From SwimSwam contributor Chris O’Linger, an assistant swim coach at the University of the Incarnate Word. (Featured Image: Michael Phelps)
For as long as there have been World Records for competitive swimming, many people have wanted to know if the females will ever catch up to the males. There have been several reasons listed for why the genders stay separated by time, and many of them are obvious, but in today’s age, we have seen several female athletes attempt to compete with the men, some even successfully.
I recently ran a study comparing the female World Record progressions to the male World Record progressions over a sixty-year period, only to find no overly significant results for the females “closing the gap”, so to speak. However, a by-product that caught my attention from my correlational and regressional analyses of male and female World Records in the same time period. Upon the interpretation of my findings, I found that there was no significant results between the mean comparisons of male and female World Records set in any of the time periods I had analyzed.
After feeling disparaged that I had not found backing for my hypothesis pertaining to females catching up to males, I began wondering if I could use these correlations and regressions to (somewhat) accurately predict the times of one gender when comparing it to the results of another. To explain my work synoptically, I will display my findings and interpretations of the current (2013) long course World Records. The data set will be shown first, the analytical tests run, and an application process to show its effectiveness.
World Records in Seconds
F WR | M WR | DIFF | |
50 FR |
23.73 |
20.91 |
2.82 |
100 FR |
52.07 |
46.91 |
5.16 |
200 FR |
112.98 |
102 |
10.98 |
400 FR |
239.15 |
220.07 |
19.08 |
800 FR |
493.86 |
452.12 |
41.74 |
1500 FR |
936.53 |
871.02 |
65.51 |
50 BA |
27.06 |
24.04 |
3.02 |
100 BA |
58.12 |
51.94 |
6.18 |
200 BA |
124.06 |
111.92 |
12.14 |
50 BR |
29.48 |
26.67 |
2.81 |
100 BR |
64.35 |
58.46 |
5.89 |
200 BR |
139.11 |
127.01 |
12.1 |
50 FL |
25.07 |
22.43 |
2.64 |
100 FL |
55.98 |
49.82 |
6.16 |
200 FL |
121.81 |
111.51 |
10.3 |
200 IM |
126.15 |
114 |
12.15 |
400 IM |
268.43 |
243.84 |
24.59 |
400 FR |
211.72 |
188.24 |
23.48 |
800 FR Relay |
462.06 |
418.44 |
43.51 |
400 M Relay |
232.05 |
207.28 |
24.77 |
As you can see, the left column contains the female World Records, the middle column contains the male World Records, and the right column contains the difference between the male and female World Records in seconds. The male records come to a mean of 173.437 seconds with a standard deviation of 216.031, and the female records come to a mean of 190.1885 seconds with a standard deviation of 200.059.
When I ran a comparison of means analysis (an independent samples t-test), I found that the difference between female and male World Records were statistically similar with regards to every event, meaning that the difference between the times for every event were equal when the length of each race is accounted for, regardless of the stroke/strokes involved. Afterwards, I ran a Pearson test of analysis, finding that the times between male and female World Records concerning all events are correlated to the extremely high degree of nearly 97% (r = 0.967).
Prediction is a terrible word to use when aerobic and human domains of performance are concerned, but the high level of correlation across the board allowed me to run a standard linear-fit regressional analysis which demonstrates a level of confidence to predict across correlated variables. What I found was a prediction rate above 99% (b-coefficient=0.991).
To be perfectly clear, this does not allow I, nor anyone else, the ability to predict exactly what any given athlete will do, but if the trend differences between male and female World Records sticks in the future as it has over the past sixty years, the time drops of one gender should very well project (with almost perfect confidence—99%) the time drops of the opposing gender.
There has been much research suggesting explanations for why female World Records have been slightly slower than males, but an evolving field is beginning to explain the trend similarities among both genders. I extend a sincere apology to all of the females in the audience as research shows that the odds of female World Records catching the male World Records is minimal, but have faith in the fact that history demonstrates a trend from inhibiting these differences from extending any further than in current times.
I look forward to continually tracking this progress as the sport’s training methods and interventions continue to evolve. We have reached a pivotal point of technological integration and philosophical change in the past fifteen to twenty years in relation to all aspects of training and performing, and the fact is, we may be entering the fastest couple of classes of all time into college swimming, and the World scene has become more competitive. These World Records should continue to be reset, and it would be interesting to see if either gender will be able to break the trend.
O’Linger is an assistant coach for the Incarnate Word swimming and diving program. He swam collegiately at both the University of Florida and University of Tampa. He earned a degree in social psychology from Tampa. He is studying kinesiology.
For a lot of you, this much research seems too easy. The fact is, this website is to lay my ideas out, and hopefully those who re uninformed on research methods will enjoy an insightful read, and those who are more rehearsed will inquire about my full work. I only reported a mean and SD just to have it noted, and plenty more was run than a simple means comparisons. Those interested in any of my works an email me individually for a conversation. I am very open to it; it’s where I get most of my ideas. The main focus here were the results from regressional analyses, not the means comparisons. It was a by-product of another study in… Read more »
Very interesting work. However, if you are doing panel data (even if you are doing cross-section), simple t-test won’t give you the whole picture. You need to control for a bunch of variables as well as ensuring that you don’t have issues with autocorrelation, heteroskedasticity which could bias your results. Now, I am not sure if you have done that or not in your study. I think is worth continuing your work and try to examine the drivers of male and female world records. A useful statistical tool would be to use a decomposition to see whether the “lagging behind in females” is due to observed or unobserved characteristics.
Congratulations on this very interesting project!
Erik
What was the reason for using the world records?
Good article. My own studies several years ago showed that the difference in National, Olympic, and World record times between men and women, whether in SCY, SCM, or LCM followed a consistent tight band of 9.5% to 12.8% over all events. The mean percentage time difference was approx 11% between men and women. This can be correlated with fluid dynamics Power/Weight and Power/Drag ratios.
First off I want to applaud the use of math/science in swimming. It is something the sport is severely lacking. However I have one suggestion: Examining the mean time and standard deviation for the events does not make sense. This gives incredible weight to the longer events like the 800/1500. You should be using a geometric mean for this as it will normalize the differences between the amount of times it actually takes to do the events.
It would also be interesting if you highlighted in red for each event which gender had more recently broken the world record. Based on the above I would guess that it should be about 50/50 but if it… Read more »
http://gizmodo.com/5977989/internet-explorer-vs-murder-rate-will-be-your-favorite-chart-today
Do you believe that this correlation is not significant?
Very interesting application of the t-test. Nicely done math, nicely written article.