Friday, September 10, 2004

Polls Schmolls

Argh, this week I have felt beseiged by polls, which I admit is in no small way connected to the fact that Dubya has seen convention 'bounce.' One out today says that the war in Iraq isn't having much of an affect on poll numbers. Over the weekend following the RNC, some polls were coming up with numbers saying Dubya got an 11 point bounce. Meanwhile, Zogby International put out polls taken in battleground states early in the week saying that, while there had been a little bounce, it was nothing like 11 points, and that Kerry still has an edge.

I admit that it's nice to see polls that go the way you want them to, but it's nicer to have a president you can get behind. Every time I hear poll numbers I have to wonder where the numbers come from. First of all, when they say that the war in Iraq isn't having much affect support for the administration I am first disheartened, because I wonder how it is that the war is seen as anything other than a complete debacle. And then I think, wait a second. What does that mean?

Well, according to the article it means that

"This support has steadily weakened over the past two years, but not in ways that suggest a direct correlation between casualties and political support."

So every dead soldier does not equal ten votes against Dubya. Instead, a political scientist interviewed said that votes may depend more on a "gut check" reaction, and that their response

will hinge on whether the public believes the Iraq war is winnable and whether the mission was worthwhile."

Part of that calculus is directly connected to the media, and whether or not it covers that topic this week, and how it is covered. As a pollster for Kerry puts it:

Bush does not frequently mention casualties, and the administration has restricted news coverage of returning flag-draped caskets -- a common image of earlier conflicts."

So, back to the polls, is it that casualties aren't affecting how people vote, or that casualties are not invading their living rooms on the nightly news, and that at 1000 dead, many people have not lost someone they know?

I also wonder what are the questions asked? All that push-polling in South Carolina in the 2000 GOP primary knocked McCain out of the race. Even when it isn't push polling, how you ask a question, what order it is asked, how it is worded all affect how people answer.

And then I think, you know, no one ever calls me for these polls. Yeah, yeah, nobody has ever known anyone who has ever been called. They make two hundred phone calls and then tell us what the country thinks. (a scary thing in and of itself). What are the chances that I would be called in a "representative sample"?

Well, zero, actually. The reason being that I don't have a landline. I only have a cell phone. Which means I'm not listed in a phone book. Which means no one calls me. It's great from the telemarketer standpoint- I've never ever been called by a telemarketer- but it also means that my voice never enters these polls. In a representative sample my one voice expressing my feeling on these issues won't change much-- but am I the only person who doesn't have a landline? I can think of five other people I know, in my age group, in my education level group, in my economic bracket, and on my side of the political fence, who aren't in the phone book for just this reason. So none of us are getting called.

This is not an argument for the polls being dead wrong. Perhaps people in their late twenties and into their thirties who don't have landlines are a statistical blip and would do nothing to change the results. But how many other people are out there not being asked who will show up at their polling stations on November 2nd?


Blogger fast eddie said...

Spoken like the sociologist that you are! Let's not forget that in 1948, all the polls predicted Dewey over Truman. Although they had stopped polling in October, the real problem was that when they took their sample, Gallup used the 1940 census to base their estimates about how many people should be drawn from rural and urban populations. But, there had been a huge migration to the cities during WWII, so they undersampled from urban areas, who were more democratic. Hmmm...what if we sampled all the people who don't have phones? Who aren't listed anywhere? Who don't vote? Might that change the numbers?

September 12, 2004 at 9:31 PM

Blogger fast & wily said...

Excellent! I read something today that said that at least some of the polls were of people "likely to vote," but that the determination of "likely to vote" was made by the pollster, rather than by a less subjective measure (such as registered voters). So what is the determination? People who answer the phone and don't hang up when their dinner is interrupted are more likely to vote for Bush? They are apparently also more likely to buy life insurance over the phone and change thier long distance plans at the urging of telemarketers. In other words, what the hell do polls mean other than that there are some people who have phones who will vote one way and some another. By the way, I also discovered today that telemarketers are prohibited by US law from calling cell phones, as the owner of the cell phone is forced to pay for minutes used on received as well as dialed calls. So none of these calls are going to cell phones.

September 13, 2004 at 8:44 PM


Post a Comment

<< Home