Sure, not everyone votes in polls or participates or prefers to speak English. Same as me standing on a random street, trying to poll people for their food preference, and not expecting everyone to participate or speak English.
However, 110 people is a relatively decent sample size (in the research lab where I work, that's a good amount that will already get you statistical significance, depending on the work done of course). And from that sample size, I would expect those that have voted in polls, participate and speak English, to already be a representative average of the whole, regardless.
If I'm on that street and after 110 participants, they're already slanting towards liking Fish Curry, Chutney and Chana Massala, I could conclude that people who like participating in my experiment and speak English, all really like Indian Food. That there's something innate about my study and Indian food, or that indian foodies like participating in studies. However, this doesn't make any sense. It also doesn't make sense that of 110 people who vote, participate and speak English, 80% just so happen to play on PC.
The more reasonable conclusion would be that I'm in India, or some major City's "India Town" for my street experiment, or that something else is going on, other then PC players just being more engaged in voting, participating or preferring English.
Though, me being a researcher also comes with having respect for other experts, and I'd say you're more knowledgeable about forum usage then me.
I don't quite take your explanation just now, but something else is going on.(That or I'm applying statistical models simply to the wrong project here. I run towards my lab's statisticians after every other project to make sure too.
)
