Second Thots

Sometimes one has to step back, take pause, and have some "second thots"

Friday, October 22, 2004


What is going on with these presidential polls?

Bush is up by eight. No wait, he's up by two. No wait, it's a dead heat. This year's election polling is as confusing as it's been in recent memory. For those of us who like to follow an election as though it was a sports league, this year's campaign has been more like a demolition derby than it's been a traditional horse race.

It would be tempting to simply dismiss some of these disparities as the result of junk science and wait for the real poll on November 2nd. However, the polling results so far haven't been as chaotic or as mystifying as a casual reading of them might suggest. There have been some trends. There continue to be some trends. Sorting through some of them might help us to keep score in this blood sport others call politics.

First, there appear to be two kinds of current polling results. One has Bush up by as much as eight points. One version of the latest Harris poll shows the race heading in this direction. The other has the race much closer, suggesting a dead heat. Ironically, another version of the latest Harris poll shows the race heading in this direction. Hey, don't blame me for the confusion.

So, judging by what these two types of polling results tell us, Bush is either ahead of Kerry by a significant margin of difference, or he's tied with Kerry with no significant margin of difference. In either instance, Kerry does not appear to have a lead in any of the major polls published since the end of the third and last debate. This is a discernible trend.

In fact, this trend has to raise an eyebrow. Because what we see is a race where Kerry seems to be hanging on to any glimmer of hope that he's still in a tied race. Indeed, it's remarkable how many of these sorts of polling results show a precise dead heat: 47-47, 45-45, 46-46, etc. If this truly was a dead heat, one would think that at least one poll would show Kerry with some kind of a lead (AP-Ipsos just came out with one - a slight Kerry lead. The only one to do so). If I didn't know any better, I would think some of these pollsters are trying real hard to keep Senator Kerry in this race.

The difference in the two kinds of results seems to stem from a difference of opinion as to what a likely voter is. Nowhere is this phenomenon more clearly illustrated than with the Harris poll. When they use a strict definition in their poll question, Bush does much better. When they use a less strict definition in their poll question, leaving the possibility that more people will vote this year than in years past, Kerry tends to pull even. Judging by this kind of pattern in the polling results, a conclusion one could draw is that Kerry can only win if he convinces enough people to get off their behinds and vote.

This conclusion seems to be supported by the behavior of the campaigns in recent days - one of the greatest indicators of polling accuracy. Bush hasn't done anything much differently than he's done over the course of this entire election season. He has tried to use the liberal label. But that's a ploy not much different than trying to use the flip-flopper label. Kerry, on the other hand, has become decidedly more aggressive. He has directly accused Bush of wanting to implement a military draft once elected. He has also accused Bush of wanting to implement private Social Security once elected. These are two tactics meant to get people scared or angry enough to go out and vote against Bush.

Kerry's other hope is his performance in the so-called battleground states, where significant victories could override national polling trends.

However, leaving aside the fact that even in these states trends tend to ultimately follow those at the national level, there is one wildcard in all of this he hope better is not in play: voters' underlying trust in the way Bush has handled matters of national security and terrorism. In many ways, it's just not trendy to vote Bush right now. Iraq is a mess; the nation divided; the world in opposition; many people angry at Bush for many different reasons. Yet throughout this period, America has not been hit again. And a lot of those terrorists are in Iraq right now - not in New York, Washington, or Los Angeles. Ironically, Bush might not seem very popular to some, but he may be a safer choice nonetheless. They know who he is and what he's done. The same can't be said for Kerry. Why take the chance at the unknown in already uncertain times?

Polls are polls. They give us numbers. It's up to us to try to determine what they mean. Kerry is hoping they mean people will come out in waves to vote for him. Some of the evidence out there suggests that might not happen, especially at this late point in the campaign.

I Believe you covered this. Theere are pollsters that favor a liberal agensda and there are pollsters that favor a conservitive aganda. However, if these polls are truely scientific why are they not coming up with more similar numbers.


Well, actually, polls are probably as scientific as so-called social sciences go.

Even in more 'straight' sciences, such as physics and biology, there is disagreement among the 'experts'. Global warming might be a perfect example of this.

When it comes to polls, the differences generally lie in the kind of sample used for the results and the interpretations made in analyzing them.

There can be genuine disagreements as to how polling should be done.

Of course, given the differences of interpretation, just as in any scientific theory or venture, there is room for things like bias or even manipulation, which is probably the greatest single source for skepticism about polls in the public.

As with most things, people should look at polls with a critical mind. Having said that, in the end they're usually right. They have to be, or the companies will go out of business. So, don't be surprised if you see some of the polls start looking very similar right before election day. The pressure of being proven right usually wins out over any kind of bias.


Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?