Perhaps the selection bias is some combination of "understands how the Internet works / not willing to accept paid results in searches / looking for real facts and data" that lead users to use both Google and Stack Exchange?
Clearly Google has an obvious and overt lead in all categories, and all search. I'm not sure how you can refute that the set of visitors to one site is not a biased sample compared to all Internet users.
Your results are different than other published search engine use comparisons. If both are factual, then the measurements must come from different populations.
"Your results are different than other published search engine use comparisons"
Link please?
The traffic for my site isn't as extreme as the stack exchange, but I'm seeing 9 times more traffic from google than bing. This leads to the same conclusion that Jeff came to. If Bing stopped sending me traffic tomorrow I wouldn't care. If google stopped sending me traffic I'd lose nearly all of my traffic.
Clearly Google has an obvious and overt lead in all categories, and all search. I'm not sure how you can refute that the set of visitors to one site is not a biased sample compared to all Internet users.
Your results are different than other published search engine use comparisons. If both are factual, then the measurements must come from different populations.