« Pfizer abandons site of infamous property rights case | Main | MoveOn.org targets Marshall »

Good polling is hard to find

It's hard to find accurate polling. Nearly every poll says something different. Some polling firms are skewed some prefer likely voters over registered voters, so it's hard sometimes to get a good read on the pulse of the electorate.

Stuart Rothenberg of the Rothenberg Political Report is giving us a hand to find out which polling firms are the best and which ones not to trust, based on the most recent election:

The "Pollster of the Cycle Award," in my opinion, goes to SurveyUSA, which once again proved its worth, at least in pre-election polls. The firm's final Virginia numbers were eerily close - the firm showed McDonnell winning 58 percent to 40 percent in its Oct. 30 to Nov. 1 poll, just shy of the actual final margin: McDonnell 59 percent, Deeds 41 percent.

In New Jersey, where Christie won by 4 points, the last SurveyUSA poll showed Christie up by 3.

Public Policy Polling was the runner-up in Virginia (it had McDonnell up by 14 points), while in New Jersey PPP (Christie by 6 points) and Quinnipiac University (Christie by 2 points) were narrowly behind SurveyUSA in accuracy.

SurveyUSA, PPP and Quinnipiac, however, dramatically overstated the support of Independent Chris Daggett in New Jersey.

Where there are winners, there are usually also losers. None of the major public pollsters was dramatically wrong in Virginia, but Research 2000, which polled for DailyKos, showed McDonnell with only a 10-point lead in late October, primarily because it overstated Deeds' support.

In New Jersey, the Monmouth University/Gannett poll erred when it showed Corzine up by 2 points in its last survey. But by far the worst-performing survey in either state was Greenberg Quinlan Rosner Research for Democracy Corps in New Jersey.

Democracy Corps polling showed Corzine pulling ahead in his race in early October and stretching his lead to 4 points (41 percent to 36 percent for Christie) among likely voters and 5 points in a higher-turnout electorate in its Oct. 29 to Nov. 1 survey. The survey showed Daggett drawing in the midteens. He actually drew just less than 6 percent.

While not mentioned in this article, Strategic Vision polling should probably not be trusted given the controversy they find themselves in.

Post a comment


About JasonPye.com

Welcome to my website. The purpose of this blog is to talk about local, state and national news from a capitalist's perspective.
- To learn more about the author, please click here.
- E-mail: jason@jasonpye.com
- AIM: jasonpyedotcom


If you have news or a tip that would like to send in, please send it to news@jasonpye.com

- Posting and Comment Policy


Recent Comments

Powered by
Movable Type 3.2