What Do the Polls Mean?

wallpaper_napoleoncouch.jpgI spent last weekend in Las Vegas with my co-blogger Heather and four other John Edwards supporters, walking precincts in Vegas and Henderson, Nevada. We thoroughly covered six large precincts in a variety of types of neighborhoods and had a great time. We did not, however, come back full of hope for Edwards’ chances in Nevada.

Wherever we went, we found the same thing: Clinton and Obama supporters. Many, particularly Obama supporters, cited Edwards as their “second choice.” But the number of individuals we met who intended to caucus for Edwards were hardly worth talking about.

Last week I watched the polls and, though I don’t now remember the numbers, kept hearing that the race was nearly a three way tie or that Edwards was expected to get up to 20% of the vote. I kept shaking my head, wondering if our experience was some sort of aberration.

On Friday I ran into another of the women who traveled to Vegas with us and compared notes. Had I been imagining things? Was there more support in Vegas for Edwards than what I had seen? She didn’t think so. We agreed it was unlikely Edwards’ support among caucus goers would get to the double digits, in spite of what the polls had been saying.

Last night we learned the experience of my friends and me on the ground was accurate, and the polls were wrong.

I’m not qualified to discuss the issues of sample size and methodology, nor am I inclined to disbelieve in polling altogether. All I know is that in the case of Nevada, my “lyin’ eyes” on the ground were a whole hell of a lot more accurate than what the pollsters predicted.

  3 comments for “What Do the Polls Mean?

  1. January 20, 2008 at 4:02 pm

    Yes, I learned a valuable lesson from this though. It’s amazing how much you can learn by canvassing for your candidate. It was well worth the experience.

    Thanks for the thoughts Gila.

  2. thesunkenroad
    January 20, 2008 at 4:21 pm

    Sorry for your disappointment. I agree with what you’re saying, but I think the polls might have had the story better than you think right now. By caucus morning, I was convinced Clinton would win and what you describe was a big part of why I thought it. It’s not science, just a gut-feeling from watching the polls move and hearing what the Clinton’s were telling union rank-n-file. Then Obama made is Reagan comment, which I thought showed he likely understood that party faithful were more in Clinton’s column. If you look at the last polls in the RCP site, Edwards’ numbers suddenly dropped like a rock in the last day or two–or at least I thought they did. Clinton’s rose more than Obama’s. And during the caucus, apparently union members broke evenly for Obama and Clinton. Media guaranteed the two-person Obama/Clinton finale, in my opinion. They mostly ignored Edwards.

  3. January 20, 2008 at 5:08 pm

    H: yes, there’s a LOT you can learn canvassing! No wonder Loretta has been out doing it!

    James: I suppose the polls do lag, which could explain part of what I saw. But Heather and I were there a full week before the caucus and at that time the support for Edwards was simply not there. Also, the records we were working with showed that he hadn’t had much support in the previous weeks either.

    I’m not what you’d call a polling skeptic in general. What I *am* a skeptic of is claims of voter fraud based on actual voting results ending up pretty dramatically different from what the polls have said.

    A poll is a model, and the accuracy of the poll is only as good as the design of the model. Sometimes, for whatever reason, the models are not so good.

    In the case of Edwards and Nevada, I don’t know of the polls were simply lagging badly, or if the models were poor, or what. But when I was “on the ground” one week before the caucus I was able to put my finger in the wind and tell pretty precisely what was going to happen a week later.

Comments are closed.