Are you a part of The WOKV Insider? Sign in or join now. Why join?
Text size: A A A
Jamie Dupree

Polling Accuracy In 2008

By
Jamie Dupree
@ November 7, 2008 12:00 AM
Permalink | Comments (0) | TrackBacks (0)

We talked a lot this year about whether we should believe the polls for the November election or not, and a post-election review shows some pollsters certainly did a lot better than others.

A study by Fordham University's political science department shows that Rasmussen and Pew were the best in terms of predicting the final popular vote.

The average of 23 polls predicted an Obama win of 7.52%. The final outcome is an Obama win of a little over 6.1%.

Realistically, the polls did not botch this election, say like they did in the New Hampshire Primary.

There were three battleground states which were shown to be basically dead-heats, and the final results in North Carolina (just called on Thursday by the AP), MIssouri (still not awarded by some news organizations) and Indiana all were settled by very narrow margins.

Among the more popular polls that did well, the George Washington University Battleground poll, done by a GOP pollster (Ed Goeas) and a Democratic pollster (Celinda Lake.)

CNN was #6 in the Fordham study, while Fox News came in 10th of 20 pollsters reviewed.

I do have to point out that the bottom five are some big names: CBS, Gallup, Zogby, CBS/NY Times and Newsweek.

Ouch.

Zogby had Obama winning by 11 points in his last poll, as did Gallup. CBS was at 9 points as was ABC/Washington Post polling.

What was interesting is that many polls overestimated the vote for Obama, and that was something we saw during the primaries as well.

Meanwhile, most of the polls underestimated the vote for McCain, with many of them giving him 42, 43 and 44 percent right before the elections.

In terms of state by state polls, the pollsters all had Obama by way too much in Iowa, and way too little in Nevada and New Mexico.

In key states, the polls were pretty close in Florida, but underestimated the Obama margin in both Ohio and Pennsylvania. The polls also were too low for Obama in Colorado.

The polls were dead on for Virginia, Missouri and New Hampshire.

As for polls that were too low for McCain in states he won, that definitely happened in Arkansas, West Virginia and Arizona. The polls were too high for McCain in South Carolina and Texas.

So overall, it was a pretty good night for the pollsters with few really bad data points.

One thing I would like to know more about would be the exit polls, because in a number of states where McCain won easily, it seemed to take the networks a long time to call the state for McCain.

Georgia, Texas, Mississippi, South Carolina and Kansas are just some of those that I jotted down during the night, while Obama's calls seemed much more speedy, occuring just after the polls closed in those states.



0 TrackBacks

Listed below are links to blogs that reference this entry: Polling Accuracy In 2008.

TrackBack URL for this entry: http://wsbradio.com/blogging/mt-tb.cgi/11895


Avg. rating: N/A

What others are saying

  • link to the report
    could you provide a link to the study by Fordham University's political science department?

    thanks.
send to a friend  view as printer-friendly  RSS feeds
advertisement
advertisement

Marketplace

Stay ahead of the storm. Find evacuation routes, safety tips and more in the Hurricane Guide.
Special Car Deals
Special Deals on Cars And Trucks! Check out these really hot internet deals!
WOKV Store
Books by Neal Boortz, Sean Hannity, and Clark Howard are now available in the WOKV store!
Experts

Our weekend experts are live and local,
and ready to answer your questions!
Gas Prices

Before you fill up check out Jacksonville's lowest gas prices!
advertisement
WJXT Partner

WOKV is a proud news partner of WJXT-TV 4
Fox News Partner

WOKV is a proud news partner of Fox News
Travel Channel
Need to get away or planning a vacation? Let us, and our partners at Travel Channel, help. Click here to find out more.