"...How fast are the subjects? The 10 Kenyans, including Martin Lel (many-time New York and London marathon champ) and Sammy Korir (the guy who missed the marathon world record by one second in Berlin one year), have an average time of 2:07:17. The nine Europeans, including Olympic champ Stefano Baldini and World Championship medalist Viktor Rothlin, have an average time (for those who've run a marathon) of 2:08:24. The two groups were statistically indistinguishable in terms of running performance and physical characteristics like body weight...."
Emphasis mine. This is a fine example of the difference between "statistical significance", and significance in the real world. They don't give out gold medals on statistically significant finishes, but on who crosses the finish line first.
And yes, the answer to why the Kenyans are faster is in the article, but not in the statistics.