« Who Is Insider Advantage, and What Do They Know About Florida? | Main | Ginny Raises $350K in Third Quarter »
Saturday, October 16, 2004
Final State Polls from the 2000 Election
Posted by DavidNYCCADem took a look at the final state polls for each state in 2000. His hard work shows that the swing states looked like this:
Arizona: Last poll showed Bush up by 10 points, outside the margin of error. Bush won by 6.Arkansas: The last three, concurrent, polls showed a tie, Bush by 1 and Bush by 3, respectively. Bush won by 6.
California: Also three polls, showing Gore up only by 3, 5 and 7, respectively. Gore won by roughly 12.
Delaware: Last poll had Bush up by 4. Gore won by 13. Great poll, guys.
Florida: The last poll, Zogby's, showed Gore up 3. Well, you know what happened.
Iowa: Two last polls showed Gore up by 1 and 2 points, respectively. Gore won by a hair.
Maine: Last poll showed a 42-42 tie. Gore won by 5.
Michigan: Last poll showed Gore up by 7, but still within the margin of error. Gore won by 4.
Missouri: A split. Of the two last polls, one showed Gore up 1 point and the other showed Bush up 4. Bush won by 4.
Nevada: Last poll had Bush up by 4. Bush won by 3.
New Hampshire: Last poll had Bush up big, by 10. Bush only won by a hair.
New Mexico: Two last polls showed a tie and Bush up by 3, respectively. Gore won narrowly.
Ohio: Two last polls, one by Zogby (who was right on nationally), showed Bush up by 9 and 10, both outside the margin of error. Gore lost by 4.
Oregon: A split. Last two polls showed Gore up by 1 and down by 4, respectively. Gore won by a fraction of a percent.
Pennsylvania: Last poll showed Gore up by 2. Gore won by 4.
Tennessee: The writing was on the wall. Two last polls showed Bush up by 4. Bush won by 3.
Virginia: Last poll showed Bush up by 6. Bush won by 7.
Washington: Last poll showed Gore up by 4. He won by 5.
West Virginia: Of three last polls, two showed Bush up by 10 points, and one had Bush up only by 2. Bush won by 6.
Wisconsin: Last poll showed Bush up by 2. Gore won by a hair.
As you can see, there are a couple of states in there (CA & DE) which truly are not swing states this time out. And some nominally swingy states this time around (CO, NC, LA) didn't make CADem's list.
Posted at 11:12 AM in General | Technorati
Comments
So the conclusion:
Don't bother watching polls?!
Posted by: dougyg at October 16, 2004 01:02 PM | Permalink | Edit Comment | Delete Comment
Maybe we should be watching yard signs and bumper stickers instead. I drove to Winston-Salem today, and saw far more signs for Bush and Burr--the exact opposite of my part of the state, where 90% of the signs and bumper stickers are for Kerry and Bowles. I've concluded that it's because of Winston-Salem and Charlotte that NC will be a red state.
Posted by: pepe at October 16, 2004 04:44 PM | Permalink | Edit Comment | Delete Comment
Ralph Nader's the only one that doesn't get the negative impact he's going to have on the progressive movement. We've got to do everything we can to stop him before it's too late!
"I am voting my conscience" and not supporting Nader
In the following statement, Winona LaDuke, Nader's 2000 running mate, announced:
I am voting for John Kerry this November. I love this land, and I know that we need to make drastic changes in Washington if we are going to protect our land and our communities. I'm voting my conscience on Nov. 2 I'm voting for John Kerry. He wants to move federal policies to support Native communities, whether Native farmers, businesspeople or tribal governments. We are on his radar; this is a beginning. Kerry offers other reasons for hope. He opposes converting Yucca Mountain into a nuclear waste dump. By Nov. 2, 2004, John Kerry will have earned my vote
Posted by: Ace Parsi at October 16, 2004 05:45 PM | Permalink | Edit Comment | Delete Comment
If we look at the average difference between last poll and result, the _average_ is 1.5% better for Gore than the polls, which is close to the national polls result. But the variance of the result is much greater than that. I think the most important thing to learn from this is that the margins of error in the polls are not the _maximum_ error, but just slightly more than the _average_ error. If we give Gore this 1.5% on all polls, there are still 9 polls (out of 32) that differ more than 4% from it (a typical margin of error for a state poll).
I think that that is the most important lesson from this: one quarter to one third of all polls will be off by more than the margin of error, even after the best possible correction for the effects of the period between last poll and the actual election.
Posted by: Andre at October 16, 2004 06:17 PM | Permalink | Edit Comment | Delete Comment
I would suggest the banning of publication of polling data 7 days prior to election day. This limits the impact polling has on the outcome. Of course, the campaigns can still conduct private polling for strategic purposes. Any comments?
Posted by: bigguy at October 16, 2004 10:36 PM | Permalink | Edit Comment | Delete Comment
I think each state has its own character. Perhaps we should do some extra work to interpret polls correctly.
These states can be divided into 3 groups:
1 Demcrats candidate was undervalued
2 Republican candidate was undervalued
3 Neither was undervalued.(the difference between polls and the actual results is only 1% or less)
A. The following are the states falling into the first group(total 11 states):
California: Also three polls, showing Gore up only by 3, 5 and 7, respectively. Gore won by roughly 12.
Delaware: Last poll had Bush up by 4. Gore won by 13. Great poll, guys.
Maine: Last poll showed a 42-42 tie. Gore won by 5.
New Mexico: Two last polls showed a tie and Bush up by 3, respectively. Gore won narrowly.
New Hampshire: Last poll had Bush up big, by 10. Bush only won by a hair.
Pennsylvania: Last poll showed Gore up by 2. Gore won by 4.
Ohio: Two last polls, one by Zogby (who was right on nationally), showed Bush up by 9 and 10, both outside the margin of error. Gore lost by 4.
Wisconsin: Last poll showed Bush up by 2. Gore won by a hair.
Nevada: Last poll had Bush up by 4. Bush won by 3.
Washington: Last poll showed Gore up by 4. He won by 5.
Tennessee: The writing was on the wall. Two last polls showed Bush up by 4. Bush won by 3.
B. The second group states(8 states):
Arizona: Last poll showed Bush up by 10 points, outside the margin of error. Bush won by 6.
Arkansas: The last three, concurrent, polls showed a tie, Bush by 1 and Bush by 3, respectively. Bush won by 6.
Florida: The last poll, Zogby's, showed Gore up 3. Well, you know what happened
Michigan: Last poll showed Gore up by 7, but still within the margin of error. Gore won by 4.
Iowa: Two last polls showed Gore up by 1 and 2 points, respectively. Gore won by a hair.
Missouri: A split. Of the two last polls, one showed Gore up 1 point and the other showed Bush up 4. Bush won by 4.
Oregon: A split. Last two polls showed Gore up by 1 and down by 4, respectively. Gore won by a fraction of a percent
Virginia: Last poll showed Bush up by 6. Bush won by 7.
C. Third group:
West Virginia: Of three last polls, two showed Bush up by 10 points, and one had Bush up only by 2. Bush won by 6.
Posted by: brightwater at October 17, 2004 11:26 AM | Permalink | Edit Comment | Delete Comment
I believe in blue states Dems are tend to be undervalued. It's the same in red states for Republicans.
Posted by: brightwater at October 17, 2004 11:30 AM | Permalink | Edit Comment | Delete Comment
Good observation, brightwater.
Posted by: DavidNYC at October 17, 2004 12:50 PM | Permalink | Edit Comment | Delete Comment
Nader allthe way
Posted by: Cody Dickson at October 22, 2004 10:49 AM | Permalink | Edit Comment | Delete Comment
I agree with bigguy, polls should be banned just prior to an election. With the current campaign finance reform, they just become another political tool to swing public sentiment.
That being said, have any of you seen or heard of anyone else studying the accuracy of the polls vs. election results for say, the last 5 presidential elections? It would be interesting to know if the polls are becoming more or less accurate at predicting results as we transform into a more mobile society.
Posted by: PackFan at October 27, 2004 09:01 AM | Permalink | Edit Comment | Delete Comment