Rahul nails it and provides the beginning of a roadmap, maybe.
http://www.empirenotes.org/December 27, 2004
Thinking Beyond the Comfort Zone -- Polls
Here's my latest radio commentary for Uprising:
To me, there’s only one good thing that can come out of the shattering defeat we all suffered on November 2. That is a recognition that so badly have the left, the antiwar movement, progressives of any stripe failed that we must change the way we do things in its entirety. And so I continue beyond the comfort zone.
One change: the left must come to terms with American public opinion, and, in particular, with polls. Too often people confuse the descriptive with the normative; if a poll says, for example, that a majority of Americans favor continuing the occupation of Iraq (as all recent polls do), activists want to contest the results. The most common way is to contest the methodology: “How can a poll of 1000 people represent the views of 300 million?” Curiously, when polls go the other way, activists often take them as absolute signs that the people are with them.
Much poll skepticism is unwarranted. The claim that polls accurately represent public opinion is based on two things. First, an assumption that the population is being randomly sampled. Second, simple mathematical analysis. If , say you have polled 1000 people, randomly selected, 19 times out of 20, you will get within 3% of the numbers you would get if you had asked the whole country. In fact, polls do have systematic errors, like the fact that they don’t usually reach people without phones, and those can throw off results by a few percent; but there’s really no chance that polls, as currently conducted, significantly misrepresent the answers of the entire country to the particular set of questions asked.