age 11 to the trough at age 17 is 64% for women, 45% for men, and 57% overall.
By age 17, USA-S age group swimming has more men than women.
Not only are there fewer athletes after age 12, but they compete less frequently, from a peak of 28 annual swims per athlete at age 12 declining 26% to 21 annual swims per athlete at age 17. For 13/Unders, women compete more frequently than men, while for 14/Overs men complete more frequently than women.
Age group athletes attend fewer meets after age 12. At their peak, age group athletes compete roughly 8.2 days in 5.0 meets per year. By age 17, their annual participation has dropped to roughly 7.3 days in 4.1 meets. For 13/Unders, women compete more often than men, while for 14/Overs men compete more often than women.
Not only do older athletes attend fewer meets, but they compete in fewer events using fewer strokes. At their peak, both male and female athletes compete in 9 different events per year using 4 of the 5 possible strokes. By age 17, they compete in 7 different events per year using only 3 of the 5 possible strokes. Below age 14, women compete in more strokes/events than men, while above age 14, men compete in more strokes/events than women.
This age-related specialization is in part due to age-related changes in competition format. Older athletes are increasingly likely to compete in “preliminaries with finals” format with qualifying times, which encourages specialization and reduces an athlete’s average number of swims per meet day.
In this post, we’ve seen that the precipitous age-related decline in USA-S athlete participation is due to the multiplicative combination of three factors: fewer athletes, attending fewer meets, and swimming fewer events. We’ve also seen that female athletes leave the sport at 1.4 times the rate of male athletes after age 11. The likelihood that an 11 year old athlete will quit USA-S before age 17 is 64% for women and 45% for men.
These facts raise additional questions about the sport. Why are so many athletes leaving the sport? Why do female athletes leave the sport at a higher rate than male athletes? And what can USA Swimming do to retain these athletes?