Here's a classic example of why you can't trust polls. The attached Rasmussen poll trumpets "Just 53 Percent Say Capitalism Better Than Socialism."
Wow! That sounds like Americans have had it with capitalism and are leaning toward socialism.
But when you start peeling it back, you discover that only 20 percent said socialism is better -- a full 27 percent said they didn't know which is better. It goes on to say that the survey did not define "capitalism" or "socialism."
Furthermore, Rasmussen references an earlier study which showed 70% of Americans prefer a free-market economy. This suggests that there's a significant lack of understanding among the population when it comes to macroeconomics. Generally, speaking, if Americans are asked to make a value judgement between multiple choices that they don't understand, they'll usually pick the one that has "free" in the title because it sounds the most American.
So if a population is being asked to make a value judgement regarding two terms that it doesn't understand, how valid is the survey? The study reflects perception, not understanding.