Lies, Damned Lies, and Statistics:

If you see a survey** with a result split somewhere in the 66% vs. 33% range, know this:

The 66% side is "normal people" almost all the time, unless the question was worded to get the result the poll sponsor wanted.* These are the people you know personally. They have half a clue, and can generally tell the difference between their asshole and a hole in the ground.

The 33% people are the lunatic fringe. You would be surprised to know somebody who answered on this side of the survey. People who are approximately as dumb as a box of rocks fall on this side of the question.

VFD, how can you say this with such confidence?

Because I know people, and I know surveys. Normal people hate taking surveys. They are a hassle, an intrusion into your schedule, and a really good one will have annoying, repetitive-sounding questions. You don't answer surveys. Nobody you know answers surveys. Mail surveys will go straight to the trash. Phone surveys get hung-up on. The people who answer surveys
1) have land lines,
2) have nothing better to do with their time,
3) probably have no job, because survey callers call more during the day, or
4) are super-helpful, willing to participate even though it's a hassle.
That narrows your respondent pool to people are either elderly, homemakers, or Democrat Lifetime Voters (read: welfare recipients). Your average homemaker is too busy for a survey. Your average old person will tell the survey taker to get off their lawn phone. So you're left with people mostly sucking off the public teat. Even most of THEM don't want to take the survey. So your pool is reduced to the very most helpful (willing to take a survey regardless) and the most feckless, willing to put up with the same amount of hassle as they are used to going through at the welfare office.

After postulating all that, you need to know that most of these people are still mostly-normal folks. They want the government to leave them alone (except for "_______" that they take from the government). They want to do what they want, within reason, and they want you to be able to do what you want, too (except keeping your own money). They want a reasonable amount of risk and security for themselves, businessmen, and their country. If you get a minority of THESE type of people on one side of an issue, rest assured that it is a very small fringe of the population feeling that way. For example: when President Obama's approval rating is tanking amongst this lot, his approval rating in the population at large is very low indeed.

This is also the reason you should completely disregard polling results when making important decisions. The sample is skewed right off the bat, regardless of how good or bad the poll itself is.

********

*Survey Question: would you rather have women raped and murdered, or beaten by their husbands?
Published Survey Result: 66% of survey respondants want women beaten by their husbands!

**Unless the survey is about some "popular culture" theme. Sports, Celebrities, TV shows, movies, etc., which can cause unreasonable levels of emotional involvement in otherwise normal people.