What do the polls really say about the NY-23 race?

One of my favorite polling websites is Pollster.com, the site acquired recently by Huffingtonpost.  Scanning their data on Friday, I stumbled across an interesting — and sketchy — bit of information.

Pollster lists NY-23 as leaning Republican, suggesting that challenger Matt Doheny has a clear edge over Democratic incumbent Bill Owens.

Digging deeper, I found that the site claims that the “latest poll” gives Doheny a 46-39% edge in the race and claims that he’s very nearly a lock, with an 85% chance of winning.

The trouble is, so far as I can tell, that number is based on older surveys conducted by Republican pollsters.  (One such poll gave Doheny a 51-37% advantage.)

The latest poll of the race that I can find — and certainly the most recent survey conducted independently — was Siena’s survey released October 13th.

They found that Owens was leading by 5%.

And when Siena factored in the lingering Doug Hoffman effect (a significant number of voters weren’t aware that the Conservative had dropped out of the race) Owens’ lead climbed to 11%.

That’s a huge swing, from Siena’s 5-point lead for Owens to Pollster’s 7-point lead for Doheny.  So what gives?

I really don’t have an answer.  There has been a lot of fresh skepticism lately about polling methodology, the influence of partisan polls, etc.

In the end, I suspect that the outcome of this race will be very, very close.

Tags: election10