Over the course of the past several months, the Sparks research team has examined the complex relationship between international recruiting and college rowing performance. In our October 26 article, titled The Significance of International Recruiting in Rowing, we found that there was a statistically significant relationship between international athletes and team performance. This effect was only present in IRA championship racing. For NCAA women’s rowing, international athletes provided no boost to performance over the time frame of our study. Although we cannot be certain, it is possible that this is related to the recent international dominance of the US women. Results from both elite international racing and championship collegiate racing may indicate that the United States produces relatively better female rowers than male rowers.
This finding motivated us to dive deeper into the subject. In our second article concerning international recruiting, released on February 26 and titled Differences in International Athlete Retention Between Countries, we defined the recruiting cost as the time, effort, and financial expenditure required to recruit an athlete. Due to the relatively higher recruiting cost of international junior athletes relative to domestic junior athletes, the rational coach will recruit international athletes only if these athletes provide high enough returns to team performance to justify the cost. These returns can be expressed both in terms of retention and performance. Retention is an easier metric to parse out. If a given international athlete carries a higher recruiting cost than a given domestic athlete, then the recruiting decision makes more sense if the international athlete is more likely to stick around. Through analysis of both IRA and NCAA roster data, we found that international athletes are indeed more likely to be retained than domestic athletes. For both men and women, German athletes were retained at the highest rates, followed by athletes from New Zealand.
When recruiting international athletes, coaches are concerned with more than just retention. Coaches want to win. By splitting international athletes into their home countries and running a multiple linear regression analysis between team points and the presence of athletes from certain countries, the Sparks research team sought to rank countries based on how well their athletes perform at IRAs. The NCAA was not included in this analysis, as our first article concluded that international athletes provide no significant advantage in NCAA women’s rowing.
As in prior studies, data from teams that sent an eight to the IRA National Championship in 2015, 2016, and 2017 was incorporated into the model. This was necessary because our dependent variable is team points; if a team does not have any points in 2016 due to not attending the regatta, for instance, they cannot be fairly compared to other teams.
In a similar manner to our first international recruiting study, the relationship between team performance and country of origin was modeled through a multiple linear regression analysis, with the dependent variable being team points, and the independent variables initially consisting of the number of athletes from each country on the team. As a control, prior year performance was also included as a variable. By incorporating prior performance, we hoped to control for external factors that could correlate with strong current team performance.
However, we quickly encountered a problem. By including variable terms for each individual country, we found that the level of detail was too granular to provide any kind of meaningful result. For instance, some countries had only two or three athletes recruited in total, while others had several dozen. We then decided to consolidate based on geographic region. By doing this, we were able to consolidate countries into statistically meaningful subregions. As a guide, we used the United Nations subregion classifications. Some subregions, such as Central Asia and West Africa, had no recruits. As such, we only included the subregions from which we had recruiting data: North America, South America, Northern Europe, Southern Europe, Western Europe, Eastern Europe, East Asia, Southeast Asia, South Asia, Western Asia, Oceania, and Southern Africa.
After we had reorganized our data (for instance, athletes from Croatia, Greece, Italy, Serbia, and Slovenia were consolidated into Southern Europe), we ran the regression.
To our surprise, the majority of geographic subregions showed no significant statistical effect on team performance; these subregions were indistinguishable insofar as driving better team outcomes. In other words, we cannot prove that recruiting an athlete from Northern Europe will lead to a better team result than recruiting an athlete from Oceania.
However, there was one exception. Athletes from Western Europe showed a statistically significant positive effect on team performance. Teams that recruit athletes from Western Europe, defined here as Austria, France, Germany, the Netherlands, and Switzerland, tend to perform than teams that do not. Cross-referencing these countries against the retention numbers, we see that Germany is the top ranked country for retention, with the Netherlands coming in at number five.
It is important to clarify that these results are derived from a limited data set, and so the results may change when more data is collected and studied. However, the combination of our respective studies of retention and performance yield interesting implications and conclusion. Based on our results, American coaches would do best to recruit junior athletes from Western Europe. Within Western Europe, athletes from Germany appear to provide the best returns to both retention and team performance.
