Search This Blog

Tuesday, October 26, 2010

Breaking: Siena switched likely voter method for today’s Gibson/Murphy poll, but failed to point it out

By Sam Foster

Improbable 26 point Chris Gibson surge?  Or, Siena switching their methodology for sampling likely voters?  Brian Mann at NCPR is reporting that Siena changed up their likely voter methodology prior to the Siena Poll on NY-20:

Pollster Steve Greenberg says you can't draw an exact apples-to-apples comparison between this survey and the last one in September.

The reason is that Siena used a slightly different voter model in this survey, so that it's difficult to draw a clean contrast.

"I would say it's a Macintosh to a Red Delicious [comparison]," Greenberg says.

"This poll today has a more stringent 'likely voter' sample in it," he added.

Greenberg is confident that their methodology was able to capture a good snapshot of which voters are more likely to vote,  and how that shapes the race now seven days from the November midterms.

The revelation is actually quite stunning, since Siena is pretty much the only non-partisan polling releasing data for NY.  Just look at the numbers from some of the previous polling:

10/20 – Cuomo, 37 points over Paladino
10/17 – Maffei, 12 points over Ann Marie Buerkle
10/13 – Owens, 5 points over Doheny

If Siena had previously skewed so far Democrat in the Gibson/Murphy race, what surprises will we find on Nov. 2nd?


Post a Comment