|
I always think of surveys and polls like this: They say, e.g., "80% of respondents think such-and-such" but what I read into that is "80% of respondents stupid enough to get trapped into their poll think such-and-such". Poll results inherently show you what stupid people are thinking. Not a statistic I'd want to base policy or any important decision on.
I'm not talking about friendly polls on forums such as ours. I'm talking about the political polls that say "We called 1000 random citizens..." What they don't mention is that 997 of those citizens hung up on them, leaving the poll with results based on the three stupid ones who stayed on the line.
|