Comments on: Club For Growth Poll Shows Three-Way Statistical Tie In NY-23

By: Nancy Hanks

Nancy Hanks — Sat, 31 Oct 2009 09:55:00 +0000

Yes, this poll brings new meaning to the term “margin of error”!!

By: Trescml

Trescml — Fri, 30 Oct 2009 11:30:41 +0000

Never trust a poll that the margin of error goes out to a 100th of a percent.

By: kranky kritter

kranky kritter — Thu, 29 Oct 2009 02:09:08 +0000

Very good advice, Doug.

By: Nick Benjamin

Nick Benjamin — Tue, 27 Oct 2009 07:06:33 +0000

The problem with most partisan and campaign polls isn’t that the campaign is necessarily lying. The problem is they only release polls that fit their narrative. DailyKos releases all their polls. They have to. No organization with that many members could release a poll just to members. They could try but it wouldn’t work too good.

I’m very skeptical of this particular poll. Club for Growth does not release most of their polls. This particular one has a minuscule sample size, and shows a 14-point swing from Owens to Hoffman in 3 days. It also fits the narrative Club for Growth wants perfectly.

So I wouldn’t be surprised if Club for Growth actually commissioned several polls, nudged the questions so that Hoffman voters were disproportionately represented, and then released the best one.

That said polling a special election well is virtually impossible. You just don’t know whose going to show up. Polling a three-way special election is even harder. This poll seems to imply the Republican is dead at 20%, but if you nudge the numbers his way five points he’s actually 2nd, behind the Conservative by a single point.

By: Gaucho Politico

Gaucho Politico — Tue, 27 Oct 2009 05:37:16 +0000

i wouldnt discount the r2k poll simply because it comes from kos. your going to need more than the source to discount polling especially when its conducted by research 2000 and not the site itself. this is especially true when the r2k poll is in line with the sienna poll you find is reliable.

the r2k poll is very transparent. it asks the follow up question about where the votes go if you eliminate hoffman. a third wont vote and the rest are undecided.

the club for growth poll isnt releasing any of the cross tabs which would be useful in evaluating how useful it is. i agree that the small sample size is disconcerting. in addition it is a “likely voter” poll so that also changes depending on how they are deciding what a likely voter is. so i wouldnt wait it that heavily but i wouldnt fully discount it.