In any good poll, no matter how seemingly intelligent the sample, there will always be at least 21% (or around 1/5th) of the people that don’t have the slightest clue what they’re talking about. This theory is based on a Pew survey from 2008 which clearly showed that 21% of atheists believe in God. That fact was both startling and enlightening, and I have henceforth decided that all opinion polls are inherently terrible at determining what people actually think. It doesn’t matter whether these people don’t understand the question, or are simply stupid; there will always be at least 21% whose answers are indicative of nothing.
This has been most clearly demonstrated recently by another Pew poll showing that 18% (often “quoted” as something around 1 in 5 or 1/5th) of Americans think Obama is Muslim. This is demonstrably untrue. Sadly, only 34% (about 1/3rd) believe he is what he claims to be (Christian). But there is a lot more to this poll than those surface statistics, so let’s dig a little deeper.
Of the people who think he is Muslim, 26% approve of his performance as President. Despite having been bombarded with this Muslim misinformation/propaganda, these people still think he’s doing a good job. Thus, the 21% theory stands strong even within this subgroup. This means that no matter how much you subdivide a group, you can never truly assume that the answers you get are accurately indicative of anything. 21% of Atheists believe in God, and 26% of people who think Obama is Muslim, also think he’s doing a great job. One could take heart in the idea that these people don’t seem to think being Muslim is inherently a bad thing, if using polls to come to conclusions about what people think was a valid endeavor. We can’t come to conclusions, however, because opinion polls, not matter how detailed, are never an accurate representation of what people think.
Polls are only an accurate representation of what people are likely to answer on a poll. An election is a type of poll, so it’s easy to argue for the validity of voter polling during an election, but opinion polls should never be treated as accurately indicative of anything. If you feel the driving need to make broad generalizations about groups of people, then I suggest you rethink your position, because chances are you’re falling prey to prejudice.