The WaPo Pattern
Saturday, October 29, 2005 :::
Now I see what Chad's been up to the last few days.
Here and, graphically, here, Chad takes a look at WaPo polling during the three previous gubernatorial elections and discovers that not only do they underestimate GOP strength, but the final spread as well.
There could be any number of reasons for this, some of which can be found in the comments to the first link, but this is not uncommon. Polling is an art that wraps itself in the cloak of science. And like any art, it reflects the viewpoint of its creator. Some polls are more accurate than others, to be sure. But they all have similar failings -- the biggest of which is the growing refusal rate. Blame it on the Do Not Call list if you will, but for several years, the refusal rate has risen to the point where pollsters have to contact more and more people just to reach minimum participatory levels. And that has an impact on results.
And then you have a group of evil people like me, who will outright lie to pollsters (particularly if it's a poll from a candidate's campaign). Usually say I'm voting for the third party candidate (not this year). Once in 2000, a crestfallen GOP campaign pollster asked whether I intended to vote for George Bush. I said no, I was leaning toward Nader. "But you're a Republican" went the shocked reply. Maybe I am, maybe I'm not. Usually, I don't make the decision until I'm standing over the ballot.
::: posted by Norman Leahy at 10/29/2005