Yes, the numbers could be slightly skewed, ...
Yes, they
might be -- but you have no way of
knowing this, knowing that they're
slightly skewed, or knowing that they're so
totally skewed that you're reaching the
complete opposite of an accurate conclusion. So it's best to look at this as purely an entertaining game -- like fortune cookies!!
Maybe it's just me -- or the years spent using data analysis to make accurate inferences and draw reliable conclusions. I just can't get any fun from thinking I've arrived at meaningful numbers when in fact there's absolutely no reason to believe that.
Stykbow said:
It should still give us a fair idea of our age range on the site.
Such a survey MIGHT turn out (by
accident of available data or biased sampling) to represent the same underlying truth that a reliable analysis would -- but there is
no way of knowing that, even if it were to happen. If you don't grasp this distinction, then I guess that the surveys can be a lot of fun. But if you do grasp the distinction, you have to realize that such a survey is just like picking what's true by throwing dice or spinning a roulette wheel or throwing a dart -- which can also be fun.
And I guess that's mostly why the survey feature is provided anyway.
I know that people will continue to insist on missing the point about these "surveys" and what they "show", but (again) in fact they don't provide a "fair idea" of
anything -- or at least you can
never have any confidence that they do. And I do know that perhaps most people don't know enough about elementary statistics to understand this. Perhaps we should do a survey on
that!