QandO

June 10, 2004

Polling Irregularities
Posted by Dale Franks

Tom Bevan tries to make sense of today's Los Angeles Times poll results. He's not having much luck at it.

Bush with an eleven point lead in Missouri? Not a chance. Even the biggest Bush booster in the country wouldn't claim that the President would win the Show-Me State by double digits as things stand right now. The latest polls show Bush has, at best, a tiny lead over Kerry.

In Ohio, three consecutive polls conducted during the last three weeks - including a Mason-Dixon one with a pretty big sample - show Bush with a small lead, but the LA Times has Kerry up by three.

Likewise, the last three polls in Wisconsin - albeit two from Zogby's online operation and one by a Dem firm - have Kerry ahead by sizeable margins. The LA Times has Bush up two. I suppose it's possible Bush has a small lead in Wisconsin, but not likely. With only two exceptions over the past three months (both Badger polls which I've been told over samples Republicans, by the way) the state has been leaning toward Kerry.

Finally, just from a common sense standpoint the LA Times state data is at odds with its own national results. If John Kerry really is leading by 6 or 7 points nationally then there is simply no way Bush is winning Missouri by 11 and Wisconsin by 2.

One ofthe things we saw during the Gray Davis recall election was that no matter what the Field Poll or Opinion Dynamics, or anyone else said, the Times always had Gray Davis slightly ahead, or within striking distance.

Naturally, Davis got stomped like a hippie at Altamont.

There is something decidedly odd about the Times polling unit. I think it pretty consistently oversamples Democrats, and gives skewed results. And, when you see similar outlier-type, in polls taken over the course of whole year, on a number of different issues and elections, you have to wonder if the LA Times isn't perfectly aware of this, too.

I mean, if your polls are always the outlier numbers, you have to wonder about the reliability of your results, don't you?

TrackBack