Liberals are happily quoting a New York Times poll that says that 60% favor the right to collective bargaining for public sector workers. However, 25% of the 984 respondents say that their family contains someone who works in the public sector. Is that rather small sample a typical cross section?
Out of 300 million people in this country, you can interview 984 people, 25% of whom can be considered to be clearly biased, and claim that you have determined the opinion of those 300 million people. And I have some property in Florida that I'd like to talk to you about. It has an ocean view.
Update, Friday 7:30am: And Misinterpreting Them
Polls are all too often taken, that is data gathered, without context. During the "health care reform" debate a poll revealed that the "public option" was immensely popular, being favored by some 85% of respondents. That was taken to mean that "health care reform" itself was popular. But when the question about "health care reform" itself was asked in another poll, only some 43% of respondents said they had a favorable view of the subject, and only 3% said it was their "top priority."
Polls mean whatever the pundit quoting them wants them to mean.