Subway, maybe you didn't understand my thought.
Elite is something subjective, depending on the level of the specific research.
If I want to see what happens to "average joggers" in the total field of people running marathon at any level, (everybody able finishing a marathon inside 6 hours, for example), and the number of runners under 6 hours is, maybe, one million from the beginning of athletics, I can suppose that Elite can be considered who was able to run at the level between 90% and 100% of the runners, that means a position between WR and the runner number 100,000 all-time. IN THIS CASE, THE PERFORMANCES OF THE BEST ATHLETES INCREASE THE DEVIATION FROM THE AVERAGE, contributing to create a wrong picture compared with the goal of the research.
If I want to see what happens to "average runners" in the field of athletes running marathon at good level, including not professional runners, able to run under 2:20 (about 80,000 all-time), we can consider "Elite" the athletes in top 8,000 in the World (runners under 2:14).
But, if we want to investigates the hypothesis that FOR RUNNING A WR ATHLETES MUST BE DOPED, BECAUSE IS NOT POSSIBLE RUNNING AT THAT LEVEL WITHOUT EPO, we need to consider a small number of athletes, because more far we go from the WR, less significative are the statistical numbers that we can produce.
In the case of a marathon runner able to run 2:10, his position in the all-time list at the moment is number 875, and frankly the number 875 can't have any connection (statistical, but also technical, practical and physiological) with who is able to run at the top.
If we want to consider "Elite" the top 100 all-time (the current limit is 2:06:22), who is expert of statistics extrapolates the best 10 (at the moment Haile Gebrselassie with 2:03:59) in order to have a proper group of subjects to analyse.
You must think that 2:10 are 7,800 seconds, while 2:02:57 (WR) are 7,377. The difference (423 seconds) represents