Scientific polls are typically constructed with more forethought as they are generally part of a funded study. There will be questions that will seem biased to the person being interviewed, but they are strategically placed within the poll to ensure overall, a more accurate result. I have been part of putting together research polls and if conducted properly, they can be very accurate and very informative. Now the ones we see here on the TFP, television, in magazines, etc. are a different story. That is why you see that margin of error percentage - to account for poll placement and those who may have not answered truthfully or who may have felt pressured to give a certain answer. For instance, if you post a poll about beef on a site frequented by vegetarians, your outcome will be much different than had you posted that same poll on a website geared toward the hamburger buying public. Research - grants, studies and soforth - depends upon poll accuracy, so they can't all be crap.
