What data tells us about Olympic Athletes.

A look at trends in age and gender of top athletes over the past 40 years.

Introduction

The Olympic games showcase peak performing athletes from around the world effectively defining the limits of human performance. Data from the games can therefore be explored to reveal trends in the top performing athletes over time.

I explored an Olympic Games data set compiled by the Kaggle user rgriffin (link) which includes data on the age, gender, height, weight, country of origin, and medals won. I limited the data set to the summer games since 1980 and looked for interesting observations regarding athletes over the past 40 years.

I. How does the number of athletes representing a country influence the number of medals won by that country?

As seen in the correlogram, one of the strongest, positive correlations observed was between the total number of athletes a country sends to the games and the number of medals that country wins.

This result is not at all surprising, since more athletes allow a country to compete in more events and have more athletes attempting to qualify per event, both of which increasing the odds of taking home medals. As seen in the diagram, total Athletes per country had a convincing 0.82 correlation coefficient with the weighted medals total of the country.

A scatterplot further visualizes this correlation, with a best fit line revealing that more athletes leads to more medals with a convincing R squared value of 0.68. Setting the color scale to the year of the Olympics indicates that the relationship has been fairly stable over time.

II. Are there trends in the average age of athletes over time?

Another interesting observation from the correlogram was a positive correlation with the year of the games and the average age of the athletes of both genders, 0.13 for male athletes and 0.18 for female athletes. This suggests two things, that the average age of the athletes has slowly increased over time and that female age is increasing faster than males. The line plot below visualizes these trends. Indeed, the average age of athletes is increasing over time with larger increases seen in female athletes.

III. What are the trends in the athlete gender?

A line chart of the percentage of athletes who are female in the games since the 1980s clearly shows an increasing trend. While women athletes are making considerable gains in equal representation, the age difference between genders shown above is quite striking and relatively stable over the years. This begs the question, why is the average female athlete 2.12 years younger than the average male. One possible answer is that the sports which attract more female athletes tend to favor younger ages. There are indeed differences in age per sport as shown in the bar plot below.

While women are making considerable gains in equal representation at the games, the age difference between genders shown in section II above is quite striking and relatively stable over the years. This begs the question, why is the average female athlete 2.12 years younger than the average male. One possible answer is that the sports which attract more female athletes tend to favor younger ages. There are indeed differences in age per sport as shown in the bar plot below.

There are also clear differences in the distribution of each gender by sport and summing the product of the gender distribution and the average age by sport does add up approximately to the observed age differences (23.8 vs.23.3 for females and 25.0 vs. 25.3 for males).

However, we need to consider the possibility that the difference in average sport age is a reflection of the gender distributions within each sport. A look at the average age by gender by sport indicates that with only a few exceptions (cycling, boxing, and football) female athletes are on average younger than their male counterparts.

Conclusion

The number of athletes per country has a clear correlation with the medal count for that country, which is not a surprising result but does point out the advantage that the larger, wealthier countries enjoy in the games.

The percentage of athletes who are female has risen dramatically since the 80s, increasing from 26.1% to 45.6%. The average age of female athletes has also increased over the same time period from 20.6 to 24.2 years of age while male athletes only increased slightly (24.4 to 25.9).

Male and female athletes are distributed differently across different sports and the sports do have different average ages of the athletes; however, these distributions do not account for the difference in age between the genders. Female peak performing athletes are, in general, younger than their male counterparts, though the gap is decreasing over time.

This leads one to wonder why is there a difference in age by gender and why is the gap closing over time. While those answers cannot be found in the data set used in this project, a review of studies on the topic of gender and the Olympics point to the passage of Title IX (a federal civil rights act that prohibits gender-based discrimination in education) in 1972 as the catalyst for a shift in the peak performance age for females (Elmenshawy et al., 2015). There has been a substantial increase in participation of women in competitive sports since the passing of Title IX and it appears that several other countries enacted similar legislation shortly after the U.S. accounting for a similar effect internationally (Lopiano, 2000).

To see more about this analysis, see the link to my Github available here.

• Elmenshawy, A. R., Machin, D. R., & Tanaka, H. (2015). A rise in peak performance age in female athletes. Age (Dordrecht, Netherlands), 37(3), 9795. doi.org/10.1007/s11357–015–9795–8
• Lopiano DA. Modern history of women in sports: twenty-five years of Title IX. Clin Sports Med. 2000;19:163–173. doi: 10.1016/S0278–5919(05)70196–4.