The first sign of a scam poll is if it samples "adults."

There are many reasons to sample adults' opinions on various issues during the year, but less than thirty days before an election, to conduct or publish such a poll on the election is nonsensical. We know for a fact most of these people won't vote, so the point of asking how they might vote IF they did is . . . what, exactly?

The only sample which matters this close to an election is the "likely voters," because they are the ones whose opinion counts at the ballot box. The category isn't difficult to screen, but performing an accurate screening would eliminate many respondents and therefore force more contacts to generate a true sample, driving up the costs for the polling firm.

They have the choice to provide a more accurate poll at a higher cost, or a less accurate one which costs less to produce. It seems many of these polling outfits have clearly made that choice, and decided to poll on the cheap.

D.J.'s earlier examination of polls is also recommended.

~~~~~~~~~

Some may speculate there are ulterior motives at work, that left-leaning media and Beltway pollsters wish to reinforce their meme of a Democratic sweep this year by depressing Republican turnout. Whether or not such intent is present, Republicans can best serve their country by turning out in force Novemeber 7th, and letting the chips fall where they may.

Even in these polls, over 60% say they think their own representative is doing a good job. How that meshes with the results saying they will support the Democratic candidate in their district is a wonder!! Unless their rep is a Democrat. Similar contradictions were seen in other polls. For example, most say economically their region and themselves are doing well, yet they say Bush is doing a bad job on the economy??? They say the Iraq war was a mistake at the same time they say it has increased our security.??? Do they not like security then?? Quite simply, I think the average poll answerer is distracted, in a hurry, not understanding the questions clearly, or just so uninformed that they are easily led. Most won't vote anyway. Do you believe that 75% number out today in one of the polls?? Then I have a bridge to sell to you. We didn't get that number in the presidential election. Take home message: ignore polls and just go out and vote.

I've been saying all along the polls aren't screening likely voters "hard enough." Obviously, if you come up with a number like 75% "sure" they will vote, you are only "screening" by asking the respondent. Most people will tell you they will vote, even if they aren't registered, because they may be a bit ashamed to admit they don't.

I believe the last time a majority of eligible voters voted in a midterm was either 1966 or 1970 - my memory could be faulty on that point, though. I do know that the 2002 midterm's participation of about 40% was a high point in recent midterms.

http://newsbusters.org/node/8221

Also, look at some of the samplings used.

Hmmmmmmm.....