Adumbrodeus:
I already controlled for popularity by using only the tournaments with the most entrants (150+). The most popular character in Melee was Link; we did not see a majority of Link mains in the top 10. As time goes on, yes, better characters can get better numbers, but this is irrelevant when you are considering the top 8 of a national tournament. The best characters played by the best players get there. It doesn't matter if they are the only ones playing the character or not; there could be 8 MKs or 8,000 MKs and you'd have the same top players there. Popularity isn't a factor because using only the top 8 of a large tournament naturally uses a weighted scale; only the GOOD characters are up there.
If you're saying that MK is just popular BECAUSE he's so good and that's why we see so many, you're claiming there is another problem: that no one else is good.
In addition to this, you need an alternate explanation as to why the results happened the way they did. You can't dismiss a statistical trend spanning about two years, two different entry brackets, and only the highest level of play at the time of each tournament by saying "that character was popular". If the character was merely popular, you'd have to explain why the character was able to do what other characters could not.