A note about polling
Now that the election results have sunk in, we thought we'd see how they matched up with pre-election polling, including the Herald-Leader/Action News 36 Election Poll. Here's how they break down:
With 99.9 percent of precincts reporting, Gov.-elect Steve Beshear got 58.7 percent of the vote, compared to 41.3 percent for Gov. Ernie Fletcher. Beshear's margin of victory was 17.4 points.
The final H-L/WTVQ poll, conducted by Research 2000, showed the race at 55-40, with 5 percent undecided. If you allocate the undecideds based on the ultimate outcome (that's what pollsters tell you to do), our poll would have shown the race at 57.9-42.1, a 15.8 point margin. That puts the H-L/WTVQ poll 1.6 percentage points away from the ultimate outcome.
The final Rasmussen Reports poll showed the race at 54-39, with 7 percent undecided. With the undecideds allocated, the poll would have shown the race at 58.1-41.9, a 16.2 point margin. That puts the Rasmussen Reports poll 1.2 percentage points away from the ultimate outcome.
The final SurveyUSA poll showed the race at 59-39, with 2 percent undecided. With the undecideds allocated, the poll would have shown the race at 60.2-39.8, a 20.4 point margin. That puts the SurveyUSA poll 3 percentage points away from the ultimate outcome.
The final Courier-Journal Bluegrass Poll showed the race at 56.2-33.4, with 10.4 percent undecided. With the undecideds allocated, the poll would have shown the race at 62.3-37.7, a 24.6 point margin. That puts the Bluegrass Poll 7.2 percentage points away from the ultimate outcome.
So, Rasmussen was the most accurate poll, followed closely by the H-L/WTVQ poll.
If you're uncomfortable allocating the undecided voters and would rather throw them out, the H-L/WTVQ poll and Rasmussen Reports poll tie as the most accurate. The 15 point margin in those polls was 2.4 percentage points away from the ultimate outcome, compared to 2.6 points for SurveyUSA, and 5.4 points for the Courier-Journal Bluegrass Poll.
Needless to say, we're pleased with the accuracy of our polling and hope to be so lucky next time.
- John Stamper


The SUSA poll was far and away the most accurate, John.
Posted by: Media Czech | November 08, 2007 at 12:38 PM
John - Your argument suffers because the polls rounded their numbers, so you can't honestly argue 1.2 versus 1.6.
How do you know that Survey USA wasn't 58.6 to 39.4? You don't. There's no need to be defensive about your poll. It did a good job as did Rasmussen and SUSA.
Posted by: | November 08, 2007 at 12:58 PM
12:58:
I just thought it would be interesting to go back and look at the numbers. I agree with your assertion that there's no real difference between 1.2 points and 1.6 points. Rasmussen and SurveyUSA were both very accurate. The C-J was a bit off, but that happens to everyone sometimes.
I'm not trying to be defensive. If that were the case, I would have highlighted the fact that BluegrassReport.org chose to repeatedly question the accuracy of our poll. The numbers speak for themselves.
Posted by: John Stamper | November 08, 2007 at 01:19 PM
Not only were you pleased with the accuracy of your polls, you were pleased with the outcome of the election. C'mon, H-L staff, admit it. Each and every one of you probably voted for Beshear.
Posted by: | November 08, 2007 at 01:28 PM
Oooh gurl, they some drama up in here!
Posted by: | November 08, 2007 at 01:30 PM
John, the fact of the matter is, BluegrassReport.org has faded into obscurity, while the number of comments here has skyrocketed and far outdistanced BGR.
Speaking as someone who was delighted with the results Tuesday, whoever was right in the polls, I suspect the criticism from there is simply sour grapes, from an online location that used to matter much more than it does now.
BGR has no institutional continuity, and doesn't reveal when it is receiving payment to take a particular position. As such, I suspect its credibility has dropped severely. That site is surely aware of that problem, and doesn't have an answer to that weakness. Which is why I wouldn't be surprised to see it once again announce it is giving up its efforts.
Despite all the criticism you receive from both sides of the aisle, you've got a good thing going here. Keep it up.
Posted by: | November 08, 2007 at 02:18 PM
2:18. Bluegrass Report gets it funds from a number of sources, I am sure. As an invividual, I can tell you that I personally made a contribution to them and will again do so. Bluegrass was and is on top of every major story as it is unfolding. Yes, a little fluff, but 90% on target every time. Somebody has to counter the opposition. They will call a spade a spade and they do it extremely well. Best a kickin butt that I have ever seen.
Posted by: | November 08, 2007 at 05:34 PM
BGR will probably cease to exist at the end of this month. Wasn't the original funding from Brown, Miller, etc., to last only through November?
Posted by: | November 08, 2007 at 10:26 PM
I HATE TO BE THE BEARER OF BAD NEWS, BUT I TOTALLY AGREE WITH "MEDIA CZECH".
THE SURVEYUSA POLL IS SO faaaaaaaaaaaaaaar SUPERIOR TO THE OTHER POLLS IT'S NOT EVEN FUNNY.
DOING FUZZY MATH WITH THE POLL NUMBERS WILL not CHANGE THAT!
http://osi-speaks.blogspot.com/2007/11/once-again-surveyusa-poll-proves-its.html#links
Posted by: KYJurisDoctor | November 10, 2007 at 10:58 AM